Keywords: Human-Machine Interaction, Human Factors, Human Performance Modeling
Abstract: Introduction of sentiment analysis, AI and machine learning (ML) based modeling of human-machine interactions can act as an excellent guide to assess the current strengths and weaknesses of human teams in a given paradigm and recommend targeted areas for improvement in team performance dynamics and situational awareness. This case study is conducted on human-machine interaction data from a team-based fire-fighting simulation environment. The primary goal of this research is to conduct sentiment analysis and ML on team communication data to gain deeper understanding of how teams behave and respond in stressful scenarios, how the levels of individual and team situational awareness influence the team dynamics, decision-making capabilities, and in turn impact the team performance (high vs. low) in the context of fire-extinguishing tasks. The AI/ML analysis of the relationship between the team performance variables and situational awareness scores reveal contributing factors that were the best predictors for team performance.

Keywords: Virtual and Augmented Reality Systems, Virtual/Augmented/Mixed Reality, Human-Machine Interface
Abstract: This brief paper describes an augmented reality tool as proof of concept of a system for increasing situational awareness during remote site exploration: the system allows the operator see an avatar of its vehicle through the obstacles, as if they were transparent, and adds nearby obstacles as seen by a Lidar sensor as 3D features to facilitate the perception of depth, a key point in making the whole concept usable. The paper presents briefly the idea, highlights the need and importance of a calibration procedure, and shows preliminary results achieved in a controlled laboratory environment using a Meta Oculus Quest 2 and an indoor motion tracking system.

Keywords: Passive BMIs, BMI Emerging Applications, Other Neurotechnology and Brain-Related Topics
Abstract: Situations of high cognitive workload can induce errors. This is widely prevalent in safety-critical domains where operators must make high-stake decisions while facing complex, dynamic situations. Therefore, it seems critical to detect instances of high cognitive workload to provide timely assistance to operators when they face such error-prone situations. This, however, requires capacities for real-time evaluation, which can be performed using brain-computer interface systems driven by measures of the central nervous system. Functional nearinfrared spectroscopy (fNIRS) measures have been shown to be associated with variations in cognitive workload, but these are typically processed and analyzed offline in a post-hoc fashion. Besides, to be used in operational situations, workload evaluation should be performed with mobile devices, supported by edge-computing devices. The goal of this paper is to present the first steps toward a real time, mobile assessment of cognitive workload using fNIRS data. We present two models developed using the Open Neural Network Exchange format that allows edge inference of cognitive workload with fNIRS data. We present prediction performance of the models and show the possibility to integrate the model into the Sensor Hub platform, a real-time sensor-agnostic data integration, synchronization, and processing nexus that allows sampling data from multiple sensors and users simultaneously for operational use cases. More specifically, we show that the model can produce workload inference from the fNIRS signal within the one-second latency requirement of the Sensor Hub platform.

Keywords: Decision Support Systems, Consumer and Industrial Applications, Enterprise Information Systems
Abstract: We consider the trend or pattern forecasting for demand timeseries in a business-to-business supply chain where demand exhibits high volatilities, non-stationarities, and skewness. We develop a pattern forecasting system by designing a data driven, feature dependent Markov chain-based framework. To increase adoption of AI based techniques among the various stakeholders we address the aspect of explainability. We define two metrices to evaluate the quality of explainability. To provide guidelines on selecting different attributes of our pipeline, we compare between feature selection methods from two families, one advanced and one traditional. We evaluate the proposed strategy on a real dataset and observe a sparsity promoting feature selection show a sparse feature selection method outperforms a conventional decision tree-based feature selection method.

Keywords: Human Performance Modeling, Human Factors, Human-Machine Interaction
Abstract: Technological advances in the aircraft continue to improve aircraft performance and extend the duration of operational missions; however, the physiological demand on pilots increases as well. With the increasing length and complexity of missions, monitoring constructs such as fatigue, attention, and stress become safety critical. This paper specifically looks at measuring fatigue and the experimental design to yield various levels of fatigue. There were 3 lighting conditions during data collection to represent varied lighting that occurs in operations. The experimental tasks and design are described, as well as the behavioral coding used during data collection. Lastly, current progress and future works are discussed.

Keywords: Service Systems and Organizations, Consumer and Industrial Applications, Communications
Abstract: Due to the global trend of digital government transformation, it is essential to understand the predictors of consumer adoption of e-government services and the underlying mechanisms to accelerate the digital transformation process. Although e-government services often collect personal information, no research examines its impact on e-government service adoption. Drawing on the stimulus-organism-response (S-O-R) theory, this research investigates the influence of the level of personal information sharing on consumers’ attitudes and intentions to use e-government services. Based on two experimental studies with 669 participants, this research finds that a high level of personal information sharing hinders the service adoption both directly and indirectly by worsening consumers’ attitudes toward service usability and the government. However, when consumers view the service as highly useful, their sharing of personal information enhances their service adoption by improving their attitudes toward the government’s personal information management. For female consumers, favorable attitudes toward service usability and personal information management and the availability of diverse functions lead to higher adoption intentions. These findings provide valuable insights for government officials and policymakers in designing effective e-government services that strike a balance between personal information requirements and consumer acceptance.

Keywords: Image Processing and Pattern Recognition, Deep Learning, Application of Artificial Intelligence
Abstract: Biomedical images are frequently used in articles to illustrate medical concepts and highlight regions-of-interests (ROIs). In many cases multiple annotation markers such as different arrows, letters or symbols overlaid on figures are often pointing to different Region-of-Interests (ROIs) relevant to the article. This work presents a proof-of-concept based comparative evaluation of three popular Deep Learning based object detection techniques (Mask R-CNN, YOLOv5, and YOLOv7) to identify such ROIs in a dataset of 450 Chest CT images which are appeared in biomedical articles. The results demonstrate that all three DL-based object detection techniques achieved high accuracy in recognizing and localizing the ROIs in biomedical images, whereas YOLOv7 exhibited the highest precision of 92.5%, indicating its ability to accurately identify ROIs.

Keywords: AI and Applications, Neural Networks and their Applications, Deep Learning
Abstract: Financial analysis reports contain a wealth of information, including details about a company's assets and performance, as well as analyst assessments. For subsequent analysis and valuation, information regarding a company's assets and performance is crucial. As a result, numerous securities firms seek to leverage artificial intelligence technology to profit from this data. We intend to resolve the issue of extracting and classifying table data from financial analysis reports by converting them to images. Financial report table images contain less information that can be extracted than typical images used in image classification research, and their sizes vary, making classification difficult with simple models. As a result, we tested a variety of classification models to determine which ones performed the best, and we propose a new model that combines their strengths. This new model achieved 94.5% accuracy and an F1-Score of 0.9414 on our self-collected financial analysis table image dataset, demonstrating its effectiveness in classifying table images in financial analysis reports.

Keywords: Autonomous Vehicle, Control of Uncertain Systems, Intelligent Transportation Systems
Abstract: The decision making and motion planning continue to pose many challenges to the reliability and safety of autonomous vehicles (AV), especially under complex and nondeterministic traffics. Nowadays, growing research on AV attempts to solve this problem by means of reinforcement learning paradigms while endowing machines with how to behave through environmental interaction. In this study, a deep reinforcement learning approach with a modified version of the deterministic policy gradient (DDPG) is proposed to implement an interactive driving control for AV. Under the developed co-simulation framework, the deep controller to achieve interactive driving safety of AV can be generated automatically with the aid of model-free deep reinforcement learning.

Keywords: Fuzzy Systems and their applications, Computational Intelligence, Soft Computing, Socio-Economic Cybernetics
Abstract: Recently, many studies on Werewolf Game have been reported in the field of games with incomplete information. In this paper, we confirm the effectiveness of the SIRMs fuzzy inference model, which has small number of rules than the simplified inference model, and the deep SIRMs fuzzy inference model, by applying them to Werewolf Game.

Keywords: Intelligent Transportation Systems, Autonomous Vehicle, Decision Support Systems
Abstract: The depth information provided by LiDAR 3D point clouds provides help for object detection when the depth direction is occluded. However, large amounts of data from lidar point clouds requires unusual computing resources. Inspired by images and 3D reconstruction, we propose to use normal vectors of the triangular mesh as a new texture feature of point clouds. Our current work indicates that the texture features we created have a certain contribution in the classification task of single-view point clouds. Key word: Classification, Single-view, reconstruction, Normal vectors.

Keywords: Cybernetics for Informatics, Computational Life Science, AIoT
Abstract: The detection of driver fatigue is a longstanding issue in the field of biometric measurements. In this study, we examined the possibility of detecting the fatigue state of drivers using their body accelerations. Nine healthy subjects operated a driving simulator for one hour in the morning and one hour in the evening. A Holter ECG was worn and body acceleration was measured using a built-in triaxial accelerometer. The results of the comparison of body acceleration in the morning and evening showed that the variance of the data was larger in the afternoon for most subjects. By visualizing and comparing the time series within the same subject, this study showed that the variance of the driver's body acceleration increases with fatigue. It is expected that this method will be applied to improve driver safety in the future.

Keywords: Medical Informatics, Assistive Technology, Human-Machine Interaction
Abstract: The global market for surgical robots is expected to experience rapid growth in the next decade, primarily attributed to the increasing demand for healthcare services driven by population aging and improved living standards. Although robotic surgery offers many advantages compared to traditional methods, lack of public awareness of these benefits poses challenges to its widespread adoption, necessitating an understanding of factors that influence patients’ choice of robotic surgery. Many studies in the literature examine the facilitators and barriers to adopting robotic surgery, focusing on surgical techniques and surgeons’ perspectives. However, limited research explores patients’ perceptions and reactions, mostly emphasizing personal characteristics. Drawing on the unified theory of acceptance and use of technology (UTAUT) model, this study investigates two situational determinants and three personal determinants of patients’ robotic surgery adoption. Analyses of Japanese data reveal that patients’ interest in cutting-edge tech, trust in AI, and surgery hospital access improve their attitude toward robotic surgery and, thereby, stimulate their robotic surgery adoption. The availability of a trusted surgeon influences robotic surgery adoption only when patients are interested in cutting-edge technology. In addition, the need for social connections directly impedes the choice of robotic surgery. These findings extend scholars’ understanding of the UTAUT model to the robotic surgery context, and they guide both surgical robotics firms in market segmentation and hospitals in decisions on the adoption of robotic surgery equipment.

Keywords: Media Computing
Abstract: This paper proposes graphical lasso-guided principal component analysis (GLIPCA) with interrupted time series analysis (ITSA), that is, GLIPCA-ITSA. This enables us to reveal trend clusters influenced by a social event (intervention) from Japanese tweets related to coronavirus disease (COVID-19). We regard daily changes in the frequencies of each word in tweets as trends and define trends with similar wave forms as trend clusters. First, we need to identify trends influenced by the intervention. ITSA, which can quantify the effect of the intervention even if its counterfactual does not exist, is useful for this aim. However, ITSA cannot exclude trends influenced by an event on a day close to the intervention. To overcome this difficulty, we newly adopt GLIPCA that combines sparse structural learning with network clustering. We can divide multiple trends, which may have been influenced by the intervention, into trend clusters with different peaks. This helps us judge the trend cluster strongly influenced by the intervention. In the experiment, we verify the effectiveness of GLIPCA-ITSA through considerations on the state-of-emergency declaration.

Keywords: Human Performance Modeling, Wearable Computing, Medical Informatics
Abstract: Digital biomarkers (dB) provide valuable information for the continuous assessment of disease status in clinical practice. Hemodynamics is an important endpoint for evaluating a patient's status and, therefore, a continuous monitoring system using dB needs to be developed. In this study, we developed supervised learning modeling approaches for estimating hemodynamic scenarios of new patients using a time series dataset obtained from contact and contact-free sensors.

Keywords: Human Performance Modeling, Human Factors
Abstract: In United States healthcare, there are persistent problems associated with infections, especially in relation to hand hygiene in anesthesia. While the sterile field of an operating room is well defined, cross contamination can still occur due to the complexity of the anesthesia work environment. This study aims to address this deficiency using a novel computational technique. Formal methods will be used to verify the cleanliness and properties associated with the anesthesia induction process as they dynamically evolve. Our hypothesis is that formal verification will be able to predict and identify how infection can spread in the operating room during the anesthesia induction process. This paper describes preliminary work demonstrating the feasibility of the modeling underlying this approach.

Keywords: Human-Machine Interface, Human-Computer Interaction, Human-Machine Interaction
Abstract: Dysarthria is a kind of speech impairment. As well-known, the people with dysarthria cannot speak clearly or his speech intelligibility is poor; so that cause them to seldom communicate with others. In addition, since each dysarthria patient has a different degree of language nerve damage, it is not easy to collect enough voice files to train ASR (Automatic Speech Recognition) for certain dysarthria patient, although the deep learning-based ASR system is very mature. It is not easy to construct a good ASR system for these people with dysarthria. Therefore, the speech recognition accuracy of this kind of ASR is not very good. How to compose meaningful Chinese sentences in this situation has become an important issue. In this paper, based on our previous work, a N-gram based Chinese sentence composition method is proposed to compose a meaningful Chinese sentence for the people with dysarthria. To address this issue, we introduce a N-Gram repository that include a lot of N-words based term, and a HSWD (History Successive Word Dictionary) to record the follow-up term for the certain term. While an ASR model recognize the voice of a sentence and generate a N-gram candidate matrix based on the N-gram repository, the system will compose possible candidate sentence according to the HSWD and recognized N-gram candidate matrix. In the work, we propose a framework and algorithm for inferring and compositing the possible candidate sentence. The experimental results show that the proposed approach can achieves over 80 % accuracy rate of composed sentence in ten candidate sentences.

Keywords: Machine Vision, Image Processing and Pattern Recognition, Biometric Systems and Bioinformatics
Abstract: Abstract—Intracorporeal suturing is one of the most crucial skills in the Fundamentals of Laparoscopic Surgery. Surgical residents are evaluated by their supervisory surgeon, but surgical assessment demands a substantial amount of time from the surgeons and can result in biased evaluations. In addition, distinguishing sealed tissue among multiple trainees would be a subjective decision. Therefore, we propose an autonomous assessment support system, which can supervise the execution of the suturing task by using YOLOv8 instance segmentation and object detection, tracking the tip of the suturing instruments and monitoring tissue seals and knots. We used mean average precision and inference time metrics to evaluate the performance of the instance segmentation for our proposed suturing assessment system. It was found that the precision of all suturing instruments was 95% and that the mask precision of the tissue was 98.8%. Our proposed autonomous laparoscopic training system saves the supervisor surgeons' time, and the outcomes of the proposed methodology may also be utilized in the development of surgical robots.

Keywords: Digital Twin, Modeling of Autonomous Systems
Abstract: This paper introduces a novel method, called robust nonlinear state space modeling (RNSSM), for predicting hemodynamic responses in fluid resuscitation. The RNSSM approach integrates autoencoder learning and Gaussian inference in a unified framework to address the challenges associated with identifying reliable models with limited and noisy critical care data. Simulation results demonstrate the initial feasibility and performance evidence of the RNSSM approach, which serves as a digital twin of an animal study, in fluid resuscitation scenarios.

Keywords: Virtual/Augmented/Mixed Reality, Assistive Technology, Augmented Cognition
Abstract: In an emergency setting, patient status can change in a matter of seconds, and failure to react in a timely manner can pose significant risks to patient outcomes. Nurses are responsible for processing large volumes of complex information in a limited time frame, which can result in high cognitive load (CL). With the advent of new technology, tools such as augmented reality (AR) can be utilized as an aid to nurses in decision-making (DM). Previous interviews indicated experienced nurses thought AR could be beneficial in nursing education. Based on this feedback, the team constructed a virtual checklist to leverage clinical judgment for novice nurses. This pilot study aims to determine whether using the Microsoft HoloLens, an AR device, to display a virtual checklist will reduce CL in novice nurses.

Keywords: Biometrics and Applications,, Medical Informatics
Abstract: Preterm birth (PTB), the leading cause of infant death, is mainly predicted by cervical length (CL). But using CL alone, even medical experts have difficulty predicting PTB. Therefore, we aimed to identify new predictors of PTB except CL using eXplainable Artificial Intelligence (XAI). Moreover, we conducted a comparative study to investigate the impact of the model's learning pattern and features according to the differences in expert guidance and XAI methods on the same ultrasound images. As a result, we figured out that these influences may be a hurdle to finding new prediction factors for preterm birth.

Keywords: Robotic Systems, Autonomous Vehicle, Consumer and Industrial Applications
Abstract: Unmanned aerial vehicles (UAV) are nowadays an important piece of equipment in search and rescue missions depending to their flexibility and convenience. Computer vision algorithms such as machine or deep learning are often used in such tasks. However, such approaches rely on many aerial view data to train the machines, but most of the existing publicly available aerial datasets belong to land or traffic scenes, and a few maritime aerial datasets focus on object detection, tiny object detection, and so on. Therefore, this paper introduces a model for marine pollution detection. We use a video of a stranded cargo ship to extract 142 images as a dataset and use the PIDNet[1] as semantic segmentation model for training and validation to demonstrate the applicability to maritime oil pollution detection. Source code is released in https://github.com/HengWeiBin/Oil-Polution-Dataset-with-PID Net.

Keywords: Autonomous Vehicle, Robotic Systems, Infrastructure Systems and Services
Abstract: This paper contributes to the development of mechanical modeling and motion control of autonomous mobile robots (AMRs) using artificial fuzzy neural networks. An interval type-2 fuzzy neural network (IT2FNN) algorithm is incorporated with the mechanical modeling of mobile robots to develop intelligent control schemes. The metaheuristic genetic algorithm is employed to determine the initial IT2FNN structure. Taking the mechanical modeling, genetic algorithm and IT2FNN control strategy, the two-wheeled nonholonomic mobile robots are steered to accomplish time-varying trajectory tracking tasks. Numerical simulations were conducted to illustrate the efficiency and superiority of the proposed intelligent IT2FNN control method for autonomous mobile robots.

Keywords: Manufacturing Automation and Systems, Robotic Systems, Intelligent Transportation Systems
Abstract: This study proposes an intelligent control strategy for automatic guided vehicle (AGV) trajectory tracking control for application in manufacturing facilities Based on the fuzzy-neural network (FNN) method combined with the newly proposed error algorithm in the design of the Loss function, virtual models with the working environment and the AGV platform can be used to learn controller's parameters in advance from the simulation environment. Finally, the real-time trajectory tracking task of navigation is evaluated to verify the effectiveness of the proposed control strategy integrated with a reflector-assisted localization algorithm.

Keywords: Affective Computing, Human-Collaborative Robotics, Human-Machine Interface
Abstract: Companion robots become familiar year by year. This manuscript describes our collaboration project related to the design and implementation of virtual and real kawaii companion robots by affective evaluation using EEG and ECG by Japanese and American university students. Each group consisted by the combination of Japanese and American students designed and implemented several robots which cause different affective reactions estimated by EEG and ECG.

Keywords: Human-Collaborative Robotics, Human-Machine Interaction, Human-Machine Cooperation and Systems
Abstract: Robot-assisted surgery has revolutionized the healthcare industry by providing surgeons with greater precision, reducing invasiveness, and improving patient outcomes. However, the success of these surgeries depends heavily on the robotic system's ability to accurately interpret the intentions of the surgical trainee or even surgeons. One critical factor impacting intent recognition is the cognitive workload experienced during the procedure. In our recent research project, we are building an intelligent adaptive system to monitor cognitive workload and improve learning outcomes in robot-assisted surgery. The project will focus on achieving a semantic understanding of surgeon intents and monitoring their mental state through an intelligent multi-modal assistive framework. This system will utilize brain activity, heart rate, muscle activity, and eye tracking to enhance intent recognition, even in mentally demanding situations. By improving the robotic system's ability to interpret the surgeon's intentions, we can further enhance the benefits of robot-assisted surgery and improve surgery outcomes.

Keywords: Intelligent Transportation Systems, System Modeling and Control, Cooperative Systems and Control
Abstract: To improve the operational efficiency of civil aviation, advanced technology in the interval management of aircraft needs to be developed. In this manuscript, the design of a fuel-efficient interval management algorithm is addressed: ensuring proper interval of arriving aircraft with altitude control. The effectiveness of the presented algorithm is verified in a simulation with arrival aircraft at Kansai International Airport.

Keywords: Intelligent Power Grid, Electric Vehicles and Electric Vehicle Supply Equipment, Cooperative Systems and Control
Abstract: The rapid increase in penetration of electric vehicles demands the widespread installation of fast charging stations. These stations require a very high level of electrical power, drastically changing the electrical load profile by increasing its peak. This results in increased system losses and voltage drops throughout the network and limits the number of electric vehicles that can charge at the same time. This paper presents a reinforcement learning-based optimization of vehicle charging location. This novel approach uses optimal distribution network reconfiguration to train an electric vehicle charging coordinator, implemented as a reinforcement learning agent.

Keywords: Human-Machine Cooperation and Systems, Human-Machine Interaction, Systems Safety and Security
Abstract: As the number of highly automated and autonomous vehicles on public roads continues to rise, gaining trust becomes a crucial aspect for their acceptance in society. Particularly challenging situations arise when these vehicles drive at low speeds, as they frequently encounter interactions with vulnerable road users. Consequently, addressing these interactions becomes essential for the vehicle's automation. This abstract has two key contributions: the introduction of an approximate vulnerable road user model and the comparison of the model-based and model-free decision-making algorithms. The simulation results showed the effectiveness and practicality of the proposed model-based algorithm even for real-world applications due to the traceability of this decision-making algorithm.

Keywords: Fault Monitoring and Diagnosis, System Modeling and Control, System Architecture
Abstract: UAV(Unmanned Aerial Vehicle), which has advantage of flexibility, cost-effectiveness, easy to deploy and risk avoidance, is emerging as a communication relays for building network infrastructure not only in military sector but also in the civilian sector. However, these networks lack countermeasures for situations in which network disconnection occurs after construction due to UAV’s own defects, such as variables that occur in nature, shooting down from enemy force in battlefield environments, operation shutdown and malfunction. Therefore, in this paper, recovery UAV is additionally deployed in a multi-UAV based FANET(Flying Ad-Hoc Network) environment composed of UAV-BS(UAV-Base Station). As a result, when a defect in the UAV-BS occurs, the recovery UAV moves to the corresponding coordinate. And Restores communications on behalf of the UAV-BS mission.

Keywords: Human-Machine Interaction, Virtual/Augmented/Mixed Reality, Human-Centered Transportation
Abstract: It is important that passengers of autonomous personal mobility vehicles (PMVs) do not feel anxious while moving. Previous works have shown that presenting information about the PMV’s future behavior and intentions effectively reduces anxiety. Mixed reality (MR) devices are promising tools for realizing this information presentation. However, current devices’ limitations may influence the visual information’s effect. Therefore, the effect of information presentation using an MR device on reducing the anxiety of autonomous PMVs was investigated using a PMV simulator. Investigation results showed that the anxiety of autonomous PMV passengers could be reduced by visual information presented through an MR device.

Keywords: Cyber-physical systems
Abstract: Cyber-physical systems (CPSs) have been deployed rapidly and widely for Industry 4.0. This work develops an AI-based autonomous diagnostics and maintenance system for CPSs to reduce operations & maintenance costs through detecting artefacts such as faults and equipment degradation, eliminating unnecessary downtime, and performing non complicated maintenance. This system is made possible by recent rapid technological advancements in robotics and computing to perform dynamic sensing and remote manipulation to reduce repetitive work and maintenance tasks, especially in harsh environments dangerous or hostile to humans. This autonomous system aims to: (1) uses mobile robots with portable sensors to dynamically sense in-the-field data, (2) uses machine learning (ML) and AI to analyze the integrated dynamic sensor data with sensor data from sensors installed in CPSs, and (3) accomplish ancillary maintenance tasks via robots. The current work progress includes a power plant simulation for robotic navigation and manipulation simulations, trajectory optimization and task planning algorithms development, and ML-based diagnostics system. This work also discusses the research challenges which need to be solved to achieve the proposed autonomous diagnosis and maintenance system.

Keywords: AI and Applications, Evolutionary Computation, Transfer Learning
Abstract: This paper presents a method of deploying a complex machine learning model on a mobile device which has limited computing resources. With the rapid advancement of machine learning and deep learning, Natural Language Processing (NLP) has made considerable progress in recent years. The transformer approach has been widely used in NLP tasks with the ability of parallel processing in recurrent neural networks through the use of a self-attention mechanism. However, the excessive pursuit of accuracy leads to higher complexity in a model, and it becomes difficult to deploy such a model on mobile devices. To solve the problem of high computational cost, this research uses Neural Architecture Search (NAS) with Genetic Algorithm as the search strategy. It trains the supernet for performance evaluation to automatically design efficient architecture for different hardware platforms while considering network compression. We further reduce the size of the model with little impact on accuracy by applying K-means clustering. To show our method’s effectiveness, we have deployed a model on a mobile device to perform machine translation. We use BLEU, FLOPs, and Latency as evaluation metrics. BLEU is used as an evaluation of sequence generative tasks while FLOPs and latency are used to observe whether the architecture found by the NAS is suitable for the hardware.

Keywords: AI and Applications, Computational Intelligence, Application of Artificial Intelligence
Abstract: The neuromusculoskeletal modeling and simulator (NMMS) have been widely utilized in various fields and applications. The deep reinforcement learning (DRL) algorithm is a promising method to study human motor controls and movement biomechanics via NMMS without experimental data. However, existing research lacks exploration of the DRL implementation for controlling neuromusculoskeletal simulators, and only a few have presented myoelectric control systems applied to the DRL-based NMMS. In this work, an off-policy DRL algorithm, Deep Deterministic Policy Gradient (DDPG), was implemented on an upper limb NMMS with two different types of action space – direct muscle activation output and PD-based internal model, and compared their control performance. In addition, we evaluated the performance of proportional myoelectric control systems implemented on the DRL-based upper limb NMMS. The results indicate that the DRL-based NMMS can execute upper limb movements accurately, and the proportional myoelectric control system reduced the muscle activation under both types of action space. Moreover, the PD-based internal model action space shows better learning and error-tracking performance than the direct muscle activation output action space.

Keywords: AI and Applications, Image Processing and Pattern Recognition
Abstract: Ultrasound kidney image segmentation presents significant challenges due to missing or ambiguous boundaries. In this study, we introduce a coarse-to-refinement approach incorporating four novel aspects. Firstly, we leverage the properties of a principal curve (PC) to automatically fine-tune the curve shape and employ a neural network's learning ability to reduce model error. Secondly, a deep fusion learning network is utilized for the coarse segmentation step, incorporating a parallel architecture to enhance deep-learning performance. Thirdly, addressing the limitation of standard PC-based methods in determining the number of vertices automatically, we propose an automatic searching polygon tracking method using a mean shift clustering-based approach to replace the projection and vertex extension step in standard PC-based methods. Lastly, we develop an explainable mathematical map function for the kidney contour, as denoted by the neural network output (i.e., optimized vertices), which aligns well with the ground truth contour. We conducted various experiments to evaluate our method's performance, demonstrating its effectiveness in ultrasound kidney image segmentation.

Keywords: Hybrid Models of Computational Intelligence, Metaheuristic Algorithms, Swarm Intelligence
Abstract: We aim to develop a simple motion capture system for home environments. As users need to install their own cameras, a calibration-free system is required. Therefore, we propose a 3D pose estimation method based on 3D joint angles of humans using multiple smart devices with a hybrid optimization method that combines Particle Swarm Optimization and steepest descent method. We also estimate the relative angles between humans and cameras to facilitate camera calibration. In this paper, we discuss the impact of the combination of global and local search capabilities of the optimization method on the system's performance. Specifically, we propose an optimization method that gradually changes the number of iterations of Particle Swarm Optimization and Steepest Descent Method and compare it with a simple sequential combination.

Keywords: Application of Artificial Intelligence, Machine Learning, Biometric Systems and Bioinformatics
Abstract: sRPE and ACWR are valuable tools for controlling fatigue levels and minimizing injuries in performance sports. Their ability to provide individualized assessment, integrate subjective and objective measures, and inform data-driven decision-making makes them essential components of a comprehensive sports safety and performance monitoring system. The goal of this study was to provide first a computer-assisted solution to predict the fatigue level and then to expand the currently existing manual calculation models with an artificial intelligence-supported solution for a more advanced pipeline in performance sports to control fatigue and minimize the injury level.

Keywords: Expert and Knowledge-Based Systems, Hybrid Models of Computational Intelligence, AI and Applications
Abstract: Theories, methodologies, practices, and software for problem solving by model definition, analysis and simulation, and communication were essentially developed in engineering during the past decades. This paper first gives an analysis of the above main advances which are tied to paradigm shifts and essential for PhD and other engineering related research programs. Following this, scenario of a proposed new style of research is introduced. This research is relied upon methods and software tools offered by high end engineering modeling platform. Another main contribution in this paper is full research in a purposefully configured autonomous and reactive research centered experimental model (RCEM) which is continuously developed during the lifecycle of research. RCEM collects and organizes model and simulation representations which accommodate results of the research. Research process is also controlled by contexts from industrial and other application environments. Essential representations in RCEM are behaviors to achieve active knowledge and intelligent computing intensive solution. New concentration area is model representation for medical problem related solutions in this paper. Virtual Research Laboratory (VRL) was founded to realize globally participated collaborative space organized high level PhD research on the cloud utilizing world level engineering modeling platform. Communication with other collaborative spaces and outside world sources and controls is also a main concern. Implementation and initial application of the results in this paper is done at the VRL.

Keywords: Deep Learning, Machine Learning, Knowledge Acquisition
Abstract: The aim of this study is to model the cognitive processes based on a perceiving-acting cycle in patients with unilateral spatial neglect (USN). USN is the inability to perceive features of the environment, body, or objects on one side. To extract the cognitive characteristics of USN patients in a multifaceted manner, we constructed a multimodal sensing system using immersive VR. In this paper, we present a system that predicts　movement of a subject while performing a search task using the measurement results and an LSTM neural network.

Keywords: Autonomous Vehicle, Intelligent Transportation Systems
Abstract: The increasing volume of traffic worldwide enlightens the problem of ensuring driving safety and preventing collisions at unsignalized intersections. In this regard, with the advent of Connected Autonomous Vehicles (CAV), Collision Avoidance (CA) and Autonomous Intersection Management (AIM) problems have been intensively studied and many cooperative and optimization-based approaches have been proposed. In this paper, we introduce a double-level collision-free control system, performed by a Central Optimizer (CO) and local CAV controllers, that allows CAVs to safely cross the intersection at the same time. The CO collects data from CAVs and imposes waiting times at specific waypoints that are then used by the low-level controllers to regulate the vehicle speed. The main advantage of the presented method is that a less complex optimization problem is formulated by imposing waiting times rather than determining the speed profile. A simulation campaign conducted on Matlab shows the performance of the proposed approach.

Keywords: Discrete Event Systems
Abstract: This work addresses the security problem of protecting secrets in discrete-event systems modeled by deter- ministic finite automata. We first characterize a global secret that composes of one or several states, in which each state is associated with a security level. In addition, we assume that the protected event labels must be recovered within a bounded of consecutive protected events (called as K-protection). The aim of this work is to design an K-protection event policy such that there exists event sequence from the initial state that reaches a secret state piece contains a number of protected events no less than the required level of security, such that the protected secret pieces satisfy protection setting value. To this end, a security automaton that integrates the information of the system state and its current security level is given, and then a K-protection automaton that lists all the possible protections of event sequences is established. Then by using the supervisor control theory technique, the valid protecting policy to enforce the security requirement is obtained. Finally, examples are used to illustrate the proposed protection method.

Keywords: Adaptive Systems, Autonomous Vehicle, Decision Support Systems
Abstract: Highly automated or autonomous systems utilize machine learning-based approaches to perceive the environment and make decisions based on the obtained information. These approaches are highly depended on the model used, training data, and environmental conditions. While high performance can be expected in certain and suitable situations, unknown or uncertain situations might lead to undetected underperformance. In case of object detection, the quantification of reliability of a particular prediction is usually given by a detection score predicted by the trained model. While a higher score indicates higher confidence, it does not reflect the actual uncertainty of the prediction considering situational variations. In this paper, the predicted detection score is converted into a true-positive rate to decide whether to accept or reject a prediction. Furthermore, the detection rate is used as a function over distance to determine the dynamic performance expectations using situational knowledge and reliability requirements.

Keywords: Cooperative Systems and Control, Autonomous Vehicle, Control of Uncertain Systems
Abstract: In this paper, we propose a reinforcement learning-based safe control strategy for uncertain heterogeneous multi-agent systems. The objective is to achieve collision-free time-varying formations under switching topologies and with limited network resources. Without requiring global communication information, an event-triggered observer is designed to decouple the heterogeneous dynamics from switching networks and reduce data transmission frequency. A data-driven off-policy reinforcement learning algorithm is developed for addressing the robust safe formation control problem. The algorithm is capable of solving non-quadratic optimization problems without requiring model information. The proposed strategy is applied to a multi-unmanned ground vehicle (UGV) system for a patrolling mission, and the experimental results verified the effectiveness of the proposed strategy.

Keywords: Autonomous Vehicle, Modeling of Autonomous Systems, System Modeling and Control
Abstract: Vehicle Dynamics Models (VDMs) face a trade-off scenario between accuracy and speed. More complex models can generate more accurate predictions of vehicle state but are more computationally slow. Additionally, many VDMs rely on the explicit estimation of unknown parameters. To avoid these limitations, we propose a feed-forward Time Delay Neural Network (TDNN) which surpasses the physics-based VDMs in accuracy and speed without explicit estimation of unknown quantities. Notably, the TDNN model was able to accurately predict the vehicle state on various road surfaces despite no knowledge of the tyre-road friction. The TDNN predicts the longitudinal and yaw accelerations of the vehicle with a Root Mean Square Error (RMSE) of as low as 0.0387 rad/s².

Keywords: Robotic Systems, Mechatronics, System Modeling and Control
Abstract: This paper presents an intuitive human-robot interface system (iHRIS), where the deep deterministic policy gradient (DDPG) with hindsight experience replay (HER) training method is proposed to accelerate model training. The system consists of a motion capture system and a motion learning network. The motion capture system includes an RGB-D camera and an operating glove. The position of the human operator's hand is estimated by Openpose using the RGB-D images, while the hand's posture is determined by the information captured by the glove. An inertial measurement unit (IMU) and a microprocessor are equipped on the glove. The IMU data is calibrated using the Recursive Least Squares (RLS) method and computed using Madgwick's algorithm. Based on the observed position and posture of the human operator's hand, a motion can be generated. This motion is then trained using the DDPG network with the Reverse Curriculum Generation (RCG) method. The network has an Actor-Critic structure and a replay experience buffer, which makes it more feasible and helps avoid overfitting. Furthermore, HER is integrated into the network to enhance convergence speed and performance. Finally, the experiments demonstrate that the proposed iHRIS enables real-time imitation of the human operator by the robot, and the imitated motion can be learned by the DDPG network.

Keywords: Evolutionary Computation, Computational Intelligence
Abstract: We propose a method that combines multi-dimensional scaling (MDS) with a regression model to approximate fitness landscapes and enhance the efficiency of evolutionary search algorithms. The approach involves projecting the original population onto a low-dimensional space using MDS, thereby preserving pairwise distances between data points. The optimal individual is determined using the approximated fitness landscape in the low-dimensional space and subsequently mapped back to the original space using our designed rules. Furthermore, several prediction points are generated in the proximity of the optimal individual to replace certain individuals. This iterative process is repeated multiple times to accelerate the search speed of evolutionary computation (EC). Experimental results demonstrate the superiority of our proposed method over the other two algorithms across most tested functions.

Keywords: AI and Applications, Agent-Based Modeling, Deep Learning
Abstract: A multi-agent decision network based on QMIX is proposed in this paper to cope with the coordination decision problem of multiple UAV air combat missions. To speed up the training process, three improvements are introduced: 1) An improved epsilon-decaying method that enable some tutor to help in action selection at the early stage of the training. This measure greatly improves the exploring efficiency when the network are far from being fully trained; 2) State pruning and action mask measures are applied during the training. The former improves the effectiveness of the input state information, and the latter reduces unnecessary action exploring. 3) A gradually training configuration is used to make the training process more robust, where the combat adversaries are configured as the static targets, the randomly maneuver vehicles, and the Min-Max strategy vehicles respectively. The multi-UAV air combat scenarios are built up and the experiments are conducted. The results shows that these improvements have significantly improved training efficiency.

Keywords: Deep Learning, Machine Vision, Transfer Learning
Abstract: With the rise of AI, the recognition of Sign Language (SL) through sign-to-text has gained significance in the field of computer vision and deep machine learning. However, there are only a few medium to large open datasets available for this task, as it requires a vast dataset of thousands of signs for words/phrases in different environments, which is a time-consuming and tedious process. Furthermore, there has been very little effort towards Arabic Sign Language Recognition (ArSLR). This research paper presents the results of fine-tuning the Vision Transformer (ViT) model on a small-scale in-house dataset of ArSL. The main goal is to attain satisfactory results by utilizing minimal computing power and a small dataset involving less than 10 individuals, with only one recording made for each sign in every environment. The dataset comprises 49 classes/signs, all of which were made with two hands and belong to the Level I category in terms of popularity. To enhance the dataset, three types of augmentations - translation, shear, and rotation were employed. The ViT model, pre-trained on the Kinetics dataset, was trained on the variation of augmented datasets with 2 to 10 times samples for each original video, where the training set includes solely augmented videos of 8 volunteers and the test set includes all original videos of one particular volunteer. Experimental results reveal that the combination of rotation and shear outperformed the others, achieving an accuracy of 85% on the 10 times augmented dataset. This study sheds light on small-scale dataset-based SLR tasks and video/action recognition in general.

Keywords: Cognitive Computing, Human-centered Learning
Abstract: The field of artificial intelligence research has witnessed remarkable advancements in recent decades. However, conventional approaches in AI research primarily depend on fixed datasets and stationary settings, which have limited applicability to real-world scenarios. In order to address this limitation, there is an increasing need to develop and study algorithms and methods for continual learning, which enables artificial systems to learn from a continuous stream of data. One of the key challenges in continual learning is to strike a balance between transfer and interference, and to identify an equilibrium solution that can effectively learn from nonstationary data. This paper presents a canonical model that is specifically designed for continual learning, utilizing the derivative of the loss function to evaluate parameter changes between tasks and achieve dynamic equilibrium. Additionally, to improve the efficiency of limited training samples in continual tasks, a feature learning method based on meta-feature disentangling is proposed. By leveraging the second derivative term of the canonical model, the parameter vector can be decoupled and meta-features can be discovered. Experimental results demonstrate the superiority of the proposed method over state-of-the-art methods in continual lifelong supervised learning benchmarks. The validity of the proposed canonical model is further supported by these experimental results. As the demands of available settings become increasingly stringent, the advantages of disentangling meta-features become more prominent, resulting in a significant performance gap with other continual learning methods.

Keywords: Medical Informatics, Human-centered Learning, Assistive Technology
Abstract: This study aims to explore the potential of a reconstructive approach in detecting and analyzing hemiplegia in human walking motion using a musculoskeletal model. Hemiplegia, a neurological disorder characterized by partial paralysis of one side of the body, is a common complication of stroke that can significantly impact its victim’s mobility and overall quality of life. We employed a combination of surface electromyography (sEMG) and motion capture data collection techniques to study healthy individuals’ muscle activation and kinematics both in normal conditions and while wearing a hemiplegia simulator. The collected data were analyzed using a non-negative matrix factorization (NMF) algorithm to decompose the muscle activation into distinct muscle synergies and associated weights. We then compared the number of muscle synergies used between the healthy and simulated hemiplegia conditions to identify changes in muscle coordination patterns. Using the information obtained from the comparison, an attempt was made to construct the synergies of the hemiplegia gait from the synergies of the healthy gait by adjusting the weights. The results of this study suggest that a reconstructive approach is a feasible tool for detecting and analyzing hemiplegia in a human walking motion. Furthermore, this study provides a framework for understanding the underlying muscle coordination patterns in hemiplegia and can serve as a basis for developing new rehabilitation strategies. This study is expected to contribute to the advancement of research in this field by providing insights into the neural control of human walking motion.

Keywords: Multi-User Interaction, Human-centered Learning
Abstract: Data trading has been hindered by privacy concerns associated with user-owned data and the infinite reproducibility of data, making it challenging for data owners to retain exclusive rights over their data once it has been disclosed. Traditional data pricing models relied on uniform pricing or subscription-based models. However, with the development of Privacy-Preserving Computing techniques, the market can now protect the privacy and complete transactions using progressively disclosed information, which creates a technical foundation for generating greater social welfare through data usage. In this study, we propose a novel approach to modeling multi-round data trading with progressively disclosed information using a matchmaking-based Markov Decision Process (MDP) and introduce a Social Welfare-optimized Data Pricing Mechanism (SWDPM) to find optimal pricing strategies. To the best of our knowledge, this is the first study to model multi-round data trading with progressively disclosed information. Numerical experiments demonstrate that the SWDPM can increase social welfare 3 times by up to 54% in trading feasibility, 43% in trading efficiency, and 25% in trading fairness by encouraging better matching of demand and price negotiation among traders.

Keywords: Machine Learning, Neural Networks and their Applications, Deep Learning
Abstract: Being able to predict the remaining useful life (RUL) of an engineering system is an important task in prognostics and health management. Recently, data-driven approaches to RUL predictions are becoming prevalent over model-based approaches since no underlying physical knowledge of the engineering system is required. Yet, this just replaces required expertise of the underlying physics with machine learning (ML) expertise, which is often also not available. Automated machine learning (AutoML) promises to build end-to-end ML pipelines automatically enabling domain experts without ML expertise to create their own models. This paper introduces AutoRUL, an AutoML-driven end-to-end approach for automatic RUL predictions. AutoRUL combines fine-tuned standard regression methods to an ensemble with high predictive power. By evaluating the proposed method on eight real-world and synthetic datasets against state-of-the-art hand-crafted models, we show that AutoML provides a viable alternative to hand-crafted data-driven RUL predictions. Consequently, creating RUL predictions can be made more accessible for domain experts using AutoML by eliminating ML expertise from data-driven model construction.

Keywords: Human-Centered Transportation, User Interface Design, Human-Machine Cooperation and Systems
Abstract: Cooperation between driver and vehicle automation can be utilized to overcome failures and problems of the automated system. Failures do not always indicate mandatory intervention scenarios but can be within the optional decision space and have a negative impact on user experience. In this paper, we evaluate one user interface (UI) based on a joystick and one based on a touchscreen that enable an increased degree of cooperation compared to state of the art UIs in automated vehicles. Both UIs allow users to adjust parameters and initiate as well as abort maneuvers of the vehicle motion within the optional space of action. In a driving simulator experiment, participants experienced the new concepts in comparison to a state of the art UI. Participants rated both newly developed concepts positively regarding usability and user experience. Overall, they preferred the novel concepts compared to the baseline interface. An objective evaluation of interaction issues and visual distraction revealed flaws that can be resolved by adapting the designs. Participants preferred an increased level of driver vehicle cooperation (DVC) and indicated a high intention to make use of it for optional interventions. The indicated usage intention decreased with increasing automation level.

Keywords: Robotic Systems, Autonomous Vehicle, Intelligent Transportation Systems
Abstract: With the development of technology, the automation industry has gradually improved, and the application of the visual-inertial system has become increasingly prosperous. Self-driving cars and autonomous mobile robots often use the visual-inertial system as a state estimator, and their image information can be used in the artificial intelligence part to provide more information about the robot. The combination of a monocular camera and an inertial sensor has the characteristics of light weight and low power. Monocular Visual-Inertial SLAM is a system that relies on monocular cameras and inertial sensors to perform localization and mapping. Monocular Visual-Inertial SLAM restores its vector to scale with the help of known extrinsic parameter information. However, compared with other methods, it is necessary to be careful in estimating the gyroscope bias of the inertial sensor. The bias and noise of the inertial sensor will directly affect the results of pre-integration, as well as the accuracy of the visual-inertial scale recovery. This phenomenon is called scale uncertainty, which results in feature points with uncertain scale or depth, rapidly reducing the accuracy of pose estimation. This paper proposes an architecture of visual-inertial based on Structure-Scale Adaptivity. In this study, the tightly coupled residual between the structure scale and the pose is designed to provide adaptivity according to the feature conditions, providing system structure and scale constraints, and thus suppressing the problem of scale uncertainty. This study is compared with state-of-the-art algorithms and succeeds in scenarios where other algorithms have failed. It also obtains the best accuracy results in the overall test. Finally, we evaluate our proposed method on VCU-RVI and KITTI datasets, demonstrating the proposed algorithm's effectiveness and feasibility for monocular visual-inertial SLAM.

Keywords: Kansei (sense/emotion) Engineering, Virtual/Augmented/Mixed Reality, Entertainment Engineering
Abstract: In this study, a two-step dropping model was proposed to replicate a long-distance fall in a virtual reality (VR) space, while realizing only a short drop distance in reality. In a previous study, a two-step dropping model representing a falling motion was proposed and presented to humans in real space. This concept replicates the sensation of a long-distance fall by presenting the dropping motion to the wearer only at the start and landing of the fall in the VR space. A two-step dropping device was fabricated using an air cylinder. A two-step dropping motion was realized by combining the open/close control of the conduit with a direct-acting solenoid valve and pressure control with a proportional solenoid valve. In the sensitivity evaluation experiment, participants were presented with a combination of the device's operation and VR images. The evaluation scores for the sense of reality of falling in the two-step dropping condition tended to be higher than those in the other conditions, compared to the condition in which the device was not operated or only one-step dropping was presented. This confirmed that the two-step drop was an effective operation for experiencing a free-fall sensation.

Keywords: System Modeling and Control, Cooperative Systems and Control, Fault Monitoring and Diagnosis
Abstract: This paper presents a new framework for robust relative position and velocity estimation of satellite swarms in Low Earth Orbit (LEO) in the presence of statistical outliers in sensor measurements. The proposed framework is based on the Robust Generalized Maximum Likelihood Kalman Filter (RGMKF) and uses a minimal sensor configuration and inter-satellite communication. The satellites are assumed to be equipped with GPS receivers that obtain signals from high-altitude navigation satellites, as well as relative positioning sensors based on radar, cameras, or lasers. Measurements from the latter set of sensors will aid in keeping reasonable accuracy when GPS is unavailable or unreliable and improve estimation accuracy when GPS is available and reliable. The paper highlights that practical scenarios may present scaling challenges when applying RGMKF, which could be effectively addressed by properly reordering data in the measurement matrix. The proposed framework and the outlier detection scheme are evaluated through simulations demonstrating their effectiveness in addressing challenges in state estimation of satellite swarms with sensor outliers.

Keywords: Machine Learning, Metaheuristic Algorithms, Evolutionary Computation
Abstract: Methods have been proposed to maximize the efficiency of Automated Machine Learning (AutoML) systems while simplifying the search for solutions. Multi-fidelity approaches have been shown as an alternative to achieve this since they are straightforward to implement and reduce the computational cost when modeling large datasets. However, they are not suited to shifting data distributions, and they remove configurations too fast. Therefore, they must be combined with other techniques. Combining multi-fidelity methods with evolutionary algorithms and island models helps to adapt better and can contribute to maintaining diversity. This paper presents a comparative analysis of 10 network topologies distributed over a generalized island model for AutoML. A dynamic migration model with multi-objective and multi-fidelity evaluation is proposed to reduce the complexity of the tasks. This proposal is compared against state-of-the-art AutoML frameworks. It was found that Hypercube, Grid 2-dim, and Grid 3-dim topologies have the best performance as they maintain a balance in the number of connections. Furthermore, these topologies were shown to be competitive against other frameworks in the state of the art.

Keywords: Machine Vision, Deep Learning, Representation Learning
Abstract: Class-incremental learning (CIL) tackles the problem of continuously optimizing a classification model to support growing number of classes, where the data of novel classes arrive in streams. Recent works propose to use representative exemplars of learnt classes, and replay the knowledge of them afterward under certain memory constraints. However, training on a fixed set of exemplars with an imbalanced proportion to the new data leads to strong biases in the trained models. In this paper, we propose an attentive knowledge replay framework to refresh the knowledge of previously learnt classes during incremental learning, which generates virtual training samples by blending between pairs of data. Particularly, we design an attention module that learns to predict the adaptive blending weights in accordance with their relative importance to the overall objective, where the importance is derived from the change of the image features over incremental phases. Our strategy of attentive knowledge replay encourages the model to learn smoother decision boundaries and thus improves its generalization beyond memorizing the exemplars. We validate our design in a standard class-incremental learning setup and demonstrate its flexibility in various settings.

Keywords: Knowledge Acquisition, Expert and Knowledge-Based Systems
Abstract: This paper investigated how prompt-based learning techniques can assist with relation fusion in Knowledge Graph (KG) acquisition. We created a unsupervised framework to generate a KG from a real-world dataset. The framework incorporates prompting with knowledge entity metadata and generating predicate embeddings with the pretrained Masked Language Model (MLM) RoBERTa. Predicate embeddings were clustered to form conceptual groups and feature tokens were used to derive relation labels. In addition, we conducted a comparative study on the effects of different prompting templates. The resulting relation labels were evaluated by human annotators, which indicated that prompt-based learning, if applied appropriately, can help with deducing conceptualized relations. Our framework proposed a way to improve the quality of KGs acquired using traditional Relation Extraction (RE). It can also assist human experts effectively in semi-automated knowledge acquisition.

Keywords: Quantum Cybernetics, Expert and Knowledge-Based Systems, Metaheuristic Algorithms
Abstract: With the growing popularity of online trading platforms, investors can consider investing in the global market without regard to their geographical location. This study analyzes the investment performance of Group of Seven (G7) members, who greatly influence the global economy and stock market. Regardless of the market in which they invest, investors seek long-term and stable gains in the stock market; therefore, investing in a risk diversification option is paramount. By constructing a portfolio, investment risk can be decreased. The relationship between return and risk is crucial. In this study, the proposed intelligent investing system with a novel trend ratio indicator aids in constructing an efficient portfolio. In a volatile stock market, the daily ups and downs are all profit opportunities; therefore, the proposed intelligent system constructs long-selling and short-selling portfolios. This study expands the investment discussion market to the G7 market and combines trading strategies into an innovative hybrid trading strategy to evaluate the G7 market's overall performance. As the trading strategies and markets under consideration are more inclusive, the proposed quantum-inspired algorithm with the entanglement technique can address the vast portfolio construction solution space. This study generalizes trend ratios in long-selling and short-selling portfolios and G7 markets. The experiment's result reveals that the hybrid strategy can increase portfolio profitability and efficiency.

Keywords: AI and Applications, Image Processing and Pattern Recognition, Deep Learning
Abstract: Artificial intelligence (AI) has transformed medical diagnosis and improved quality of life. But in the field of veterinary medicine has been limited due to training data and obtaining high-quality data. In this study, we propose a framework for diagnosing left atrial enlargement in dogs using AI techniques. Our framework involves generating X-ray image data and utilizing the UNet model for segmentation. The results of our experiments show excellent performance, with a mean dice score of 0.9186 for segmentation. The highest classification accuracy was achieved in trial 1 for normal and overall cases, with 0.9200 and 0.8478, respectively, while trial 2 had the highest abnormal heart classification accuracy of 0.8095. Our findings indicate that generating data and training the model with a certain percentage of the generated data can lead to high classification accuracy. We conclude that the proposed framework has the potential for clinical application in veterinary medicine.

Keywords: AI and Applications, Deep Learning, Representation Learning
Abstract: In typical search engines, users are exposed to an interface that combines free-form textual fields and hard filters. After the initial query, users implicitly express their preferences through clicks, purchases, or connection requests. In this paper, we propose leveraging implicit user feedback to learn similarity metrics between search results that receive attention together. The learned similarity model can notably be used to serve recommendations. It captures a form of latent similarity that is not known to the search engine but learned from implicit user feedback.

We introduce a method to create a dataset suited for similarity learning from search data. Using this dataset, we use contrastive learning and train a similarity model that outputs large scores for pairs of results that tend to receive the same attention from users and low scores otherwise.

Our experiments show that BERT models and their siamese equivalents (Sentence BERT) produce meaningful similarity metrics when fine-tuned on the dataset built from search data.

Keywords: AI and Applications, Deep Learning, Evolutionary Computation
Abstract: Text-to-image generation using deep learning based models has become a popular research topic, allowing users to generate custom artworks from specified text input. However, generating suitable prompts that produce creative and desirable images remains a significant challenge. To address this challenge, we propose a novel method that incorporates interactive evolutionary computation (IEC) to evolve the latent array. By integrating human perception into the system, our approach enables users to search for and generate images that align with their desired specifications through interaction with the system. We demonstrate the effectiveness of our proposed method in generating images that align with the users' mental images from an initial image using genetic algorithms through a series of experiments. Furthermore, the results from our user studies show that our proposed method enables users to generate images that match their desired mental images with less effort and in less time compared to conventional generation methods. Overall, this study contributes to the field of text-to-image generation by introducing a human-in-the-loop approach that enhances user control and specificity in the image generation process.

Keywords: Electric Vehicles and Electric Vehicle Supply Equipment, System Modeling and Control, Intelligent Transportation Systems
Abstract: This paper proposes a backlash control algorithm using Terminal Iterative Learning Control (TILC) in the angle domain. The proposed method addresses the issue of control time variation in the Iterative Learning Control (ILC) method, which makes it impractical for vehicle control. By controlling backlash in the angle domain, the control interval remains the same for each iteration. The backlash impact is proportional to the velocity at the end of the backlash mode. However, by bringing the reference value nearer to zero, the impact was mitigated. Additionally, the utilization of TILC enhances its resilience to sensor noise. The proposed method is evaluated through simulations and experimental results, demonstrating its practical applicability to vehicles and high accuracy in various initial conditions. This paper provides a novel approach to backlash control in-vehicle systems, contributing to the advancement of control methods for improved ride comfort and safety.

Keywords: Autonomous Vehicle, Intelligent Transportation Systems
Abstract: We present a method, which incorporates knowledge awareness into the symbolic computation of discrete controllers for reactive cyber physical systems, to improve decision making about the unknown operating environment under uncertain/incomplete inputs. Assuming an abstract model of the system and the environment, we translate the knowledge awareness of the operating context into linear temporal logic formulas and incorporate them into the system specifications to synthesize a controller. The knowledge base is built upon an ontology model of the environment objects and behavioural rules, which includes also symbolic models of partial input features. The resulting symbolic controller support smoother, early reactions, which improves the security of the system over existing approaches based on incremental symbolic perception. A motion planning case study for an autonomous vehicle has been implemented to validate the approach, and presented results show significant improvements with respect to safety of state-of-the-art symbolic controllers for reactive systems.

Keywords: Image Processing and Pattern Recognition, Neural Networks and their Applications, AI and Applications
Abstract: Automated brain tumor classification is an intensively investigated problem, which recently attracted significant attention. Convolutional neural networks (CNN) and deep learning represent the standard for the foundation of any recent solution. This paper proposes two simplified VGG architectures and investigates their capabilities and limitations, in comparison with state-of-the-art CNN networks deployed via transfer learning. Various parameter settings are involved in the evaluation process, including different kernel sizes, dropout rules, loss functions, etc. Networks are trained and tested on a public brain tumor classification data set consisting of 3064 images and three tumor classes (meningioma, glioma and pituitary tumor). The thorough evaluation process revealed that the proposed CNN models can achieve competitive performances with regard to state-of-the-art methods in several scenarios. The best achieved accuracy benchmarks are 98.2% overall Dice similarity score and correct decision rate, and AUC values over 99.6% for each of the three tumor classes.

Keywords: System Modeling and Control, Control of Uncertain Systems, Adaptive Systems
Abstract: In this paper, a novel control architecture is presented by integrating an Output Recurrent Fuzzy Neural Long Short-Term Memory (ORFNLSTM) and a Broad Learning System (BLS) for a class of single-input-single-output (SISO) nonlinear dynamic systems. This new controller, abbreviated as the ORFNLSTM-BLS controller, is especially proposed by combining the techniques of deep learning and broad learning method to establish an adaptive intelligent controller with an online deepest gradient descent learning algorithm to online update its weights. The ORFNLSTM-BLS controller aims to improve the performance of the ORFBLS controller by incorporating the memory of LSTM to handle time-series data more effectively. A sufficient condition of the proposed controller is established to accomplish its uniformly asymptotical stability. The effectiveness and superiority of the ORFNLSTM-BLS controller are well exemplified by carrying out one comparative simulation in comparison with a fixed-gain proportional-integral-derivative (PID) controller, an adaptive predictive PID controller augmented with ORFBLS (ORFBLS-APPID), and an existing ORFBLS controller in terms of two types of control performance indexes: the overall performance indexes and transient state performance indexes. The results show that the proposed ORFNLSTM-BLS controller outperforms the three existing control methods. The developed techniques would provide useful references for professionals working in the fields of process and servomechanism control.

Keywords: Intelligent Transportation Systems
Abstract: Different applications for advanced driver assistance systems (ADAS) have been increasing rapidly. With the advances of machine learning techniques, traffic risk assessment is becoming possible by understanding the complex driving scenes. This paper presents a technique to classify possible risks in four levels using deep neural networks. The important regions in road scene images are first predicted, followed by the risk assessment using a series of weighted RGB and optical flow images obtained from the proposed network model. To achieve better understanding of driving scenes, driver visual attention is incorporated to enhance the feature extraction. An LSTM model is then employed to learn the temporal information from road scene video clips. In addition to the experiments using the public TRA dataset, we also present a new LTDR dataset for performance evaluation. Compared with existing techniques, the result has demonstrated the effectiveness of the proposed method.

Keywords: Communications, Smart Sensor Networks, Technology Assessment
Abstract: Abstract—Based on IoT network security considerations, the BT mesh specification defines twice encryption for each data packets. One of the encryptions is in Network layer, it uses NetKey to separately encrypt DST (destination address) and Lower Transport PDU with AES encryption, and obfuscates SEQ

(sequence number), SRC (source address), and TTL (time to live). However, the security mechanism may increase the packet process overhead of all relay nodes in mesh networks, and reduce packet forwarding efficiency. This paper will design a BT mesh routing, tag switching and adopt CRC-24 chained computing mechanisms concurrently to avoid the packet process overhead caused by the security mechanism and improve the forwarding

efficiency of data packets at all relay nodes

Keywords: Cyber-physical systems, Digital Twin, System Modeling and Control
Abstract: Laser remelting (LRM) is one of very few universal technologies (i.e., no material removal and no material addition) used in a wide range of manufacturing applications, spanning from surface polishing to functional structuring. Given that LRM is a nonlinear, non-stationary thermodynamic process, its analysis, monitoring, control, and optimization require a comprehensive and fundamental understanding of multiple interlinked laser-material interaction phenomena. Such an understanding is derived from on-line information gained from measurements made during the process by various sensors. Toward this end, this study presents a concept and recent achievement in the development of a statistical digital twin and cyber-thermophysical system towards their use to stabilize, control, and optimize the process. In particular, the digital twin statistically describes the transformation of the initial surface into the LRM topography in terms of thermodynamic transfer functions of remelting and resolidification of surface topography and bulk material. In addition, a cyber-thermophysical system was developed interconnecting the thermodynamic heat-transfer model with the laser beam position, thermal-emission distribution, and overall temperature of the laser-material interaction zone measured and synchronized on-line in space-time position coordinates. The efficiency of the developed cyber-thermophysical system was demonstrated by analyzing the surface formation and process thermodynamics during LRM of H13 tool steel. The preliminary results open new research directions in thermophysics-supported advanced analysis, monitoring, and optimization of LRM, including the implementation of artificial-intelligence methods.

Keywords: Computational Intelligence, Neural Networks and their Applications, Hybrid Models of Computational Intelligence
Abstract: In order to realize a super-smart society (Society 5.0) where humans and robots coexist, there is a need for a perceptual system that can recognize the environment quickly and flexibly in an environment that changes from moment to moment. In an unknown environment, the characteristics of objects cannot be known in advance, and thus prior learning-based recognition methods such as deep reinforcement learning may not be able to cope with this situation. In this study, we construct a 3D topological map of the environment in real-time using Growing Neural Gas (GNG), which can learn 3D topological structures even for unlearned objects. However conventional GNG have the problem that they cannot generate nodes with high-density for distant objects and cannot identify whether an unknown object is static or dynamic. Therefore, by directly adding useful input data as a new node (reference vector) based on the object category labels of the winner nodes (nearest nodes) to the input vector (3D point cloud), it is possible to generate high-density topological structures even for distant objects. We proposed the Add-if-Silent rule-based GNG with Amount of Movement (AiS-GNG-AM), which can identify between static and dynamic objects based on the past amount of movement of a node. The effectiveness of the proposed method is verified through experiments using a 3D dynamics simulator.

Keywords: Decision Support Systems, Discrete Event Systems, Large-Scale System of Systems
Abstract: This paper presents an improved Delaunay-Triangulation-based Variable Neighborhood Search (DVNS) using a new solution data structure, Satellite List (SL), to solve large-scale colored traveling salesman problem instances. SL encoding can reduce the time complexity of three basic operators in DVNS, i.e., insertion, swap, and flip to O(1), dramatically promoting the iteration efficiency of DVNS. Extensive experiments are conducted to validate the improved DVNS by comparing the presented method with the state-of-the-art algorithm DVNS. The results show that the improved DVNS outperforms the original in algorithm convergence, iteration efficiency, and solution quality.

Keywords: Brain-Computer Interfaces, Systems Safety and Security, Human-Machine Interaction
Abstract: Machine learning models are known to be vulnerable to adversarial perturbations in the input domain, causing incorrect predictions. Inspired by this phenomenon, we explore the feasibility of manipulating EEG-based Motor Imagery (MI) Brain Computer Interfaces (BCIs) via perturbations in sensory stimuli. Similar to adversarial examples, these adversarial stimuli aim to exploit the limitations of the integrated brain-sensor-processing components of the BCI system in handling shifts in participants' response to changes in sensory stimuli. In this paper, we first define adversarial stimuli and enumerate the characteristics of such stimuli. Second, we study the feasibility of adversarial stimuli as an attack vector in a series of human subject experiments, and report the findings on the impact of visual adversarial stimuli in the form of random partial flickers on the MI task of playing a video game. Our findings suggest that adversarial stimuli can significantly deteriorate the performance of MI BCIs across all participants, and also that such attacks are more effective in conditions with induced stress. Furthermore, we provide a preliminary analysis on the underlying dynamics of adversarial stimuli attacks by investigating the variations in Alpha, Beta, and Gamma bands. Lastly, we present a discussion on the impact of our findings, and enumerate directions of further research in this area.

Keywords: BMI Emerging Applications, Active BMIs, Passive BMIs
Abstract: —Brain-computer interfaces (BCIs) are positioned to help individuals with physical disabilities, yet training data for these systems are time-consuming and expensive to collect. BCIs could thus benefit from cross-user data sharing, but as neurophysiological signals used in BCIs are personal health information, the protection of user privacy and data sovereignty will be of the utmost importance for end-user BCI applications. Federated learning is a novel privacy-preserving machine learning technique that decentralizes training, leaving end-user data on the device/institution of origin. In this work, we explored the feasibility of federated learning for BCIs, profiling convergence and performance for two federated learning algorithms (FedAvg and FedDC) on both identically and independently distributed (IID) and non-IID partitions of data. Federated learning for a 4-class motor imagery BCI decoding task was found to be feasible, although came at a cost of reduced performance longer convergence rate and reduced accuracy). The FedDC algorithm, which introduces drift correction for heterogeneous data, well out-performed the FedAvg algorithm. Larger datasets with more subjects and data per subject would be beneficial for further investigations, and future work should explore federated transfer learning techniques for combating inter-subject data heterogeneity and improving global model performance.

Keywords: BMI Emerging Applications, Other Neurotechnology and Brain-Related Topics, Passive BMIs
Abstract: Brain-Computer Interface (BCI) and Internet of Things (IoT) systems have recently been amalgamated to create BCIoT. Most of the early applications have focused on the healthcare sector, and more recently, in education, virtual reality, smart homes, and smart vehicles, amongst others. While there are many transversal developing stages that can be satisfied by a single system, no common enabling technology or standards exist. These challenges are address in the proposed platform, Brain-eNet. This technology was developed considering the constraints-space defined by BCIoT real-time mobile applications. This is expected to enable the development of BCIoT systems by providing modular hardware and software resources. Two instances of this platform implementation are provided, a motor intent detection for rehabilitation and an emotion recognition system.

Keywords: Telepresence, Human-Machine Interaction, Human-Machine Cooperation and Systems
Abstract: Telepresence is the technology that allows human beings to remotely monitor, supervise, and control a wide variety of tasks. These tasks range from medical examinations to space discovery. Teleoperation, as the underlying foundation of telepresence, saves money and lives by replacing and relocating human elements from dangerous or harmful environments such as factories’ production lines, mines, ports, logistic centres, and so on. Teleoperation also finds its place in autonomous systems, including training an autonomous mobile robot via imitation learning (learning from demonstrations) and human-in-the-loop performance evaluation. On the other hand, this technology can improve life quality by offering healthcare services to remote and rural areas, where access to clinical experts is dangerously limited.

However, teleoperation frameworks generally suffer several threats and challenges, including time delays, uncertainties, safety, and cyber-attacks. Time delay is one of the most critical factors that should be taken into account in the design and development of teleoperation systems of any kind. Transmission of command/control and sensory information signals through a communication medium between the two main sides of a teleoperation platform, being the operator and teleoperator, is critically vulnerable to latency and uncertainties. Delays in those transmissions will negatively affect the stability and performance of the teleoperation task. Moreover, designing a safe and secure teleoperation framework is another challenge considering the topology of the network on which the teleoperation is being conducted. In the case of remotely commanding an automated vehicle, it is down to the teleoperation strategy to determine the safety and feasibility of the desired path and manoeuvring before driving it; or in clinical applications and telehealth applications, cyber-attacks can maliciously steal and manipulate patients’ data or jeopardize the clinical teleoperation (remote surgery, medical imaging, rehabilitation, etc). In all these cases, a robust and reliable teleoperation strategy is a must to guarantee the stable, safe, and effective performance of the desired tasks in a remote environment.

Keywords: Human-Machine Interaction, User Interface Design, Telepresence
Abstract: The world population exceeded 8 billion by 2020. On the other hand, there are concerns about shortage of labor population in in many developed countries due to the declining population with declining birthrate and an aging population. AVATAR technology is expected to solve the social problem of uneven population distribution by enabling everyone to live and work where and with whom they want to live. There are a lot of functional Avatar in the world. Because of the same body structure, humanoid avatars are suitable for replacing human tasks, especially when they are remotely controlled by wearing HMDs to perform a variety of tasks. However, humanoid robots are currently too complex and expensive and wearing HMDs may disturb the spread of tele-operation services.

We at LAST MILE aim to realize a tele-operation service that realize anyone to live and work where they want and with whom they want. In order to popularize the service, it is important to develop flexible design technology that enables low cost and interface technology that can be operated by mobile smart devices. In this paper, we introduce two main strategies developed to make this service begin true. i. The first is the proposed Augmented Avatar technology, which can be flexibly designed for each service by separating the "mind" and the "body" of the avatar. By a selection of modularly designed communication avatars and manipulation avatars in accordance with the service, the combined avatar which we named “Augmented Avatar” can communicate and manipulate with people in the surrounding area at a distance. ii. The second is a mobile smart device interface technology which operate our new prototyped Augmented Avatar. It makes an operator communicate and manipulate remotely only by a smart phone.

Keywords: Telepresence, Haptic Systems, Human-Machine Interaction

Keywords: Human Performance Modeling
Abstract: Decoding Part of Speech(POS) tagging directly from electroencephalography (EEG) signals whilst the user overtly spoke (voiced speech) sentences could improve direct speech brain-computer interfaces (BCIs) using imagined or inner speech. To our knowledge, earlier work uses a machine learning approach using 74,953 sentences/tokens recorded in 75 EEG sessions. The tokens can be found in 4,479 phrases consisting of terms from the English Online treebank, which contains the record of weblogs, newsgroups, reviews, and Yahoo Answers. The results demonstrated the feasibility of POS decoding from EEG based on word class, word frequency, and word length with an accuracy of 71%,86%, and 89%, respectively. We believe there is significant room for improvement with more advanced artificial intelligence. In this paper, we further extend the existing work with end-to-end transformers. Our results presents transformer model outperforms benchmark traditional ML results with +20% in length, +13% for the open vs closed class and +12% in frequency. In our empirical analysis, we find the decoding performance was better when using multi-electrode recordings as compared to single-electrode recordings.

Keywords: Human Performance Modeling, Augmented Cognition, Information Systems for Design
Abstract: Dual tasks with extended and natural bodies can enhance and complement human body functions and improve productivity. In recent years, methods for sharing information between two points have been investigated to realize dual tasks. However, even with these methods, when the required attentional load increases due to time and environmental factors, dual tasks may not be possible when the human cognitive ability reaches its limits. Therefore, the purpose of this study is to determine the intervention method among individuals by analyzing and categorizing how the performance on a dual task decreases when the required attentional load increases through a validation test. The validation test consisted of a dual task with two different tasks: the Button task, which needed to be performed with priority, and the Tracking task, which needed to be performed lower priority but with a gradually increasing amount of attention. As a result of the validation test, we hypothesized that there exists a one-sided state in which response delays occur in the low-priority task during the transition from the dual-task state, where the task is performed stably, to the saturation state, where all tasks are disrupted. Therefore, considering the one-sided state, we classified participants' attention distribution state into four states during the dual-task and analyzed the transition of the collapse model for individual. As a result, we were able to identify the one-sided state, classify the collapse models of the dual-task into three patterns, and propose appropriate timing of intervention methods for each pattern. The appropriateness of the timing of intervention methods needs to be verified, which is a topic for future research.

Keywords: Human Performance Modeling, Team Performance and Training Systems, Resilience Engineering
Abstract: This paper aims to classify and quantify individual growth processes during pilot training to understand the variations in their growth processes. Various software reliability growth models were fitted to growth processes, and the best-fit model, which showed the minimum Akaike information criterion, was clarified for each applicant. As a result, it was shown that the individual growth processes were classified into concave and S-shaped processes. The concave and S-shaped processes were attributed to the cases where an applicant completed the training without and with wandering, respectively. Since the best-fit models included models with imperfect debugging and testing efforts these factors were discussed from the viewpoint of flight training. It was also indicated that the estimated growth models after the training can be utilized to monitor the applicant's competencies depending on operation-service flights done up to recurrent training.

Keywords: Intelligent Internet Systems, Neural Networks and their Applications, Hybrid Models of Neural Networks, Fuzzy Systems, and Evolutionary Computing
Abstract: Efficient and accurate identification of network anomalies is of great significance to the construction of network security systems in the information age. It is highly challenging to accurately detect abnormal behaviors in the increasing network data. Currently, classification methods based on feature extraction of autoencoders have been proven to be suitable for network anomaly detection. However, traditional detection models with autoencoders have poor detection accuracy in the face of massive network features. In addition, the hyperparameter optimization of their models cannot be effectively solved. For network anomaly detection, a new network anomaly detection method named SAXP is proposed in this work. SAXP integrates Stacked sparse shrink Autoencoders and an unbalanced XGBoost model based on genetic simulated annealing Particle swarm optimization (GSPSO). Specifically, features extracted by stacked sparse shrink autoencoders are introduced into the XGBoost model based on improved unbalance parameters for classification, and GSPSO is used to optimize the hyperparameters of XGBoost. Experimental results based on two real-life data sets demonstrate that the proposed SAXP achieves higher recognition accuracy than several state-of-the-art algorithms.

Keywords: Intelligent Internet Systems, Cloud, IoT, and Robotics Integration, Evolutionary Computation
Abstract: Mobile edge computing (MEC), as a promising paradigm, delivers computation and storage capacities at the edge of the network. It supports delay-sensitive services for mobile users (MUs). However, dynamic and stochastic characteristics of MEC networks necessitate constant migration of installed services across edge servers to keep up with the mobility of MUs. As a result, the cost of maintaining the network increases significantly. Existing studies of MEC rarely consider the cost of service migration due to MU mobility. To minimize the long-term cost for microservices in a hybrid cloud-edge system comprising of MUs, small base stations (SBSs), and a cloud data center (CDC), the total cost minimization is formulated as a constrained mixed-integer nonlinear program. To solve it, this work designs a novel meta-heuristic optimization algorithm called Multi-swarm Gray-wolf-optimizer based on Genetic-learning (MGG), which effectively combines strong local search capabilities of gray wolf optimizer with superior global search capabilities of genetic algorithm. MGG simultaneously optimizes service request routing among MUs, SBSs, and CDC, CPU speeds of SBSs, service deployment of SBSs, service migration cost of SBSs, as well as MUs’ transmission power and channel bandwidth allocation. Simulation results with Google cluster trace demonstrate that MGG outperforms several state-of-the-art peers with respect to the overall cost of the hybrid system.

Keywords: Knowledge Acquisition, Machine Learning
Abstract: Active semantic mapping is essential for service robots to quickly capture both the map of an environment and its spatial meaning, while also minimizing the burden on users during robot operation and data collection. SpCoSLAM, a method of semantic mapping with place categorization and simultaneous localization and mapping (SLAM), is well suited to environmental adaptation, as it is not limited to predefined labels. However, SpCoSLAM presents two issues that increase the burden on users: 1) users struggle to efficiently determine a destination for the robot's quick adaptation, and 2) providing instructions to the robot becomes repetitive and cumbersome. To address these challenges, we propose Active-SpCoSLAM, which enables the robot to actively explore uncharted areas and employs CLIP for image captioning to provide a flexible vocabulary that replaces human instructions. The robot determines its actions by calculating information gain integrated from both semantics and SLAM uncertainties. We conducted experiments in a simulated environment, comparing the proposed method to other methods in terms of efficiency and applicability to object discovery tasks. Additionally, we tested the proposed method, which combines user instruction and CLIP, in a real environment. Our results demonstrated that the robot explored its environment with approximately five fewer iterations and 11 minutes faster compared to the case of random exploration. Moreover, our method achieved a higher success rate in object discovery tasks during earlier stages of learning compared to other methods. In conclusion, the proposed method rapidly covers an environment while gathering useful data for object discovery tasks, thus reducing the burden on users and enhancing the robot's adaptability. The project website is https://tomochika-ishikawa.github.io/Active-SpCoSLAM/.

Keywords: Knowledge Acquisition, Machine Learning
Abstract: In recent years, the graph feature mining method of brain connection data based on graph theory has been regarded as a popular and universal technology in the field of neuroscience. How to mine valuable information from brain connection data has become a research hotspot. Current research shows that the pathogenic factors of attention deficit and hyperactivity disorder (ADHD) may be caused by the abnormal connection between brain network structures. In order to find out the pathogenic factors of ADHD patients, we also carried out frequent subgraph mining on the connectivity graph data of brain functional network. By constantly adjusting the support threshold, all the subgraphs of ADHD patients and healthy control group were mined, and the differences in brain region connectivity were successfully found out. By combining the recently introduced neural document embedding model with traditional pattern mining techniques, we regard the brain network connection structure graph as the document and frequent subgraph as the atomic unit of the embedding process. By learning the mapping, each graph can be mapped to a D-dimensional continuous vector. The mapping needs to capture the similarity between the graphs. Feature vectors can be used as the direct input of graph classification in many traditional machine learning methods. Finally, support vector machine in machine learning is used to verify the accuracy of classification, and the results show that the accuracy is high.

Keywords: AI and Applications, Application of Artificial Intelligence, Deep Learning
Abstract: In today’s era of automation, mobile robots are being used for collecting meaningful information about an ambient phenomenon such as temperature or moisture distribution in an agricultural field. Most of the studies in the literature assume that the underlying information field is Gaussian, and therefore, Gaussian Process (GP)-based models are extremely popular. Furthermore, we have found that due to the inherent computational complexity of such naive GP-based techniques, most studies in the literature do not scale well beyond small-size environments, i.e., where the number of informative points n < 1000. These render such a predictive model more or less useless in many practical applications. In this paper, we posit that a different technique, Generative Adversarial Network-based inpainting, for robotic information gathering can be useful. The state-of-art inpainting techniques 1) do not assume that the underlying data is Gaussian, and 2) easily scale to n ≫ 1000. Thus, they eliminate the two bottlenecks posed by the GP-based solutions. We have tested our hypothesis on a synthetic and a real-world crop dataset. Results show that while the inpainting technique easily scales to 1024 × 1024, GP-based predictions cannot. On the other hand, their solution qualities are shown to be comparable.

Keywords: AI and Applications, Intelligent Internet Systems
Abstract: Selecting an optimal recommendation method is crucial for an electronic commerce (EC) site. However, the effectiveness of recommendation methods cannot be known in advance. Although continuous comparative evaluation in an actual environment is essential, it results in opportunity loss. To overcome this problem, opportunity loss reduction has been studied as a multi-armed bandit (MAB) problem, and adaptive meta-recommender systems (meta-RS) were devised to automatically and continuously select the best recommendation method according to a policy. The following three factors cause opportunity loss: the context of the recommendation method, temporal variation, and response time. However, studies are yet to formulate an MAB policy that considers all three factors. Thus, reducing opportunity loss remains a problem. We propose an MAB policy that considers all three causes of opportunity loss by using a Kalman filter for a linear Gaussian state space model. We conducted extensive experiments to select the best recommendation method using data from a real EC site. The results revealed that the proposed policy is highly effective for reducing opportunity loss of the meta-RS during evaluation and increasing cumulative clicks compared with baseline policies.

Keywords: AIoT, Machine Learning, Deep Learning
Abstract: "To enable domain adaptation of AI on edge devices with fast convergence and low memory, we present a novel backpropagation-free optimization algorithm dubbed Target Projection Stochastic Gradient Descent (tpSGD). tpSGD uses layer-wise stochastic gradient descent (SGD) and local targets generated via random projections of the labels to train the network layer-by-layer with only forward passes. It doesn’t require retaining gradients during optimization, thus greatly reducing memory allocation compared to SGD backpropagation (BP) methods. Compared to other target propagation methods, tpSGD generalizes the concept to arbitrary local layer-wise loss functions. Our method performs comparably to BP gradient-descent within ∼5% accuracy on relatively shallow networks of fully connected layers, convolutional layers, transformers, and recurrent layers. tpSGD also outperforms other state-of-the-art gradient-free algorithms with competitive accuracy and less memory and compute time."

Keywords: Human-Machine Interaction, Human-Machine Cooperation and Systems, Information Visualization
Abstract: As the drone captures image targets at different flying altitudes, their scales may vary significantly, which can pose challenges for the object detection model to accurately detect them.Additionally, tiny objects in the image contain minimal information, making them difficult to distinguish from the background.To overcome these two challenges, we proposed a network architecture that aims to improve the accuracy of tiny object detection in drone images. Specially, we designed a tiny object detector(TOD) that can effectively extract features of tiny objects and distinguish between tiny object features and image background. Furthermore, this TOD module contains a Convolutional Visual Attention Network (CVAN) to better focus on the regions of tiny objects. Experimental results demonstrate that the proposed method achieves mAP@.5 accuracy of 53.9% on the VisDrone2021-test-dev dataset and improves by 2.8% compared to YOLOv7.

Keywords: Human-Machine Interaction, Human-Machine Cooperation and Systems, Human Factors
Abstract: The simulation of the human body dynamics offers a powerful way to investigate the interaction between humans and systems. This so-called physical human-machine interaction (HMI) is useful to consider during the development of new products like tools, cars or aircrafts, in order to increase comfort and safety. In this work, the Modelica Human Body Library is presented first. It is the only digital human mode (DHM) for physical HMI available in the multi-domain modeling language Modelica so far. The model considers the human body dynamics as well as the motor control of the limbs. A special feature of the model is the human inspired control scheme, which takes into account effects such as human reaction time and limb stiffness adaption. After an introduction to the DHM, its capabilities are evaluated by comparing simulation results with experimental data from a experimental study. In the study, different unstable force fields were applied to the hands of subjects as they approached a target. The resulting hand trajectories were measured. As many control tasks for humans, especially while using tools, are characterized by unstable behavior, this study is well suited to evaluate a DHM. Comparison with the experimental data shows that the resulting hand trajectories in the simulation are comparable to the ones the subjects produced due to the force fields. This applies for both, the initial tries with big deviations as well as the tries after the subjects adapted their arm stiffness to the force field.

Keywords: Human-Machine Interaction, Human-Computer Interaction, Human-Machine Cooperation and Systems
Abstract: With the rapidly growing use of Multi-Agent Systems (MASs), which can exponentially increase the system complexity, the problem of planning a mission for MASs became more intricate. In some MASs, human operators are still involved in various decision-making processes, including manual mission planning, which can be an ineffective approach for any non-trivial problem.

Mission planning and re-planning can be represented as a combinatorial optimization problem. Computing a solution to these types of problems is notoriously difficult and not scalable, posing a challenge even to cutting-edge solvers. As time is usually considered an essential resource in MASs, automated solvers have a limited time to provide a solution. The downside of this approach is that it can take a substantial amount of time for the automated solver to provide a sub-optimal solution.

In this work, we are interested in the interplay between a human operator and an automated solver and whether it is more efficient to let a human or an automated solver handle the planning and re-planning problems, or if the combination of the two is a better approach. We thus propose an experimental setup to evaluate the effect of having a human operator included in the mission planning and re-planning process. Our tests are performed on a series of instances with gradually increasing complexity and involve a group of human operators and a metaheuristic solver based on a genetic algorithm. We measure the effect of the interplay on both the quality and structure of the output solutions. Our results show that the best setup is to let the operator come up with a few solutions, before letting the solver improve them.

Keywords: Human-Machine Interaction, Human-Machine Interface, Human-Machine Cooperation and Systems
Abstract: Inspired by a common behaviour pattern from the dog-human relationship, we present a behaviour model combined with a method for controlling a real mobile robot, effectively realizing physical Human-Object-Robot Interaction. Translating social behaviours from the dog-human relationship into social robots has been already found to be a viable way. Playing, a cardinal behaviour element in the dog-human relationship, turned out to be one of the highest priority factors in forming a successful and lasting relationship between dogs and humans. Accordingly, we modeled the playing behavior when a human repeatedly initiates playing with a dog by picking up and throwing an object (toy), expecting the dog to go for the object and bring it back to the person throwing it, commonly known as playing fetch. We also developed a robot control system able to execute the commands from the behaviour model, and for gathering the poses of the participants in the real physical space we have used the What-You-See-Is-What-You-Get Indoor Localization (WIL) system. The integration of these components allowed us to create a social robotics system able to offer real physical human-object-robot interaction fast enough to be enjoyable for human participants, as a tool to be used in future HRI experiments, which was our main goal. The proposed behaviour model is intended to be embedded into social robots with the purpose of applying it in e.g. therapy, child-rearing, child-robot interaction scenarios, human behaviour assessment, recreational or entertainment scenarios.

Keywords: Human Enhancements, Cognitive Computing
Abstract: The imbalance between the supply and demand of healthcare resources is a global challenge, which is particularly severe in developing countries. Governments and academic communities have made various efforts to increase healthcare supply and improve resource allocation. However, these efforts often remain passive and inflexible. Alongside these issues, the emergence of the parallel healthcare system has the potential to solve these problems by unlocking the data value. The parallel healthcare system comprises Medicine-Oriented Operating Systems (MOOS), Medicine-Oriented Scenario Engineering (MOSE), and Medicine-Oriented Large Models (MOLMs), which could collect, circulate, and empower data. In this paper, we propose that achieving equilibrium in medical resource allocation is possible through parallel healthcare systems via data empowerment. The supply-demand relationship can be balanced in parallel healthcare systems by (1) increasing the supply provided by digital and robotic doctors in MOOS, (2) identifying individual and potential demands by proactive diagnosis and treatment in MOSE, and (3) improving supply-demand matching using large models in MOLMs. To illustrate the effectiveness of this approach, we present a case study optimizing resource allocation from the perspective of facility accessibility. Results demonstrate that the parallel healthcare system could result in up to 300% improvement in accessibility.

Keywords: Assistive Technology, Human Enhancements
Abstract: This study aims to clarify the relationship between the user's characteristics and the assist effect of wearable robot for walking aid, thereby elucidating the scope of application of robot-assistive technology. In this paper, we report on the transformation of walking movements when assisting forces are applied from a powered wear. The powered wear developed in the study assists the hip joint during walking with a wire-driven system. An experiment was conducted to measure walking movements and physical characteristics. Thirteen middle–aged and elderly persons took part in the experiment. The participants were instructed to walk at their preferred walking speed, but some of them showed a slight decrease in cadence and an increase in stride length during assisted walking. Particularly, participants with average cadence tended to increase their stride length when assisted. Furthermore, the participants with higher knee extensor muscle tended to improve their propulsive force during assisted walking. The experiment result showed changes in gait specific to the assistive robot and their relationship to the user's physical ability and gait characteristics.

Keywords: Human-Machine Interaction, Human Enhancements, Human Factors
Abstract: To achieve efficient and effortless movement with exoskeletons, it is essential to optimize human-exoskeleton fluency. Current methods for measuring fluency include energetic cost, muscle activation, and spatiotemporal gait parameters, but cortical brain dynamics have not been well studied. Understanding individual cortical responses could improve the application of personalized exoskeleton assistance through human-in-the-loop (HIL) optimization approaches. Our research aimed to enhance the understanding of individual responses and identify a potential objective function based on electrocortical activity. We found that electrocortical responses show more individual variability than muscle activations, emphasizing the need for personalized cortical optimization to maximize exoskeleton assistance. We also found a significant negative relationship between changes in TA muscle activation and EEG beta power in the motor area. These findings suggest that EEG beta power changes could be a possible objective function for HIL applications.

Keywords: Deep Learning, Application of Artificial Intelligence, Biometric Systems and Bioinformatics
Abstract: Over the past several years, deep learning technologies have made tremendous progress in medical image tasks including classification, segmentation, and object detection. However, there are two main limitations of indocyanine green(ICG) images which are often used in breast cancer related lymphedema (BCRL): insufficient sample numbers and low image quality. Consequently, the conventional deep learning based classification methods such as ResNet have faced challenges in achieving satisfactory results. To tackle the concern, this paper puts forward a deep learning method named Denoising Diffusion Probabilistic Model Self-supervised Knowledge Distillation Net (DDPM-SKDNet) for the ICG images classification task, by incorporating a contrastive learning-based approach as the network architecture and using DDPM as the image generator in the contrastive module to expand the dataset size. Furthermore, a knowledge distillation approach is utilized to increase the effectiveness of the network. The proposed method was validated on ICG datasets and achieved a significant improvement in classification accuracy, increasing it from 66.7% in the baseline method to 82.1% in the proposed method.

Keywords: Image Processing and Pattern Recognition, Biometric Systems and Bioinformatics, Deep Learning
Abstract: Accurate lesion segmentation on breast ultrasound (BUS) images is a crucial procedure in computer-aided ultrasonic diagnosis. Owing to the privacy of BUS data and the complexity of acquiring pixel-level labels, numerous researches attempt to achieve breast lesion segmentation with predefined feature-based and deep learning-based methods in a unsupervised or weakly supervised scenario. Although the former can typically extract more reliable contour information of the lesion, it is severely interfered by irrelevant tissues due to its inability to capture any semantic information. Furthermore, the weakly supervised deep learning segmentation based on class activation map (CAM) can explore the semantic information while failing to provide precise contour information. In view of the above observation, we present a weakly supervised framework merging complementary contour and semantic information for early lesion segmentation in BUS images. Specifically, guided by the prior knowledge of breast anatomy, we first extract and filter the contour information of suspected lesions located in breast parenchyma layer by clustering and morphological characteristic, respectively. Afterward, semantic information extraction is performed by a classification network to automatically explore the category information of the lesion. Finally, we selectively fuse the complementary information to facilitate lesion segmentation performance with more comprehensive features. Extensive experiments are conducted on the public dataset BUSI, and the results confirm the validity of our approach.

Keywords: Deep Learning, Neural Networks and their Applications, AI and Applications
Abstract: International Classification of Diseases (ICD) coding is an internationally unified diagnostic system, which sets a unique code for each patient who carries particular disease. The primary diagnosis indicates the most crucial disease for the patient in the hospitalization, and its coding is very important for both the patient and the hospital. The current ICD encoding methods mostly use text as the input of the model, and the unstructured nature of text data results in a relatively scattered feature distribution, which is not conducive to feature extraction and interpretability research of the models. In this paper, we utilized a knowledge graph to transform text into graph structured data, and used an improved graph convolutional model to extract features from the transformed graph, achieving automatic ICD encoding for the primary diagnosis of disease. The method was tested on a Chinese dataset with macro-averaged F1 score of 0.862, and the comparative experiments depict that the performance of method based on graph convolutional networks is generally better than that ICD coding models at the text level and node level.

Keywords: Team Performance and Training Systems, Wearable Computing, Human-Machine Interface
Abstract: In basketball free-throw practice, learning the appropriate amount and timing of applying force is crucial. However, since the throwing motion is done in a short time and the applied force cannot be directly observed, it is hard for coaches to evaluate the shooter's motion and give instructions. In this study, we propose a wearable device that measures the wearer's force control during throwing based on an electrical impedance and acceleration measurement, and provides feedback to help the wearer adjust their motion. Throughout experiments, we figured out the temporal features of the measured electrical impedance and acceleration during throwing that differentiate experts and beginners. Based on the features we also proposed a training system for force control using the proposed device, and verified that is effect to improve the free-throw skill.

Keywords: Assistive Technology, Human Enhancements, Wearable Computing
Abstract: Freezing of Gait (FOG) is one of the typical parkinsonian gait disturbances of progressive neurological disorders such as Parkinson’s disease (PD) and progressive supranuclear palsy. Previous studies showed that the wearable cyborg HAL trunk-unit improved their gait disturbances through only an autonomous sway control that provided lateral swing with constant frequency. To promote the improvement by establishing interactive Bio-Feedback loop, it is necessary to realize synchronization between the wearer’s intention and gait states for stabilizing gait, i.e. reducing variation of gait cycle. In this study, we developed a voluntary control method that responded to an intentional change in gait cycle and a method that switched 2 kinds of controls including the voluntary control method and the autonomous sway control method for stabilizing gait. The voluntary swing control method synchronized the HAL’s force with wearer’s gait in a stable gait state less prone to FOG. The autonomous sway control method provided lateral swing with constant frequency as feedback to the patients to achieve a gait state less prone to FOG. These controls were switched according to the gait stability calculated by wearer’s gait cycle. Through the gait experiments with an abled-body participant, we confirmed that the HAL’s lateral cyclic sway synchronized with the participant’s gait and each control switched based on a gait stability. These results showed that the proposed methods had the feasibility of stabilizing gait.

Keywords: Wearable Computing, Human Factors, Human-Computer Interaction
Abstract: Every human being is unique and behaves differently in any given context. Standard approaches like surveys that are used today to assess human behaviour often generate subjective responses and can be administered either before or after task/activity only. This presents an opportunity to develop solutions that can monitor a human being’s cognitive, affective and mental state in their current context continuously and in real time, during activities and in an unobtrusive manner. We have developed a composite system called Mind Indriya that can unobtrusively measure cognitive load (CL), anxiety and visual attention using a combination of frugal sensors like wrist wearable and webcam . The accuracies of each of the individual algorithms have been proven with cognitive load at accuracy 69.5% , anxiety at accuracy of 86% and the detection of eye blinks as an index of visual attention at Fscore of 0.91. Moreover, our system is person and task independent while state-of-the-art techniques are either task dependent or requires heavy personalization. The proposed system can be used in various real life applications like a) neuromarketing - to understand how customer behavior towards new products or advertisement, b) online learning - to personalize contents or understand the learning outcome, c) Website/user interface design - to understand the effect of the new design on people’s cognitive load, affect and attention span and therefore how effective the design is and so on.

Keywords: Passive BMIs, Other Neurotechnology and Brain-Related Topics
Abstract: Brain machine interfaces (BMI) are becoming increasingly prevalent in diverse applications including motor rehabilitation, virtual reality trainings, etc. Two critical aspects of an effective BMI are machine reliability and cognitive workload (CWL). Previous studies have reported a notable effect of machine reliability on the 6 factors of the CWL. However, it remains unclear whether this effect can be detected in cognitive processes. Electroencephalography (EEG) is a widely used technique to explore cognitive processes by recording brain activities as signals. Therefore, we utilized the event-related spectral power (ERSP) feature of EEG signals to determine the cognitive processes regarding the effect of machine reliability. The results revealed that machine reliability affected the CWL factor of performance which was reflected in the Gamma band activities of the right prefrontal cortex. The findings indicate the potential of cognitive processes in detecting the effect of machine reliability. The detection could pave the way for designing adaptive BMI to balance between the machine reliability and the CWL.

Keywords: Passive BMIs, BMI Emerging Applications
Abstract: This paper investigates the efficacy of a passive Brain-Computer Interface (BCI) in enabling a robot tutor to adaptively respond to a user's engagement level in real-time. The BCI system extracted EEG Engagement Index from the user's electroencephalography (EEG) signals as an indicator of engagement during Human-Robot Interaction (HRI). A within-subjects study was conducted in which the robot performed attention-recapturing behavior during a learning task under two conditions; either in an adaptive manner whenever a lapse in the user's engagement level was detected by the BCI system (Adaptive condition) or at random intervals regardless of the user's mental states (Random condition). In both conditions, users completed an information retention test following the interaction. The study found no significant difference in the post-interaction test results or mean EEG Engagement Index values between the Adaptive and Random conditions. However, analysis of 10-sec time windows following robot interventions showed that adaptively timed gestures were significantly more effective in restoring user engagement to optimal level compared to randomly timed gestures. This finding provides evidence for the potential of passive BCIs in improving user experience in pedagogical HRI settings.

Keywords: Brain-Computer Interfaces, Affective Computing
Abstract: EEG channel selection can reduce data redundancy, thereby beneficial for improving the utility and efficiency of emotion recognition. Previous studies on EEG channel selection have not considered the influence of genders despite long-standing belief in gender differences with respect to emotion analysis. In this paper, we collected EEG signals from 20 subjects containing 10 males and 10 females by letting them watch short emotional videos. Then, to reduce data redundancy, we propose an enhanced genetic algorithm to select the optimal channel subsets separately for male and female subjects by incorporating a novel evolution operation. Experimental results show that the proposed algorithm achieves higher accuracy in terms of emotion recognition than several compared methods with a smaller channel subset. Besides, experimental results also indicate that the gender differences in neural patterns indeed exist. Through this study, the gender-sensitive channel selection offers a new avenue for further development of EEG based emotion recognition.

Keywords: BMI Emerging Applications, Other Neurotechnology and Brain-Related Topics, Passive BMIs
Abstract: Scalp electroencephalography (EEG) is a neural source signal that is extensively used in neuroengineering due to its non-invasive nature and ease of collection. However, a drawback to the use of EEG is the prevalence of physiological artifacts generated by eye movements and eye blinks that contaminate the brain signals. Previously, we have proposed and validated an H∞-based Adaptive Noise Cancellation (ANC) technique for the real-time identification, learning and removal of eye blinks, eye motions, amplitude drifts and recording biases from EEG simultaneously. However, the standard electrooculography (EOG) electrode configuration requires four electrodes for EOG measurement, which limits its applicability for reduced-channel mobile applications, such as brain-computer interfaces (BCI). Here, we assess multiple configurations with varying number of EOG electrodes and compare the ANC effectiveness of these configurations to the ideal four-electrode configuration. From an analysis of the root mean squared error (RMSE) and differences in signal to noise ratios (SNR) between the ideal four-electrode case and the alternative configurations, it is reported that several three-electrode alternative configurations were effective in essentially replicating the ability to remove EOG artifacts in an experimental cohort of ten healthy subjects. For nine subjects, it was shown that only two to three EOG electrodes were needed to achieve similar performance as compared to the four-electrode case. This study demonstrates that the typical four-electrode configuration for EOG recordings for adaptive noise cancellation of ocular artifacts may not be necessary; by using the proposed new EOG configurations it is possible to improve electrode allocation efficiency for EOG measurements in mobile EEG applications.

Keywords: BMI Emerging Applications, Other Neurotechnology and Brain-Related Topics
Abstract: Conventional classification approaches for EEG-based emotion recognition cannot often adapt to different domains, such as cross-subject or cross-dataset scenarios, leading to poor performance. To handle this challenge, we introduce a novel fusion method using a combination of multiple domain adaptation techniques to improve the emotional states in EEG datasets via classification accuracy. For this aim, Our proposed approach exploits domain adaptation approaches such as Transfer Component Analysis (TCA), Correlation Alignment (CORAL), Transfer Joint Matching (TJM), Geodesic Flow Kernel (GFK), and Joint Distribution Adaptation (JDA), to enhance the overall classification performance. Later, a new fusion approach called Multiple Domain Adaptation based on a Neuro-Fuzzy Inference System (MDA-NF) is applied to combine the classifiers using proper fuzzy membership functions and deliver maximum separation between classes. The main contribution is by applying the fusion approach using MDA-NF technique, adaptability is sufficiently enhanced. Another advantage is to employ multiple adaptation techniques that improve separation between classes. In experimental test results conducted with cross-subject and cross-dataset scenarios, the MDA-NF approach demonstrates superior performance in terms of accuracy for both the valence and arousal aspects, as observed in two public DEAP and DREAMER datasets.

Keywords: Conflict Resolution, Decision Support Systems
Abstract: An evacuation navigation model incorporating a fire’s detrimental effect on evacuees’ escape capability is presented. Fire is one of the most common hazards in buildings that severely impacts evacuees' escape capability. A fire dynamics model is used to simulate dynamics of a fire at various locations in a building. An index, called Capability Deterioration Rate (CDR), is introduced to model how fire effluents, such as smoke, heat, and toxic gases, impact escape capability spatiotemporally. This index is incorporated into a modified Dijkstra algorithm by changing weights of navigation edges. An agent-based model taking into account the fire dynamics and escape capability is used to simulate an agent’s evacuation from a location to minimize the total travel cost from the location to an egress door. A three-floor building is used as a case study to demonstrate the application and effectiveness of the proposed model. The proposed model can be used to evaluate and select the best navigation plan in terms of escape capability under various fire scenarios.

Keywords: Electric Vehicles and Electric Vehicle Supply Equipment, Conflict Resolution, Smart Buildings, Smart Cities and Infrastructures
Abstract: The diffusion of electric vehicles (EVs) is affected by the spread of EV public charging stations (PCSs) and vice versa. Their interactions are often referred to as ``indirect network effects'' and analyzed with a two-sided market model. However, the consumer adoption of home charging stations (HCSs), which are considered key drivers of EV deployment, is often ignored. By modeling a two-sided market of EVs and PCSs (i.e., indirect network effects) under the adoption of HCSs, this study explores the EV diffusion process and considers effective strategies for its spread. We examine two cases: in the first, consumer adoption is determined exogenously; in the second, it is determined endogenously. In each case, we vary the strength of the indirect network effects, which correspond to the drivers' concern, referred to as ``range anxiety.'' Through numerical simulations, we evaluate the market shares of EVs, HCSs, and PCSs, as well as social welfare. Our findings have strategic implications for policymakers seeking to increase the market share of EVs in the presence of different types of charging stations without negative social impacts.

Keywords: Conflict Resolution
Abstract: This paper proposes a procedure to elicit lower and upper probabilistic preferences within the Graph Model for Conflict Resolution (GMCR) through a multicriteria perspective. This imprecise probability approach is especially recommended in situations where there is not enough information regarding how DMs evaluate or compare the feasible conflict states. Within this proposal, the elicitation process is carried out through three questions in which decision makers are asked to evaluate the performance of each viable state in each criterion. Uncertainty in this performance generates lower and higher probabilistic preferences among states. To illustrate the applicability of the new elicitation procedure, we present a real-world example from a construction project in Brazil.

Keywords: Conflict Resolution, Decision Support Systems
Abstract: This paper proposes an adaptive subgroup rescue mechanism to better balance efficiency and information loss in large-scale group decision-making. We calculate the consensus for different clusters through linguistic preferences and trust relations, find the group with the lowest consensus level, and advise the group to exit from the process. A rescue process is triggered once the group has low cohesion or can propose a new idea. A trust-based feedback mechanism is designed to ensure enough decision makers in consensus reaching. Finally, we conduct an empirical study to verify the feasibility of our mechanism.

Keywords: Autonomous Vehicle, Intelligent Transportation Systems, Robotic Systems
Abstract: The gradual deployment of automated vehicles (AVs) results in mixed traffic where AVs will interact with human-driven vehicles (HDVs). Thus, social-aware motion planning and control while considering interactions with HDVs on the road is critical for AVs’ deployment and safe driving under various maneuvers. Previous research mostly focuses on the trajectory planning of AVs using Model Predictive Control or other relevant methods, while seldom considering the integrated planning and control of AVs altogether to simplify the whole pipeline architecture. Furthermore, there are very limited studies on social-aware driving that makes AVs understandable and expected by human drivers, and none when it comes to the challenging maneuver of driving through roundabouts. To fill these research gaps, this paper develops an integrated social-aware planning and control algorithm for AVs’ driving through roundabouts based on Driving Risk Field (DRF), Social Value Orientation (SVO), and Model Predictive Contouring Control (MPCC), i.e., DRF-SVO-MPCC. The proposed method is tested and verified with simulation on the open-sourced highway-env platform. Compared with the baseline method using purely Nonlinear Model Predictive Control, the DRF-SVO-MPCC can achieve better performance under various maneuvers of driving through roundabouts with and without surrounding HDVs.

Keywords: Human-Machine Interface, Assistive Technology, Medical Informatics
Abstract: Intuitive human-machine interfaces may be developed using pattern classification to estimate executed human motions from electromyogram (EMG) signals generated during muscle contraction. The continual use of EMG-based interfaces gradually alters signal characteristics owing to electrode shift and muscle fatigue, leading to a gradual decline in classification accuracy. This paper proposes a Bayesian approach for adaptive EMG pattern classification using semi-supervised sequential learning. The proposed method uses a Bayesian classification model based on Gaussian distributions to predict the motion class and estimate its confidence. Pseudo-labels are subsequently assigned to data with high-prediction confidence, and the posterior distributions of the model are sequentially updated within the framework of Bayesian updating, thereby achieving adaptive motion recognition to alterations in signal characteristics over time. Experimental results on six healthy adults demonstrated that the proposed method can suppress the degradation of classification accuracy over time and outperforms conventional methods. These findings demonstrate the validity of the proposed approach and its applicability to practical EMG-based control systems.

Keywords: System Modeling and Control, Manufacturing Automation and Systems, Cyber-physical systems
Abstract: This work deals with the verification of a decision-making system for additive manufacturing (AM) technology adoption specified by a Petri net. An innovative verification technique of a Petri net-based system is oriented toward practical applications, and can detect errors at the early design and modelling stage. The idea is illustrated by a real-life case study of supporting decision making in AM technology adoption affecting supply chain management (SCM). Two main issues are addressed. Firstly, making optimal decisions about AM technology requires models rarely possessed within a company. Therefore, this work proposes a model supporting the decision making related to the implementation of AM technology, based on a Petri net, by utilizing its main advantages: graphical modelling, and strong mathematical support of formal verification techniques. Contrary to the most popular analysis methods (which are bounded exponentially in a general case), it is proved that the presented method is bounded by a cubic polynomial with net size. Secondly, an investment in AM technology is often financially assessed and does not affect other processes, as in our SCM case. Hence, strong connectivity within the proposed Petri net-based model is examined.

Keywords: Brain-Computer Interfaces, Assistive Technology, Cooperative Work in Design
Abstract: There is a pressing need for alternative access technologies that enable children with severe physical disabilities, as current options often require some degree of controlled movement to be used efficiently. Brain-computer interfaces (BCIs) hold significant potential for improving the lives of children with severe physical disabilities, however research must prioritize user-centered approaches and real-world applications to maximize benefits. This paper examines the integration of home-based BCIs for children with quadriplegic cerebral palsy through user-centered design, focusing on the feasibility, usability, and impact on personal goals and activities of daily living. Seven children aged 6-15 years with quadriplegic cerebral palsy and their families participated in this pilot study, using personalized BCI packages and home-based virtual sessions to help them achieve individualized goals in self-care, productivity, and leisure. We utilized a collaborative goal-setting approach and assessed satisfaction and performance changes using the Canadian Occupational Performance Measure (COPM) and the BCI-adapted Quebec User Evaluation of Satisfaction with Assistive Technology (e-QUEST2.0). Significant improvements in performance and satisfaction were observed in the COPM scores, while parents were most satisfied with professional services and least satisfied with the adjustability of the BCI system, as per the eQuest2.0 questionnaire. Despite no significant improvement in BCI consistency across nine sessions, the intervention positively impacted participants' perceived performance and satisfaction in goal-oriented activities. Future research should focus on enhancing BCI design, comfort, and effectiveness while considering user priorities and feedback for personalized goal achievement.

Keywords: Medical Informatics
Abstract: Personalized patient representation learning (PPRL) is a critical element in clinical risk prediction. It aims to obtain a complete portrait of each patient based on Electronic Health Records (EHR). Although existing works have achieved remarkable progress in healthcare prediction, there are still three major issues. First, feature correlation is crucial for risk prediction, but it has not yet been fully exploited by existing works. Second, variation pattern of dynamic feature contains useful information about patient's physical status, but adaptive pattern recognition is still a challenge. Third, existing works usually adopt a two-stage embedding process to process each dimension of the EHR data. However, some useful low-level information for PPRL will be lost. To address these issues, in this paper, we propose a fine-grained PPRL architecture named FGRL-Net for clinical risk prediction based on EHR. Specifically, we propose a Medical Feature Correlation Detection Module (FCM) to effectively learn the feature correlations for each patient and a Temporal Variation Pattern Recognition Module (TVM) to effectively detect the variation patterns of each dynamic feature. Moreover, we design a Fine-Grained Representation Mechanism (FGRM) to preserve the low-level information (from both feature and visit dimensions) useful for risk prediction. In addition, in the stage of data preprocessing, We utilize generic medical classification knowledge to classify numerical dynamic data. We conduct the in-hospital mortality experiment and the decompensation experiment on a real-world dataset. The experiment results show that the FGRL-Net outperforms state-of-the-art approaches. The source code is provided in github https://github.com/JackyChio/FGRL-Net.

Keywords: Other Neurotechnology and Brain-Related Topics, BMI Emerging Applications
Abstract: A novel 3D deep convolutional neural network (3D-CNN) model called MEGNet3D has been proposed in the paper. MEGNet3D is designed to differentiate between amyotrophic lateral sclerosis (ALS) and healthy individuals from their resting state (eyes open and eye closed condition) sensor-level magnetoencephalography (MEG) data. The raw MEG data is initially transformed into their time-frequency representation which are then used as inputs to MEGNet3D. Both magnetometer and gradiometer recordings have been investigated separately. The proposed model exhibits an accuracy of over 75% for most classification conditions. Thus, MEGNet3D is capable of handling high subject variability and shows that spectral-temporal representation of resting state MEG data yields relevant neural markers related to the existence of ALS. Furthermore, it has also been observed resting state with eyes closed yields better classification accuracy as compared to resting state with eye open condition.

Keywords: Control of Uncertain Systems, Adaptive Systems
Abstract: An adaptive estimated inverse control methodology for a class of uncertain nonlinear system subjected to unknown Prandtl-Ishlinskii (PI) hysteresis nonlinearity is studied in this paper. The largest difficulty in the inverse hysteresis compensation mechanism lies in that its parameters are hard to determine. To obviate such obstruction, the unknown hysteresis parameters are estimated online by the adaptive techniques together with uncertain system parameters. Thus, hysteresis effects as well as parameter uncertainties can both be suppressed without the time-consuming off-line identification. Meanwhile, the tracking differentiator is utilized to solve "explosion of complexity" problem in backstepping control by calculating the derivatives of the virtual control laws. Based on the Lyapunov stability analysis, it has been revealed that the boundedness of all the system signals is achieved and the tracking error along with estimation errors can all converge to an compact set around zero. Contrastive simulations are performed to demonstrate the theoretical findings of this control algorithm.

Keywords: Autonomous Vehicle, Robotic Systems, Distributed Intelligent Systems
Abstract: In this paper, a novel decentralized leader-follower control scheme for multi-agent systems is devised, where each agent communicates only with a subset of neighboring mates. The goal is to track assigned trajectories for the centroid and the formation of the system. The desired trajectories are known only by a subset of agents, named leaders: the other agents, the followers, are required to estimate the desired trajectories based on a dynamic consensus scheme. Then, the desired trajectories to be tracked by each agent are computed from the estimated trajectories for the centroid and the formation and a simple local control loop is adopted to track the former. Stability and performance are analyzed and experiments are run on Robotarium platform to show the effectiveness of the approach and the effect of different parameters on the achieved performance.

Keywords: Autonomous Vehicle, Digital Twin
Abstract: Two-dimensional (2D) simultaneous localization and mapping (SLAM) using a LIDAR is a method used to track the position and orientation of a moving platform. 2D-SLAM assumes that the platform translates in a 2D plane and can only rotate about an axis perpendicular to that plane. However, the assumption of no out-of-plane (OOP) motion does not hold true for platforms experiencing motion in six degrees-of-freedom (6-DOF), such as wearable technologies that have no 3D LIDAR. This paper proposes a new algorithm, called the Level Plane for SLAM (LPS) for removing OOP motion from 2D-LIDAR scans generated on platforms experiencing 6-DOF without requiring scan-matching in 3D. Like other existing methods, an IMU is combined with a 2D-LIDAR to determine the platform’s orientation, capture OOP motion, and generate a scan in 3D. Unlike other methods, OOP motion is removed by projecting scans onto a globally stabilized coordinate frame in 2D where both scan matching and map alignment take place. The proposed algorithm is validated over a series of experiments with different levels of induced and observed OOP motion. Experimental results show that LPS is able to handle more OOP motion than other algorithms and run in real-time.

Keywords: System Modeling and Control, Smart Buildings, Smart Cities and Infrastructures, Adaptive Systems
Abstract: Adaptive high-rise buildings use multiple sensor systems, actuators integrated into the structure, and a control unit in order to actively counteract external disturbances. In civil engineering, a distinction is made between static loads such as snow, and dynamic loads, e.g. wind and earthquakes. These result in two control objectives, each of which is achieved by a separate controller. A static load compensation method is employed to minimize static displacements, while a model predictive controller induces additional damping into the structure to suppress structural vibrations. Here, both controllers use the same set of actuators with limited forces. This paper presents an investigation of the control input requirement for static load compensation and vibration damping. An algorithm for optimal control input distribution over the different control objectives is implemented to achieve good performance of the overall system. The method is tested in simulations considering a wind disturbance. By applying the introduced control strategy, the closed loop achieves a performance improvement of 14% with regard to the building's displacements and velocities compared to the application of solely a model predictive controller.

Keywords: System Modeling and Control
Abstract: This paper is concerned with control design methods via a new Lyapunov function for fuzzy descriptor systems. Takagi-Sugeno fuzzy descriptor model, that can represent nonlinear systems, are divided into two families: one with same membership functions in both sides of fuzzy systems and another with different membership functions in right and left side of fuzzy systems. Although many papers propose control designs for the above two systems that are derived from Lyapunov stability theory, they are very conservative because they use a Lyapunov function including a single Lyapunov matrix. In this paper, to design less conservative conditions, we propose new Lyapunov functions with all Lyapunov matrices that are multiple. Finally, numerical examples are given to show the effectiveness of our methods.

Keywords: System Modeling and Control, Mechatronics, Robotic Systems
Abstract: This paper reveals that the turning radius of a tilt-rotor UAV during high-speed flights can be significantly reduced by utilizing the redundancy in control inputs. Based on a dynamic model of the UAV, the flight conditions of steady level turns are analyzed to numerically search for the minimum turning radius when all the actuators installed in the UAV are utilized. Numerical results indicate that adding lift with sub-rotors installed for rotary-wing mode makes it possible to increase the roll angle during level turns while reducing the turning radius. The tilt angle of main rotor can also be chosen within a certain range to reduce the turning radius. The feasibility of such steep turns is checked by numerical simulations using a feedback control.

Keywords: Optimization and Self-Organization Approaches
Abstract: In semiconductor manufacturing, a multicluster tool is widely employed for most wafer fabrication processes. With the demand for high-mix integrated circuit chips and shrinkage of circuit width, a scheme that multiple wafer types are fabricated inside multicluster tools is adopted by wafer foundries to earn more profits. Multiple wafer types, multiple robots, and wafer residency time constraints make these scheduling problems challenging. This work focuses on scheduling a single-arm multicluster tool to process two wafer types concurrently subject to wafer residency time constraints in which a conventional one-wafer cycle and backward strategy are not efficient. To our best knowledge, it is the first time to propose a one-wafer-per-type cycle and two-wafer-type backward sequence for such a scenario. With such properties, several necessary and sufficient conditions are presented to check the feasibility of a periodic schedule. Further, polynomial time algorithms are proposed to match the schedulability and coordinate multiple robots to handle wafers for the schedulable scenarios. The cycle time of an obtained schedule can reach the lower bound. Finally, a practical example is used to show the effectiveness of the proposed algorithm.

Keywords: AI and Applications, Deep Learning, Optimization and Self-Organization Approaches
Abstract: Robust and performant controllers are essential for industrial applications. However, deriving controller parameters for complex and nonlinear systems is challenging and time-consuming. To facilitate automatic controller parametrization, this work presents a novel approach using deep reinforcement learning (DRL) with N-dimensional B-spline geometries (BSGs). We focus on the control of parameter-variant systems, a class of systems with complex behavior which depends on the operating conditions. For this system class, gain-scheduling control structures are widely used in applications across industries due to well-known design principles. Facilitating the expensive controller parametrization task regarding these control structures, we deploy an DRL agent. Based on control system observations, the agent autonomously decides how to adapt the controller parameters. We make the adaptation process more efficient by introducing BSGs to map the controller parameters which may depend on numerous operating conditions. To preprocess time-series data and extract a fixed-length feature vector, we use a long short-term memory (LSTM) neural networks. Furthermore, this work contributes actor regularizations that are relevant to real-world environments which differ from training. Accordingly, we apply dropout layer normalization to the actor and critic networks of the truncated quantile critic (TQC) algorithm. To show our approach's working principle and effectiveness, we train and evaluate the DRL agent on the parametrization task of an industrial control structure with parameter lookup tables.

Keywords: Artificial Immune Systems, Evolutionary Computation, Swarm Intelligence
Abstract: This paper classifies and summarises the historical literature on carousel systems in automated storage and retrieval systems in recent years. As an automated storage and retrieval system for distribution centers and production facilities, carousels facilitate the storage and dispatching of goods, significantly improving warehouse turnover efficiency. Their performance have been investigated by many scholars and experts. As carousels evolve and upgrade, more and more innovative algorithms have been used to improve the efficiency of outbound carousel storage. In this paper, we collate articles investigating how the carousel system is stored inbound versus retrieved outbound. We then discuss articles on the dualcommand model of automatic storage retrieval systems as a whole. By reviewing over 50 papers, we summarise research on how to store and unload goods, focusing on the performance of automatic storage retrieval systems under dual-command conditions. On this basis, we review the current research’s limitations and suggest future research directions.

Keywords: Consumer and Industrial Applications, System Modeling and Control, Modeling of Autonomous Systems
Abstract: The accumulation of bird droppings on photovoltaic (PV) farms reduces power generation efficiency and necessitates manual cleaning on a regular basis, which is a challenge in large power plants. To solve this problem, this paper proposes an automatic Unmanned Aerial Vehicle (UAV) based bird droppings detection, localization, and cleaning method on large PV power plant. An automated flight route is first created, and use an UAV to fly over the solar farm to capture images of the solar panels. The captured images are then stitched together to create a high-resolution orthomosaic image of the solar farm, which enables to precisely locate the bird droppings on the solar farm. An improved YOLOv7-based model is proposed to detect the bird droppings because they are quite small in comparison to the stitched image. Then, using the ground sample distance, we calculate the distance between each of the bird droppings and the drone’s takeoff point, which is used to clean the bird droppings from the solar panel. Last, the proposed model is verified by high-resolution orthomosaic images and the experimental outcomes unequivocally show that it is successful for detecting and cleaning of bird droppings on PV farms.

Keywords: Fault Monitoring and Diagnosis, Cooperative Systems and Control, Distributed Intelligent Systems
Abstract: In recent years, significant progress has been made in the application of data-driven, learning-based approaches to fault detection in distributed networks. These methods are optimized for quickly detecting and identifying faulty instruments, whether originating from within a single vehicle or from a network of connected vehicles. This paper provides a preliminary review of typical Fault Detection and Identification (FDI) techniques, with a focus on platoons of vehicles arranged in a rectilinear formation using a leader-follower architecture. Specifically, this paper discusses the advantages and disadvantages of data-driven versus model-based methods for addressing the FDI problem. In particular, the main characteristics of a novel immunity-based bio-inspired data-driven technique are highlighted, and numerical simulations of a multi-vehicle system under normal and faulty conditions are presented to support the discussion.

Keywords: Robotic Systems, Autonomous Vehicle, Mechatronics
Abstract: Injuries during the practice of sports in gymnasiums have been reported, and one of the causes of injuries is due to environmental factors as tiny wood splinters on the gymnasium floor. Although it is important to regularly inspect gymnasium floors, it is difficult for humans to inspect the entire gymnasium floor, as it is done manually and visually, and requires a lot of time and manpower. We have developed an automatic inspection system to detect tiny splinters on the gymnasium floor. The system attaches cotton to tiny splinters and detects the attached cotton by using an image processing technique. Using this system, the entire gymnasium floor can be inspected automatically by using simply creating a 2D map. After the inspection, the system can show where splinters are located on the map. In this paper, the method for detecting splinters attached to cotton using deep learning object detection-YOLO was proposed. The detection ratio of the proposed method was improved by 25.0 % compared to the conventional method of threshold color segmentation process. In an inspection of an entire gymnasium, the proposed method detected 33 markers and was able to detect splinters that could cause injury.

Keywords: Fault Monitoring and Diagnosis, System Modeling and Control
Abstract: Anomaly detection aims to detect data that do not conform to regular patterns, and such data is also called outliers. The anomalies to be detected are often tiny in proportion, containing crucial information, and are suitable for application scenes like intrusion detection, fraud detection, fault diagnosis, e-commerce platforms, et al. However, in many realistic scenarios, only the samples following normal behavior are observed, while we can hardly obtain any anomaly information. To address such problem, we propose an anomaly detection method GALDetector which is combined of global and local information based on observed normal samples. The proposed method can be divided into a three-stage method. Firstly, the global similar normal scores and the local sparsity scores of unlabeled samples are computed separately. Secondly, potential anomaly samples are separated from the unlabeled samples corresponding to these two scores and corresponding weights are assigned to the selected samples. Finally, a weighted anomaly detector is trained by loads of samples, then the detector is utilized to identify else anomalies. To evaluate the effectiveness of the proposed method, we conducted experiments on three categories of real-world datasets from diverse domains, and experimental results show that our method achieves better performance when compared with other state-of-the-art methods.

Keywords: System Architecture, System Modeling and Control, Decision Support Systems
Abstract: In this work, we propose a novel anomaly detection approach in images based on normality scores using transformations. By applying various transformations to the input image such as rotation and flipping, we train a classifier to predict the transformation label applied to the images. Then, we represent the output of the classifier by a softmax vector. Thanks to the flexibility of multivariate Beta in fitting the data compared to other conventional distributions such as the Dirichlet distribution, we approximate the softmax vector by this general form of the Beta distribution to construct the normality scores. Moreover, we use the Maximum Likelihood to estimate the parameters of the proposed distribution. To show the power and the effectiveness of our approach, we conduct experiments of detecting anomalies in various public datasets. Furthermore, the proposed method is compared with state-of-the-art techniques and results demonstrate its superiority in terms of Area Under Receiver Operating characteristics (AUROC).

Keywords: Autonomous Vehicle, Intelligent Transportation Systems
Abstract: This study reports the robust sideslip angle estimation of vehicles with an uncertain tire cornering stiffness and fading measurements. The missing measurement and possible inaccuracy in the measurement of the vehicle’s yaw rate are considered by using a random variable distributed over [0, 1]. Norm-bounded uncertainties are considered in the vehicle’s tire cornering stiffness. Next, the Lyapunov stability theory is used to design a sideslip angle estimator such that the filtering error dynamics is stochastically stable and the H∞ performance criterion is met. The desired parameters of the proposed H∞ sideslip angle estimator are gained by solving a linear matrix inequality (LMI) problem. Simulation results show that the proposed novel estimator can efficiently estimate the sideslip angle while it demonstrates robust performance to uncertainties and fading measurements.

Keywords: Robotic Systems, Trust in Autonomous Systems, System Architecture
Abstract: In modern factories and daily life, the application of heterogeneous robot groups is becoming more and more widespread. However, there are still many areas for improvement in the management and communication of heterogeneous robot swarms, including the need for different connection interfaces for heterogeneous robots and the use of heterogeneous robot action logs to identify possible bottlenecks in the production line or record unplanned behaviors, whether malicious or not. In order to better manage heterogeneous robot swarms, this paper introduces the Heterogeneous Robots Chain (HR-Chain), a blockchain-based solution that can manage different types of robots and prevent unnecessary changes in robot operation logs to help improve production efficiency or other management requirements. HR-Chain is a Tezos-based blockchain project that securely stores robot logs in the blockchain using smart contracts and designs a new consensus algorithm, Delegated Proof of Stake with node's Resource and Behavior (DPoRB), which is tailored to heterogeneous robot swarms to improve the efficiency and fairness of the consensus process. Finally, this paper conducts experimental research on HR-Chain, and the simulation results show that the new consensus algorithm has better performance in terms of throughput and so on. The real experimental results show that the robot system based on HR-Chain has better response speed and execution force, showing its potential application prospects in the industrial and consumer fields.

Keywords: Robotic Systems, System Modeling and Control, Smart Sensor Networks
Abstract: Localization is crucial for various automation systems to provide awareness of the robot's position and orientation. Additionally, a localization system that offers portability, flexibility, and low computational and economic costs is required by a variety of robotics applications. However, no existing system can offer all the aforementioned features suitable for motion capture tasks involving ground swarm robots. In this study, we propose MoCArU, a novel Motion Capture system based on odometry and ArUco, robustly recognized through image data with low computational cost. We have evaluated the system's performance by comparing it with the ground truth trajectory and adopting different numbers of cameras. The results show that MoCArU can achieve a root mean square error of 0.1345±0.0065 m using ten cameras. Our findings add to previous knowledge by presenting a robust and cost-effective alternative to existing localization methods. Here, we show that MoCArU's use of lightweight camera stands and wireless communication ensures ease of installation, portability, and low computational cost, making it suitable for tracking swarm ground robot systems. We anticipate this system to be used in various applications, such as robot position control, navigation, and obstacle avoidance control. Overall, MoCArU provides a reliable and cost-effective solution for the real-time localization of robots, so its wider applicability in various environments is a significant advantage in robotics. An open-source implementation of MoCArU, as well as its related details, is open for public use at https://www.rm.is.tohoku.ac.jp/MoCArU.

Keywords: Cooperative Systems and Control, Communications, Distributed Intelligent Systems
Abstract: The structure of a mobile ad hoc network changes dynamically based on node positioning. We consider a setting in which nodes can communicate if they are within a prescribed distance of one another, giving rise to a communication network. An example is a swarm of small satellites that cooperate to perform tasks; such swarms are likely to become commonplace in space missions. In this paper, we consider the problem of adding a new node or repositioning a current node in the network while optimizing a given network parameter such as network reliability. Although there are infinitely many locations to place the new node in space, there are only finitely many possible changes to the communication network. We provide an algorithm that enumerates all possible network changes in time O(n^2 log(n)) or O(n^3 log(n)), for networks in 2- or 3-dimensional Euclidean space, respectively. We apply the proposed algorithm to a satellite swarm formation planning problem, where the goal is to maximize network reliability.

Keywords: Autonomous Vehicle, Intelligent Transportation Systems, System Modeling and Control
Abstract: Modern vehicles are equipped with a growing amount of advanced driver assistance functions (ADAS), such as adaptive cruise control or collision avoidance systems. These functions heavily rely on the current state of the vehicle. Particularly for lateral vehicle dynamics functions, the state of the tires has a crucial influence. For moderate driving scenarios, the vehicle lateral tire forces are mainly described by the tire cornering stiffnesses, which give a linear relation between tire slip angles and the resulting lateral tire forces. However, the tire cornering stiffnesses heavily depend on the type of tire and on many parameters, most significantly the tire vertical load, the tire wear and the tire pressure. As such changes in tire cornering stiffness can lead to a completely different vehicle behavior, it is essential to obtain current cornering stiffness estimates for the application of ADAS. Thus, this paper proposes a combined beta-less approach for vehicle cornering stiffness and road-friction coefficient estimation based on a nonlinear tire model, which solely makes use of typical vehicle in-series sensors. Vehicle measurements on dry road as well as on ice show the effectiveness of the proposed method.

Keywords: Autonomous Vehicle, Communications, Trust in Autonomous Systems
Abstract: This paper presents a scenario-based safety assurance process for Automated Driving Systems (ADSs) combined with the Vehicle-to-Everything (V2X) connectivity aspect, system with connectivity capability is benchmarked with system without such capabilities. In addition, a novel approach to V2X modelling is introduced and implemented to obtain the required configuration of the V2X parameters for individual system, such modelling approach ensures that the V2X is effective within a system during testing by using a distance-based V2X parameter that correlates to the speed of the system. Such parameter is then used as the configuration of the V2X model when carrying out the ADS and V2X combined test. Using a pedestrian crossing scenario, a reduction in failure cases is demonstrated when combining the V2X and the ADS. Therefore, a synchronised approach of vehicle, sensor and communication sub-system can improve the overall safety function of the system.

Keywords: Mechatronics, System Modeling and Control
Abstract: Automation functionalities of mobile construction site vehicles continue to improve constantly with the growth of the industry. Advanced sensor technology and enhanced computational power allow for the development of innovative and robust assistance systems. By applying control algorithms to the machines, they are enabled to operate on a construction site autonomously. Whereas such use cases expand the utilization of these vehicles, existing features can be improved as well. For instance, load limit control avoids engine stalling of a mobile machine in critical scenarios by regulation of the engine dynamics. This work presents a feed-forward control approach for a telescopic handler's driving functionality, which is actuated by a hydro-mechanical powertrain and powered by a diesel engine. Application of the proposed method to the real system yields a quality of life improvement through a more reliable prevention of engine stalling due to a lower reaction time to external loads. Furthermore, the engine speed decrease is reduced by efficiently regulating the drive pump.

Keywords: Autonomous Vehicle, Robotic Systems, System Architecture
Abstract: The ability to perform efficient trajectory planning is crucial for UAV to carry out tasks autonomously. However, existing research on UAV trajectory planning often employs the cascade process method that involves high-precision maps, real-time positioning and path planning. These methods have limitations such as high computational complexity and time delay, which hinder the efficiency of trajectory planning. End-to-end trajectory planning methods offer a promising solution to this problem. As the core of these end-to-end methods, perception-end plays a decisive role in trajectory planning. But current multimodal fusion of perception is only post-fusion, lacks intermediate feature-level fusion and lacks attention to global visuospatial information. To solve these problems, we propose a new network architecture called FlyTransformer, which fuses the proprioceptive state and visual perception in feature-level for end-to-end trajectory planning. And the key visuospatial information can be attentioned in this architecture. We evaluate our method in forest and cuboid scenarios and their corresponding outdoor scenarios. The results show that FlyTransformer outperforms other baseline algorithms in terms of efficiency and performance.

Keywords: Autonomous Vehicle, Communications, Cooperative Systems and Control
Abstract: Recent developments in the autonomous vehicle industries have increased the significance of accurate positioning. Popular techniques for localization include the Global Positioning System (GPS). However, owing to the presence of obstructions, GPS signals are unavailable in dense urban environments. Moreover, in indoor environments (such as a parking garage below the ground), GPS signals are inaccessible to users. In this article, we introduce a novel technique for accurate indoor vehicular positioning. The first step in our proposed system is localization based on received signal strength (RSS) fingerprints of 5G New Radio (NR) downlink signals. Furthermore, to compensate for the high susceptibility of RSS fingerprinting techniques in varying environments, we propose a real-time collaborative localization scheme based on 5G sidelink device-to-device (D2D) communication. We develop extensive test campaigns to assess the efficacy of our proposed two-step scheme. According to test results, our proposed algorithm outperforms scenarios that rely solely on 5G RSS fingerprints.

Keywords: Consumer and Industrial Applications, Robotic Systems, Manufacturing Automation and Systems
Abstract: Autonomous robotic grasping of uncertain objects in uncertain environments is an impactful open challenge for the industries of the future. One such industry is the recycling of Waste Electrical and Electronic Equipment (WEEE) materials, in which electric devices are disassembled and readied for the recovery of raw materials. Since devices may contain hazardous materials and their disassembly involves heavy manual labor, robotic disassembly is a promising venue. However, since devices may be damaged, dirty and unidentified, robotic disassembly is challenging since object models are unavailable or cannot be relied upon. This case study explores grasping strategies for industrial robotic disassembly of WEEE devices with uncertain vision data. We propose three grippers and appropriate tactile strategies for force-based manipulation that improves grasping robustness. For each proposed gripper, we develop corresponding strategies that can perform effectively in different grasping tasks and leverage the grippers design and unique strengths. Through experiments conducted in lab and factory settings for four different WEEE devices, we demonstrate how object uncertainty may be overcome by tactile sensing and compliant techniques, significantly increasing grasping success rates.

Keywords: Control of Uncertain Systems, Adaptive Systems, Soft Robotics
Abstract: Harvesting mushrooms while maintaining their quality remains a challenging task due to their fast growth rate, uneven developmental posture, and soft, easily damaged characteristics. This study proposes a 3D cultivation medium that enables harvesting by deformation to eliminate the mismatch between harvesting and cultivation. The proposed medium embeds a 3D-printed anisotropic elastic support structure in the covering soil, allowing mushrooms to be generated from the sides and bottom. Furthermore, a mechanism is developed to detach the successfully grown mushrooms by uniaxial contraction of the medium, induced by the out-of-plane pressure generated during free-fall. We also discuss how our method can potentially improve yield compared to conventional flat plate mushroom cultivation.

Keywords: Digital Twin, Adaptive Systems, System Modeling and Control
Abstract: This paper introduces a novel framework for replicating manual processes in digital twins. The proposed approach allows the modeling of manual process variability, a key aspect often overlooked in the literature. Our research highlights the potential of AI to tailor the digital twin to specific contexts influenced by human factors. The model's accuracy can be improved with each simulation synchronization cycle through supervised machine learning, which enhances the alignment of the virtual and physical processes. The proposed digital twin framework aims to avoid discrepancies from the physical counterpart, disregarding decisions based on fixed parameters that do not evolve over time and ensuring a higher degree of realism of the virtual replica. Ultimately, we aim to design the digital twin as a realistic and effective decision-making aid for all human participants engaged in the digital twin loop.

Keywords: Service Systems and Organizations, System Modeling and Control, Decision Support Systems
Abstract: Gig work is a work style in which freelancers undertake one-off jobs based on their own convenience.One can see that the gig work system is a management system that matches one-off jobs with registrant freelancers under their free decision-making.This paper addresses a systematic design of the gig work system.To this end, first, we model the decision-making by the freelancers and the evolutions in remaining workloads.We express the probability that at least one of the registrant freelancers gets opportunities of one-off jobs based on their utility.Next, we design a model-based controller, which matches one-off jobs with them.Finally, the effectiveness of the proposed work management system is verified in numerical experiments.

Keywords: System Modeling and Control, Control of Uncertain Systems, Digital Twin
Abstract: The pharmaceutical manufacturing industry has been undergoing a paradigm change from the standard batch pharmaceutical manufacturing towards continuous manufacturing, which is a quicker, more effective approach. Even so, most work in continuous tablet manufacturing has focused on classic model predictive control strategies. Multi-parametric model predictive control strategies are a better fit for assuring a quality-by-control (QbC) approach and progressing towards embracing the Industry 4.0 revolution. In this work an advanced multi-parametric model based predictive control strategy for the control of a continuous manufacturing process for solid dosage forms in the pharmaceutical industry is developed. The model used to develop the advanced control strategies is validated and calibrated using real data from the Pilot Plant. This will lead to better process performances including increased process efficiency, expanded process flexibility, and decreased environmental impact despite the presence of process uncertainties, measurement noise and disturbances.

Keywords: Enterprise Information Systems, Discrete Event Systems, Decision Support Systems
Abstract: In recent years, a form of production known as high-mix low-volume production has attracted attention. Mass customization is a production strategy that manufactures produce products to meet diverse customer demands while maintaining production efficiency close to mass production. Modular production is one of the production methods to achieve mass customization, and products are manufactured by combining modules. This production method allows a manufacturer to increase its total profit by manufacturing products suited to individual customers while reducing costs. In actual production planning, production costs and customer demands are uncertain. By considering these uncertainties, robust production planning can be achieved. Various robust production planning methods have been proposed in previous studies, however, no previous studies have reported cases of modular production that reflect customer decisions. Therefore, in this study, the production planning problem is formulated as a bilevel production planning problem consisting of customers and a manufacturer, and robust optimization is applied. Finally, we verify the usefulness of the proposed method for customer demand uncertainty by comparing the conventional model with the robust optimization model.

Keywords: Decision Support Systems, Soft Robotics
Abstract: Particle Swarm Optimization (PSO) (Kennedy and Eberhart, 1995) is an evolutionary computation technique that imitates the foraging behavior of flocks of birds. Particles, with position and velocity information, explore the search space and efficiently search for solutions by sharing information among them. However, PSO has limitations such as early convergence to a local solution and difficulty in finding a global solution due to insufficient search in the wide range of the target space. To fix this issue, TCPSO (Two-swarm Cooperative Particle Swarm Optimization: TCPSO) (Sun and Li, 2014) employes two types of particles, slave particles for intensive search and master particles for global search. However, even TCPSO may struggle with high-dimensional and complex problems and fall into a local solution. This study aims to improve the performance of TCPSO by estimating the function of the target problem using Gaussian process regression during the solution search process.

Keywords: Adaptive Systems, Cooperative Systems and Control, Autonomous Vehicle
Abstract: This paper proposes an integrated software-hardware approach for automated bathymetry mapping using low-cost unmanned surface vehicles (USVs). The solution takes inspiration from adaptive sampling, aiming to minimize the time, hardware, and training costs compared to traditional coverage path planning (CPP) methods. The proposed approach implements a scalable Markov Decision Process (MDP)-based model and a planning method based on multi-agent approximate dynamic programming (ADP), which prescribes USVs to seek the most informative samples in bandwidth-limited environments with uncertain and incomplete prior map information. The control structure offloads surrogate model updates and path planning to the control station and ensures mission completion even with communication dropouts. Computer simulation results on both real-world and computer-generated bathymetry data demonstrate the effectiveness of the proposed approach in terms of predictive accuracy and efficiency. The proposed solution has the potential to significantly aid coastal development by streamlining access to accurate models of near-shore bathymetry.

Keywords: System Modeling and Control, Control of Uncertain Systems
Abstract: In this paper, we propose an efficient data-driven predictive control approach for general nonlinear processes based on a reduced-order Koopman operator. A Kalman-based sparse identification of nonlinear dynamics method is employed to select lifting functions for Koopman identification. The selected lifting functions are used to project the original nonlinear state space into a higher-dimensional linear function space, in which Koopman-based linear models may be constructed for the underlying nonlinear process. To address the potential issue of a significant increase in the dimensionality of the resulting full-order Koopman models caused by the use of lifting functions, we propose a reduced-order Koopman modeling approach based on proper orthogonal decomposition. A computationally efficient linear robust predictive control scheme is established based on the reduced-order Koopman model. A case study on a benchmark chemical process is conducted to illustrate the proposed framework.

Keywords: Mechatronics, Infrastructure Systems and Services, Robotic Systems
Abstract: In recent years, infrastructure facilities have been noticeably deteriorating and need to be inspected. The use of pipe-inspection robots has been investigated, but the robots developed so far are too large for smaller pipes. As the robot is downsized, however, the cables used for communication and power transmission disturb the movement of the robot. In this report, we utilized microwave guided modes propagating along a pipe wall as a communication method. An experimental robot platform for small-diameter pipe inspection was constructed. We confirmed the feasibility of the communication method by conducting a 3D electromagnetic field simulation and experiments on the transmission properties. In particular, the experiments on data communication and picture transmission demonstrated the possibility of communication with the robot. A small pipe-travelling robot platform for experiments on wireless communication was designed and fabricated. We confirmed that the prototype robot was able to communicate wirelessly by using microwave guided modes propagating along the pipe wall.

Keywords: System Modeling and Control, Smart Sensor Networks, Control of Uncertain Systems
Abstract: In this paper, we present a comprehensive approach for designing and analyzing control systems with min-max constraint controllers and machine learning-based virtual sensors. By leveraging the Standard Nonlinear Operator Form (SNOF), we establish the necessary conditions for global asymptotic stability and demonstrate their applicability through an illustrative example based on a modified plant model from the literature. The proposed methodology effectively handles nonlinearities and constraints, ensuring stability while providing a systematic procedure for constructing a well-formed SNOF by integrating the plant, virtual sensor, and controller. The successful application of this method in the example highlights its potential for addressing complex control problems involving min-max constraints and virtual sensors in real-world scenarios. This paper contributes to the growing body of knowledge in this area and sets the stage for future advancements, including the exploration of transforming other machine learning architectures into the SNOF and extending the stability analysis to accommodate different types of nonlinearities and constraints.

Keywords: Medical Informatics, Systems Safety and Security, Environmental Sensing,
Abstract: In recent years, the rapid development of the Internet of Things (IoT) has attracted significant interest in smart healthcare. However, such collaborative IoT applications still face three major challenges: multi-task data heterogeneity, high communication cost, decentralized computation and privacy preservation. As a result, efficiently integrating and processing such complex data has become a crucial problem to solve. In this paper, we propose COFFE, a collaborative federated learning framework that allows edge devices to participate in the model training. Our framework leverages the local computing power of edge devices and utilizes the communication resources of the IoT network to enable distributed learning. We apply a compression algorithm which optimizes the communication cost by minimizing the amount of exchanged data between the edge devices and the cloud server. We demonstrate the effectiveness of our approach through a set of experiments on a simulated IoT network. Our results show that our proposed COFFE approach achieves significant performance gains compared to traditional centralized learning approaches. Overall, our work provides a promising direction in smart healthcare for applying federated learning to IoT setting and enabling edge device collaboration.

Keywords: Medical Informatics, Biometrics and Applications,
Abstract: Biomedical named entity recognition (BNER) is a basic task of the extraction of medical information. The existing deep learning-based approaches usually represent the medical text by using words or characters. However, most of biomedical terms consist of many words (characters). Splitting them into many fragments (words or characters) while leveraging the attention mechanism to assign attention scores for each fragment maybe disperse the importance weight and cause a lower attention score for the biomedical terms. Therefore, this paper presents a dictionary-based adaptive attention network for BNER. Specifically, a biomedical dictionary is firstly constructed by integrating multiple existing medical resources. Secondly, building the guidance vectors by matching the electronic medical record (EMR) text to the constructed dictionary. Then, an adaptive attention strategy is presented to guide the attention mechanism to assign higher attention to the overall medical term by using the guidance vectors. We conduct extensive experiments on a real-world dataset, the results illuminate that our presented method outperforms all baselines.

Keywords: Systems Safety and Security
Abstract: As cyber-attacks show to be more and more complex and coordinated, cyber-defenders strategy through multi-agent approaches could be key to tackle against cyber-attacks as close as entry points in a networked system. This paper presents a Markovian modeling and implementation through a simulator of fighting cyber-attacker agents and cyber-defender agents deployed on host network nodes. It aims to provide an experimental framework to implement realistically based coordinated cyber-attack scenarios while assessing cyber-defenders dynamic organizations. We abstracted network nodes by sets of properties including agents' ones. Actions applied by agents model how the network reacts depending in a given state and what properties are to change. Collective choice of the actions brings the whole environment closer or farther from respective cyber-attackers and cyber-defenders goals. Using the simulator, we implemented a realistically inspired scenario with several behavior implementation approaches for cyber-defenders and cyber-attackers.

Keywords: Medical Informatics
Abstract: 因为在 最常见的乳房 X 光检查图像，肿瘤通常 仅占整个图像的一小部分，这 经常使深度学习模型失去对 肿瘤区域。在以前的工作中，大多数模型都解决了这个问题 使用 ROI 标记来训练模型的问题，即 价格昂贵且难以广泛应用。一些最近的 无ROI方法使用多尺度特征或多阶段 训练，摆脱模型对投资回报率的依赖 但大大增加了计算复杂度和 部署难度大，限制了深度的潜力 神经网络。因此，深 基于位置表示的本地信息保留 提出了网络（PR-LIR）。PR-LIR 是一个单一功能， 端到端乳房X光图像分类方法，其中 使用位置表示 （PR） 和多尺度 区域汇集（MRP）模块，用于定位肿瘤区域和 保留小靶点肿瘤的区域语义信息 不同规模，无需ROI标签和多级 训练，并且几乎没

Keywords: Human Performance Modeling
Abstract: Monocular single-frame 3D human pose estimation (HPE) has garnered significant interest, particularly in the domains of human-computer interaction and human action recognition. Several remarkable works have achieved excellent results in obtaining accurate 2D human poses. Building upon the foundation laid by previous advancements, this paper focuses on 2D-3D lifting. The proposed 2D-3D lifting algorithm in this paper is a transformer-based model, which demonstrates its unique advantages in processing sequential data. The self-attention mechanism in the transformer can handle global information without being limited by the receptive field. The transformer's superior ability to process global information enables it to adaptively learn the relationships between human joints across different human behaviors. However, directly estimating 3D human pose from a single 2D image is a complex task due to depth ambiguity and joint occlusion. This ill-posed problem arises from the limited information available in a 2D image, making it challenging to determine the precise 3D pose. Additionally, occlusions further complicate the accurate estimation of joint positions. This article introduces a transformer-based network that enhances the algorithm's robustness by utilizing multi-layer dual-stream blocks. One path concatenates the 2D coordinates and the focal length of the camera as input, while the other path takes the difference between the 2D coordinates as input. The two paths undergo an information fusion process at the front end of the block. Experiments conducted on various datasets verify the effectiveness of the algorithm and achieve state-of-the-art performance on Human 3.6M and MPI-INF-3DHP benchmarks. Our code will be made publicly available on GitHub.

Keywords: Biometrics and Applications,, Design Methods, Networking and Decision-Making
Abstract: Obstructive sleep apnea (OSA) is considered one of the most common sleep disorders, causing multiple organ multiple system dysfunction and leading to a series of complications such as depression, insomnia, stroke, and so on. Piezoelectric ceramic sensor array (PCSA) with the merit of easily embedding into a mattress is a quite promising tool to monitor OSA events at home. Previous efforts achieved promising results for OSA detection based on PCSA, however, there are two major challenges still on open:（1）how to screen out the high-quality PCSA signals;（2）how to alleviate the difference among PCSA signals from different individuals. To address these challenges, we propose a cross-domain deep neural network for OSA event detection named CDDNNet. To obtain the high-quality PCSA signals, we propose a new dynamic channel-selection algorithm with maximizing energy minimizing variance of PCSA signals. To make CDDNNet with the cross-domain learning ability, we employ the gradient reverse learning (GRL) technique to alleviate the difference among PCSA signals from different individuals. What’s more, we conduct a pilot study to collect overnight PCSA signals from the Peking Union Medical College Hospital. Experiment results show that CDDNNet can achieve competitive results of 54.20% sensitivity,90.30% specificity, and 84.68% accuracy for OSA event detection.

Keywords: Virtual and Augmented Reality Systems, Virtual/Augmented/Mixed Reality, Human-Computer Interaction
Abstract: The mental health of older adults is a vital issue in the era of global aging. Empathy towards older adults constitutes a crucial component of social interaction, as it engenders awareness of the physical and mental obstacles encountered by this population. Empathy promotes individuals to be more friendly, considerate, and prosocial when interacting with older adults on various occasions, such as volunteering and professional caring. Various approaches have been used to promote people's empathy, but disadvantages like low participation enthusiasm, high cost, and time-consuming cannot be ignored and are not easy to deal with. Aimed at facilitating people's empathy towards older adults, we developed EmpathiaVR, a virtual reality older adults experience simulation system. It provides a multi-sensory mixed reality experience, including vision, hearing, and kinesthesis, from an older person's perspective, which is beneficial for provoking users' empathy towards older adults. To investigate the effectiveness of the system, an empirical study was conducted with 24 participants. The experiment employed a between-subjects design with two groups. The experiment results from both the subjective reports and behavioral variables indicated that the system enhanced people's empathy.

Keywords: Cloud, IoT, and Robotics Integration, Application of Artificial Intelligence, Cybernetics for Informatics
Abstract: The growing prevalence of cybersecurity threats has increased the demand for robust intrusion detection systems (IDSs). Deep learning techniques have shown promising results in detecting and mitigating these threats, making them an increasingly popular choice in IDS design. However, evaluating the performance of deep learning-based IDSs can be challenging due to the complexity of the models and the lack of standardized evaluation metrics. This review paper presents an overview of the most common evaluation metrics used in deep learning-based IDSs, including precision, confusion metrics, accuracy, F1 score, Area Under Curve (AUC), and recall. Several studies have applied machine-learning classic algorithms like Random Forest, Decision Tree, Logistic Regression, and others, but for this paper, we used a Convolutional Neural Network (CNN) that would be independent of the features in the dataset. The studied papers did not provide AUC and none of them balanced the dataset based on the feature's proportion. The dataset utilized in this study is the CSE-CIC-IDS2018 dataset, which underwent meticulous cleansing and normalization procedures to ensure the inclusion of legitimate and useful data. Furthermore, a weighting mechanism was introduced to balance the dataset and mitigate the potential for bias in the Machine Learning process.

Keywords: Big Data Computing,, Computational Intelligence, Machine Learning
Abstract: Runoff prediction is essential for flood forecasting, irrigation planning, and sustainable water resource management. However, accurate predictions can be challenging due to the involvement of multiple variables. This paper presents a novel Graph Convolution-based Spatial-temporal Attention LSTM Multi-Task learning (GC-SALM) model for accurate runoff predictions. Our approach combines a multilayer neural network and an attention mechanism for enhanced generalization performance. The GC-SALM model employs spatial attention and graph convolutional networks to discern local and global spatial patterns, while temporal attention and LSTM are utilized to capture temporal characteristics within extended sequences. Experimental results reveal that the proposed model outperforms six state-of-the-art methods in runoff prediction and flow calibration, emphasizing its potential for real-world hydrological applications.

Keywords: Cloud, IoT, and Robotics Integration, AI and Applications, AIoT
Abstract: Video analysis typically requires a significant amount of computing resources and energy. Traditional cloud-based video analysis relies on concentrating computing resources in the cloud, which puts a tremendous load on network bandwidth and introduces poor user experience due to latency. Edge computing offers a solution by offloading a portion of the video analysis tasks to the edge, significantly reducing the pressure on bandwidth and latency. However, the fixed video configurations (resolution, frame rate, etc.) uploaded to the edge may not be optimal for specific situations. Moreover, using a low video configuration makes it difficult for edge video analysis systems to recognize environmental changes, while using a high video con- figuration poses challenges to latency and bandwidth. Therefore, in this paper, we propose the MACEdge—an edge video analysis framework by leveraging the multi-access capability of the edge gateway to enhance edge perception by accessing heterogeneous sensors. We offer an algorithm based on Q Learning for dynamic adaptive adjustment of sensor threshold and video configurations. In experiments using real-world data, our collaborative solution improved accuracy and reduced high bandwidth costs compared to other video configuration adjustment benchmarks.

Keywords: Cloud, IoT, and Robotics Integration, AIoT, Deep Learning
Abstract: In the context of the Internet of Things (IoT), ensuring public safety relies heavily on the real-time detection of anomalies in monitoring systems. Thus, video anomaly analysis has garnered significant attention. We propose a collaborative anomaly analysis system(RACCS) that leverages cloud-edge-end computing. The system receives real-time video packets generated by cameras through edge gateways and utilizes load balancing algorithms to determine the destination node, reducing system latency. Using the MQTT protocol of the IoT, the video packets are transmitted to edge nodes or the cloud for inference. To prevent highly confident video packets from being transmitted to the cloud for secondary inference, we deploy a lightweight model on the edge nodes for initial inference, thereby reducing bandwidth consumption and latency. To achieve this goal, we adopt specific quantization methods for both the input video packets and complex cloud model parameters. This converts the complex cloud model into a lightweight model, reducing model storage and improving inference speed. We also propose a joint optimization framework for compressive sensing and anomaly classification that meets both high compression rates for video packets and high accuracy for anomaly classification. We further conduct a series of experiments to determine the optimal configuration and compare our approach with other mainstream methods, demonstrating the superiority of our system.

Keywords: Cloud, IoT, and Robotics Integration, AIoT, Machine Vision
Abstract: The real-time query of surveillance video plays a significant role in many fields such as public safety, smart city, and abnormality monitoring. However, with the exponential growth of surveillance video data, traditional cloud-based intelligent video processing faces significant challenges in terms of latency and bandwidth, while the pure edge computing approach is deficient in query accuracy due to its lack of computational power. Existing edge cloud collaboration approaches, such as SurveilEdge, focus on real-time target queries within a single video stream and do not show promising results during target queries across multiple video streams. For this reason, this paper proposes QueryEdge, an edge-cloud collaborative real-time query system for multiple video streams. Specifically, we design a real-time query system based on an edge-cloud collaboration framework to achieve highly accurate and low-latency target query services in multiple video streams. In addition, we introduce a prioritization mechanism and a load-balancing strategy in the query task scheduling process to further improve query efficiency. The evaluation proves that QueryEdge has a significant improvement in query latency and bandwidth consumption compared with pure cloud computing, pure edge computing, and SurveilEdge.

Keywords: Biometric Systems and Bioinformatics, AI and Applications, Deep Learning
Abstract: High-performance anti-spoofing countermeasures (CMs) have been widely used to protect automatic speaker verification systems by identifying and filtering spoofing speech. However, their performance degrades severely when confronted with out-of-domain (OOD) samples. To solve this issue, recent papers investigate the strategy that a CM can opt for abstention when it is not confident about the decision. In this paper, we develop an effective approach to enhance the performance of CMs with abstention. Specifically, we introduce supervised contrastive learning to group samples into known classes more tightly and employ data augmentation to enhance the diversity of known samples. The experiments are conducted on ASVspoof 2019 logical access corpus and another test set consisting of different OOD samples from other databases. With our scheme, CMs with abstention can achieve an equal error rate (EER) of 1.84% on the LA test set and 0.86% on the other test set. These results demonstrate that CMs trained with our approach show better OOD detection performance and can make more confident decisions.

Keywords: Cloud, IoT, and Robotics Integration
Abstract: Assembly inspection is a crucial aspect of smart manufacturing to guarantee the quality of products. To address the challenges posed by complex workshop scenes, an assembly inspection algorithm based on multi-view stereo is proposed. Specifically, to enhance the accuracy of reconstruction, the multi-view stereo method utilizes segmentation attention-assisted depth estimation and point cloud fusion to mitigate the noise in the reconstructed point cloud. And the proposed method introduces normalized depth loss to improve the reconstruction ability of foreground pixels. Futhermore, this paper presents a novel 3D point cloud-based size estimation algorithm capable of accurately estimating the size of small objects in complex working scenes. Experimental results demonstrate the efficacy of the improved multi-view stereo algorithm in assembly scenes and verify the superiority of the proposed size estimation algorithm.

Keywords: Neural Networks and their Applications, Deep Learning, Computational Life Science
Abstract: Most colorectal cancers originate from adenomatous polyps, which start as single asymptomatic polyps and develop into malignant tumors. In clinical practice, colonoscopy is an extremely effective method for detecting polyps, and it provides important visual information for the accurate identification and removal of polyps. However, it is highly challenging to achieve accurate segmentation of various polyps due to the complex and variable size, shape, color, number, and growth background of polyps at different periods. To address these dilemmas, we propose a multi-distance feature dissimilarity-guided encoder-decoder network for automatic polyp segmentation, mainly consisting of the Multi-Distance Differential Module (MDDM) and the Hybrid Loss Module (HLM). The former mainly utilizes the multilayer feature subtraction (MLFS) operations to extract the difference information between short-distance adjacent layer features and short-distance and long-distance cross-layer features. Given this, the pyramid-inspired MDDM obtains discriminative features continuously between adjacent/cross layers, enhancing complementary features between different layers. The latter supervises the feature maps extracted at each network level to achieve finer predictions. Experiments on four challenge datasets confirm that the proposed model outperforms most state-of-the-art methods in six evaluation metrics while yielding reasonably accurate segmentation results.

Keywords: Image Processing and Pattern Recognition, Machine Vision, Machine Learning
Abstract: In this paper, we propose a dual-branch Siamese network for visual object tracking. The proposed network consists of two distinct branches: a shallow network branch and a deep network branch. The shallow network branch focuses on precisely locating the target object and improving the anti-interference ability to similar objects, while the deep network branch focuses on capturing the more abstract semantic features of the object. Additionally, a multi-scale key feature fusion module is embedded into the shallow network, enabling the model to accurately locate the target object. Furthermore, we leverage the attention mechanism to further enhance the robustness of the model. Experimental results on three different public datasets demonstrate that our method outperforms state-of-the-art tracking algorithms.

Keywords: Cognitive Computing, Human Factors, Human-Machine Interaction
Abstract: Situation awareness of human and artificial agents can be improved by the recognition and adequate representation of real-life situations. The lack of easy-understandable, easy-to-use, and effective computational models of situations hindered the adoption and diffusion of situation awareness approaches in modern human-machine systems. Context Space Theory is a context awareness approach that uses a geometric metaphor to provide integrated mechanisms for representing contexts and situations. A drawback of this approach is the expert-based definition of context and situation spaces. This process can be time-consuming and expensive. In this paper, we propose a novel approach, namely Machine Learning-based Context Space Theory, which adopts machine learning techniques and, in particular, decision trees, to semi-automatically define context spaces and situation spaces with a data-driven approach. A case study related to the monitoring and control of the Covid-19 pandemic in Italy is proposed to demonstrate the feasibility and benefits of the proposed approach.

Keywords: Machine Vision, Image Processing and Pattern Recognition, Neural Networks and their Applications
Abstract: As human life continues improving, applications like virtual reality (VR) and augmented reality (AR) necessitate increasingly higher-quality 3D reconstructions. With the advancements in neural implicit 3D surface and volume rendering, multi-view 3D reconstruction has garnered significant attention. A prevailing issue in this domain is that the direct combination of neural implicit surface and volume rendering typically considers only photometric consistency loss, leading to an under-constrained surface problem. To address this problem, we develop a double-supervised approach and an integrated network for implicit surface and volume rendering, enabling the generation of a 3D surface model of an object from a set of multi-view images. In our approach, the object surface is represented by a signed distance function, and a 3D surface reconstruction model is jointly trained with photometric consistency constraints and geometry constraints. Experiments demonstrate that by enhancing prior geometry supervision and integrating the double-supervised network architecture into neural implicit 3D reconstruction, we can achieve accurate and high-quality 3D reconstruction.

Keywords: Neural Networks and their Applications, Deep Learning, Image Processing and Pattern Recognition
Abstract: Many Computer-Generated (CG) images are spreading widely on the Internet, which may deliberately misinform or deceive the public. Therefore, distinguishing CG images from natural photographic (PG) has become a frontier research topic in the field of image forensics. Although many algorithms have been proposed, it is still very challenging to detect CG images generated by the recent cutting-edge generative methods. Besides, most existing algorithms tend to generalize poorly when facing different unseen multimodal generative models. To address this issue, a novel method based on amplified texture differences learning is proposed to tackle this problem. We first design a deep texture enhancement module for discriminative texture amplification. Specifically, a semantic segmentation module is utilized to generate semantic segmentation map for the affine transformation operation guidance, which can be further used to recover the texture in different regions of the input image. Then, the combination of the original image and the high-frequency components of the original and enhanced images are fed into a hybrid neural network equipped with attention mechanisms, which refines intermediate features and facilitates trace exploration in spatial and channel dimensions respectively. By verifying on several commonly used benchmark datasets and a newly constructed dataset with more realistic and diverse images, the experimental results demonstrate that the proposed approach outperforms some existing methods.

Keywords: Medical Informatics
Abstract: Acute respiratory distress syndrome (ARDS) prognosis has become integral to modern critical care models aimed at determining expected patient outcomes, optimizing clinical pathways, and improving resource allocation. However, existing clinical studies of ARDS prognosis face severe challenges due to the redundancy of multiple sources of heterogeneous data in current healthcare datasets, the high dimensionality of patient features, and the wide variation in the importance of each feature. In this paper, we propose a holistic ARDS prognostic framework for assessing ARDS prognosis through a standardized data governance process and an integrated feature selection approach based on embedded algorithms. Specifically, we extract ARDS patient data from medical data sources by various criteria and normalize the features. Then we apply multiple embedded feature selection methods to obtain a decision matrix based on the data-governed dataset. We conducted extensive experiments to demonstrate the efficiency and superiority of our proposed framework. The experimental results show that our proposed framework performs well in both data governance and feature selection and has wide clinical application potential.

Keywords: Fault Monitoring and Diagnosis
Abstract: In practical industrial scenarios, it is difficult to collect sufficient labeled fault data for diverse machine types. How to train a robust intelligent fault diagnosis model with limited labeled samples and work well on different machine types remains a challenging issue. This paper proposes a novel cross-machine few-shot fault diagnosis method based on asymmetric distribution measure. The proposed method first constructs a large number of meta-diagnostic tasks from source machines with sufficient labeled fault data and then trains the meta-diagnostic model based on meta-learning to learn task-independent diagnostic knowledge, which will be transferred for the diagnostic task of the target machine to achieve accurate few-shot fault diagnosis. A series of cross-machine few-shot fault diagnosis experiments based on three rotating machinery datasets have verified the effectiveness of the proposed method.