Keywords: Human Factors, Ergonomics, and Performance in Intelligent Systems, Human-in-loop Systems and Architectures, Human-Machine Integration
Abstract: Interaction with touchscreens in moving vehicles can be problematic due to biodynamic feedthrough (BDFT), which occurs when the acceleration of the environment causes unintended movement of limbs. A potential solution is to develop models that predict BDFT based on the measured vehicle accelerations. The insights gained by these models can not only be used to improve the ergonomic design of, e.g., cockpits, but may also enable active, real-time correction of unintended touchscreen inputs. Although existing research has successfully identified BDFT models during touchscreen operation, how people adapt to varying motion intensities just before making screen contact is still unknown. To this end, this paper presents the results of a human-subject simulator experiment in which the effects of varying turbulence intensity on BDFT were studied during a pointing task. The key findings are that acceleration direction had the largest effect on neuromuscular stiffness (gain, natural frequency, and damping ratio), and that increasing turbulence intensity results in reduced BDFT gains. The experimental results shed light on how, and to what extent, BDFT model estimates differ according to varying motion conditions during pointing tasks. The implication for in-the-loop BDFT corrections is that it may be important to adapt the magnitude of the BDFT correction to the motion intensity.

Keywords: Human-Robot Interaction and Collaborative Robotics, Human Factors, Ergonomics, and Performance in Intelligent Systems, Human-in-loop Systems and Architectures
Abstract: Human–robot interaction is emerging as an important paradigm for integrating persons with disabilities into the workplace. While these systems can enable individuals to work, their design is mostly personalized, hindering widespread use beyond the individual user. The universal design paradigm is a central pillar of inclusive design, describing usability of systems by all. To incorporate universal design into process design for human-robot workplaces expert knowledge is required that is often not available. To simplify process design of human-robot workplaces, we propose a persona-based design approach. First, typical impairments prevalent in the workforce or particularly relevant for the processes are abstracted into personas with disabilities. The work process is subdivided into sequential actions. For each action and persona, strategies are developed to reach the action goal by a design thinking approach. The resulting actions are ordered by level of robot assistance, i. e. robot involvement, and implemented in a behavior tree. Therefore, the macro-behavior of the workplace may adapt to individual personas online. We demonstrate the method in a collaborative box folding process with a total of seven personas with disabilities. The persona-based process design shows promising results by generating more comprehensive process strategies while enabling adaptive behavior in the sense of universal design.

Keywords: Human-Machine Integration, Human Factors, Ergonomics, and Performance in Intelligent Systems, Trust, Transparency, and Ethical Governance in Human-Machine Systems
Abstract: Automation reduces pilot workload but introduces risks such as mode confusion and automation surprise, which can impair situation awareness and delay responses. Haptic shared control, delivered via active sidesticks, may mitigate these risks by providing continuous force feedback to convey automation intent and warnings. This study examined how the haptic information in shared control systems affects trust, workload, and awareness during automated take-off scenarios involving normal and abnormal autopilot behaviors. Twenty participants monitored automation performance while completing a secondary 2-back task under two shared control conditions: haptic shared versus input-mixing shared control. Results show that abnormal autopilot behaviors significantly increased workload and reduced trust, with haptic shared control having context-dependent effects - supporting trust and reducing workload during over-rotation but lowering trust in normal conditions. Haptic shared control had no significant effect upon participant situation awareness or secondary task performance. These findings suggest that the provision of haptic information within shared control can support pilots during automation anomalies but require human-centered implementation and training to avoid unintended effects under normal conditions.

Keywords: Human-Robot Interaction and Collaborative Robotics, Trust, Transparency, and Ethical Governance in Human-Machine Systems, Human-AI Collaboration and Decision Support
Abstract: Artificial intelligence (AI) has emerged as a central driver of transformation within the tourism industry, influencing destination planning, visitor engagement, business operations, and service delivery. This evolving tourism reality necessitates a rethinking of existing conceptual frameworks to explain how knowledge from computer science, business, marketing, and psychology converges within tourism systems, not only to enhance efficiency and personalisation, but also to ensure ethical business practices and consumer protection. In response, current paper introduces the Tourism AI Multiverse (TAIM), a novel taxonomy that captures the multilayered impact of AI across the tourism ecosystem while explicitly addressing risks associated with deceptive influence, discriminatory pricing, and covert behavioural steering. TAIM highlights the fundamental needs of tourism stakeholders for AI-driven solutions that shift the industry from a linear model towards a multidimensional and interconnected system grounded in transparency, accountability, and user autonomy. By revealing the growing dependence of tourism on AI and its implications for future development, the present framework provides researchers and practitioners with a structured lens for understanding and strategically leveraging this cutting-edge technology. It optimises performance and competitiveness, as well as opens avenues to prevent manipulative practices and to build trust-centred, AI responsible tourism systems.

Keywords: Trust, Transparency, and Ethical Governance in Human-Machine Systems
Abstract: Explainability of artificial intelligence (AI) is seriously questioned, despite the alarming legislation measures for global standards to ensure the safety and transparency of the systems in use. Comprehension of explainability perception, therefore, calls for thorough investigation. Present study addressed this challenge in the attempt to provide the very needed understanding on key determinants shaping AI explainability. EU citizens were asked to evaluate their experience with AI agent(s) after actual interaction in real life scenarios. Accuracy and competence emerged as key drivers of explainability perception. Explainability further interacted with functionality and quality perception. These effects were modulated by the level of satisfaction. While respondents in high satisfaction group seem to find the AI agent being explainable, respondents in low satisfaction group were hesitant about explainability. Naturally, then, future use intention was a function of explainability perception and satisfaction. These outcomes invite reconsideration of explainability offered by the AI systems currently available on the market, should we want to lift user satisfaction and future use, guaranteeing ethically sound applications.

Keywords: Human-AI Collaboration and Decision Support, Human Factors, Ergonomics, and Performance in Intelligent Systems
Abstract: Emotions play a critical role in learning by shaping motivation, attention, cognitive processing, and persistence. Emotionally Intelligent Tutoring Systems (EITS) address these dynamics by explicitly modeling learners’ affective states and integrating them into adaptive instructional decision-making. This paper presents a comprehensive survey of artificial intelligence (AI) techniques across the core functional components of a modular EITS architecture. Building on and extending an established reference architecture, we discuss AI-based approaches for multimodal emotion recognition, evaluation of affective states, adaptive tutoring strategy selection, and longitudinal personalization. Particular emphasis is placed on the role of neural methods and large language models, which enable nuanced natural language interaction and contextaware adaptation. Overall, this work illustrates how AI can be leveraged across system modules to establish a closed, affectsensitive feedback loop, providing a conceptual foundation for the development of effective, personalized, and emotionally adaptive intelligent tutoring systems.

Keywords: Human-Robot Interaction and Collaborative Robotics, Human-in-loop Systems and Architectures, Applications in Healthcare, Smart Infrastructure, and Environments
Abstract: Shared control robotic architectures have shown promise in their ability to compensate for the inaccuracies seen in non-implanted brain computer interface control devices, yet these existing methods have tended to have a fixed user-robotic authority. This paper evaluates a discrete-input shared-control artificial-potential-field algorithm that connects to a motor imagery BCI for indoor wheelchair navigation. We carried out a small exploratory feasibility study with 2 participants, both completing 8 trials, as well as by multiple simulated BCI studies, in varying levels of user authority. We also attempt to classify error related potentials collected during the study, to investigate new possible ways of identifying mistakes, and provide a supporting metric for a dynamic authority BCI system. We found that both participants found the low user authority system less challenging to use in comparison to the high user authority system, however, in simulation, we showed that the low user authority system was able to travel a significantly greater distance than the high user authority when simulated inaccuracies were introduced. From our results, we have discussed a few possible future studies from this exploratory study, as well as suggesting a method towards the development of the first dynamic authority BCI shared controlled wheelchair system.

Keywords: Human-in-loop Systems and Architectures, Human Factors, Ergonomics, and Performance in Intelligent Systems, Autonomous Systems: Control, Safety, and Reliability
Abstract: Human-machine symbiosis is a meta pattern of dynamic interactions between humans and machines, characterized by a subjective oneness between the human and the machine and a high degree of alignment in physically coupled shared control. Reaching and maintaining this symbiotic state is challenging because conflicts can arise from mismatched objectives, internal models, goals, and control actions. Therefore, this paper focuses on pre-symbiotic arbitration: a mediator algorithm is proposed that identifies and resolves conflicts in order to facilitate the transition toward the symbiotic state. We formalize symbiotic state, conflict state, and arbitration, and propose a pattern handler capturing directional opposition and effort mismatch between human and automation inputs. Using this pattern handler, we introduce a conflict-aware mediator model in which automation can arbitrate by goal yielding of its reference to adapt toward the human reference. In a simulation study using 2-DOF shared navigation with goal mismatch, we show that the proposed pattern handler can successfully model the goal mismatch. In addition, the conflict-aware mediator algorithm can resolve this conflict by arbitration of the automation goal, which leads to convergence of the human and the automation goals.

Keywords: Human Factors, Ergonomics, and Performance in Intelligent Systems, Human-in-loop Systems and Architectures
Abstract: This paper compares different multivariate signal processing methods - both based on classical statistical or machine learning models - for detecting human operator distraction in a preview tracking task. Binary Time Series Classification (TSC) methods proposed in prior work, which require both 'non-distracted' and 'distracted' training examples, were found to be out-performed by dedicated Anomaly Detection (AD) methods that aim to detect outliers from a learned 'non-distracted' prior distribution only. Even when applied in an ensemble learning framework, the TSC methods suffer from high false alarms and are highly sensitive to domain shift, due to their explicit reliance on representative 'distracted' training data. The considered statistical AD method, based on calculation of the Mahalanobis Distance (MD) of tested multivariate time-series samples, was found to be outperformed by a machine learning One-Class Support Vector Machine (OCSVM) method, as the latter more faithfully captured the (non-normal) distribution of the 'non-distracted' training data. Combining both AD methods in a meta detector only gave marginal improvement in the attained F1 score (0.5694) compared to the OCSVM (0.5610). The presented findings show that statistical and machine learning methods for AD can enable relevant detection of potentially impactful moments of distraction from human control data, which can be tested in real-time monitoring and adaptive support frameworks.