Human Factors Daily is a curated guide to recently published work across human–AI collaboration, intelligent mobility, accessibility, adaptive interfaces, and embodied AI.
New literature is scanned daily. High-confidence matches may be added automatically, while uncertain matches remain hidden for editorial review.
Research from outside BAT LabThese papers are not BAT Lab publications. They were written by researchers outside the lab and are shared as a free resource for learning and study.
Collection updated August 15, 2026·Literature last scanned August 31, 2026·14 selected papers
HFscan
Scanning recent literature
Current reading
New work, oriented for human factors readers.
Each entry orients human factors readers to a recent paper from the wider research community, with a source-based summary and a note on why it may matter.
Showing 14 papers
Intelligent Mobility
Examining Drivers’ Sense of Agency When Operating Level 3 Automated Vehicles: A Driving Simulator Study
Heather Aiken, Stephen G. Onyeukwu, Yue Hu, Youngbin Kwak, & Anuj K. Pradhan · Proceedings of the Human Factors and Ergonomics Society Annual Meeting
External publication · not a BAT Lab paper
Thirty-three participants reported greater perceived control during takeover scenarios than during scenarios in which the automated vehicle retained control. During takeovers, stronger agency was associated with lower perceived danger and greater agreement with the vehicle’s handoff decision; these relationships were absent during automated scenarios.
Why it matters
The findings show how takeover requests alter drivers’ perceived control and interpretation of danger, informing safer transitions of authority in Level 3 automation.
Human-Centric Composite Field Motion Planning for Ergonomics-Aware and Demonstration-Informed Human-Robot Collaboration
Chenzui Li, Yiming Chen, Xi Wu, & Fei Chen · International Journal of Social Robotics, 18, Article 81 · Open access
External publication · not a BAT Lab paper
The authors propose a composite-field motion planner that combines task geometry learned from human demonstrations with a continuous representation of ergonomic comfort. The planner generates smooth robot motions through gradient flow and executes them with compliant control. In simulations and physical collaboration experiments, it preserved demonstrated movement patterns while reducing ergonomic strain and muscle activation, showing how embodied systems can balance task fidelity with human comfort.
Why it matters
It connects embodied intelligence with real-time, ergonomics-aware collaboration in shared work.
This study develops a transferable mental-workload model for human–machine collaboration that uses eye-tracking data to help calibrate noisy EEG signals. Dynamic graph learning, cross-modal attention, and self-supervised pretraining identify features that generalize across people. Tests on two benchmark datasets showed strong cross-subject performance, including under noisy conditions, suggesting a path toward adaptive systems that can estimate operator workload without retraining for every user.
Why it matters
Transferable workload sensing can help adaptive interfaces respond to cognitive state without retraining for every user.
Driver Adaptation to Partially Automated Driving in Urban Environments: Effects of Repeated Exposure and System Capabilities on Drivers’ Trust, Monitoring, and Response
Elena Malaika Nkusi, Jasmin Verena Schneider, Maximilian Wiegand, & Klaus Bengler · Human Factors: The Journal of the Human Factors and Ergonomics Society · Open access
External publication · not a BAT Lab paper
Across two repeated-exposure simulator studies with 45 drivers, encounters with system limits helped calibrate trust and sustain monitoring, leading to faster and safer responses to a later system failure. The findings also indicate that current driver-monitoring systems may not fully protect against safety-critical adaptation during repeated urban use.
Why it matters
Repeated exposure shows how system limits recalibrate trust, visual monitoring, and intervention, informing adaptive warnings and driver-monitoring strategies for automated vehicles.
Enhancing End-user Engagement in Human–Robot Interaction by Performing LLM-driven Expressive Behaviors
Shipeng Lyu, Fangyuan Wang, Weiwei Lin, Guodong Guo, & David Navarro-Alarcon · ACM Transactions on Human-Robot Interaction
External publication · not a BAT Lab paper
The framework interprets dialogue with an LLM and maps inferred intent to physically executable speech and movement on a 58-degree-of-freedom humanoid robot. Video-based and live interaction studies found stronger expressiveness, appeal, and engagement than rule-based and handcrafted baselines.
Why it matters
It connects language-based intent inference with physically realizable robot behavior and evaluates whether expressive responses improve engagement in live human–robot interaction.
DiG-Net: Enhancing Human–Robot Interaction through Hyper-Range Dynamic Gesture Recognition in Assistive Robotics
Eran Bamani Beeri, Eden Nissinman, & Avishai Sintov · ACM Transactions on Human-Robot Interaction · Open-access preprint
External publication · not a BAT Lab paper
DiG-Net uses a single RGB camera and distance-aware spatiotemporal modeling to recognize dynamic gestures at distances up to 30 meters. It achieved 97.3% accuracy under varied distance and visibility conditions, supporting remote and nonverbal interaction with assistive robots.
Why it matters
Long-range gesture recognition can expand accessible, nonverbal control of assistive robots across home, rehabilitation, industrial, and remote-support settings.
Amplifying the Good, Muting the Bad: A Social Amplifier Model of Embodied Role Adaptation in Human–Robot Interaction
Yankuan Liu & Pei-Luen Patrick Rau · ACM Transactions on Human-Robot Interaction
External publication · not a BAT Lab paper
A confidence-aware NAO robot changed roles according to users’ confidence during an ambiguous decision task. Across 512 interactions, embodied communication clarified roles and amplified the benefits of socially appropriate robot behavior for trust, collaboration quality, and influence.
Why it matters
The study links confidence-aware robot role adaptation with embodied communication, showing how social appropriateness shapes role clarity, trust, and influence.
The effects of using a Virtual Agent to Signal Danger on Hazard Prediction ability in Conditionally-Automated Driving
Thomas Goodge, Frank Pollick, & Stephen Brewster · ACM Transactions on Computer-Human Interaction · Open access
External publication · not a BAT Lab paper
Across two experiments, attention-capturing visual and virtual-agent cues maintained drivers’ hazard-prediction performance at levels comparable with undistracted driving. The findings support socially expressive cues as a viable method for preserving situational awareness during non-driving tasks.
Why it matters
Virtual-agent cues offer a socially expressive approach to preserving hazard awareness when drivers are engaged in non-driving tasks.
An IMOI model for human-AI teams in critical care settings
Jenna Korentsides, Zander N. Miller, Elizabeth R. Merwin, Brooklyn Costinett, & Joseph R. Keebler · Human Factors in Healthcare, 9, 100137 · Open access
External publication · not a BAT Lab paper
The paper adapts the Input–Mediator–Output–Input model to describe recurring human–AI teamwork in intensive care and emergency settings. It links system and team conditions to processes such as trust and coordination, then connects those processes to team performance and patient outcomes through a feedback loop. The model is conceptual and has not been empirically validated, but it offers a structured foundation for training, interface design, and future simulation studies.
Why it matters
It offers a systems view of roles and responsibility in high-stakes human–AI teams.
Exploring the Role of Individual Characteristics in Shaping User Experience during Human–Robot Collaboration in Manufacturing Contexts
Riccardo Gervasi, Matteo Capponi, Luca Mastrogiacomo, & Fiorenzo Franceschini · Journal of Intelligent & Robotic Systems · Open access
External publication · not a BAT Lab paper
This study examines how personality and attitudes toward robots shape experiences with a collaborative robot during manufacturing assembly tasks. The researchers combined self-reported workload and affect with physiological indicators while participants completed tasks with and without robot assistance. Results suggest that individual characteristics meaningfully moderate the experience of collaboration, supporting adaptive cobot designs that account for differences among operators instead of assuming one interaction style will suit everyone.
Why it matters
The results point toward embodied systems that adapt to meaningful differences among users.
Human–AI collaboration: trade-offs between performance and preferences
Lukas W. Mayer, Sheer Karny, Jackie Ayoub, Miao Song, Danyang Tian, Ehsan Moradi-Pari, & Mark Steyvers · Cognitive Research: Principles and Implications, 11, Article 18 · Open access
External publication · not a BAT Lab paper
In two behavioral experiments, participants worked with AI agents that varied in how strongly they adapted to human actions during a shared decision task. People preferred agents that respected their intentions and allowed them to contribute meaningfully, while objective performance alone did not drive preference. The findings show that human-centered collaboration strategies can improve how AI teammates are perceived without necessarily reducing team performance.
Why it matters
The findings sharpen how role allocation can protect human agency without sacrificing team performance.
A Shared-Control Framework for A Human-Robot Front-Following Behaviour in Unknown Dynamic Environments
George Moustris & Costas Tzafestas · International Journal of Social Robotics, 18, Article 31 · Open access
External publication · not a BAT Lab paper
The authors present a mobile robot that follows a person from the front while navigating unfamiliar, obstacle-filled environments. Its shared-control framework combines local planning with intention recognition, shifting control between the person and robot when route choices become uncertain. Field trials showed highly accurate intent recognition and walking patterns closer to participants’ natural gait, while also revealing that leader and follower roles can change fluidly during assistive navigation.
Why it matters
It links assistive mobility with intent inference and responsibility handoff.
Bridging the gap: generating a design space model of socially assistive robots for older adults using participatory design methods
Adi Bulgaro, Ela Liberman-Pincu, & Tal Oron-Gilad · Universal Access in the Information Society, 25, Article 43 · Open access
External publication · not a BAT Lab paper
Using participatory design, the researchers worked with older adults and other stakeholders to identify expectations and concerns surrounding socially assistive robots. The work progressed from interviews and focus groups to direct interactions with a commercial robot, informing a model that connects robot characteristics with users’ emotional responses. The resulting process helps translate lived experience into design requirements and can be adapted to different applications and cultural settings.
Why it matters
The work demonstrates how participatory design can translate lived experience into accessible robot requirements.
Partner or burden? The dual pathways linking perceived attributes of intelligent cockpits to human–machine collaboration willingness via cognitive load
Silian Li · BMC Psychology, 14, Article 313 · Open access
External publication · not a BAT Lab paper
Using survey data from 503 licensed drivers, this study examines how perceived cockpit usefulness, entertainment value, and anthropomorphic design relate to willingness to collaborate with an intelligent vehicle system. Usefulness and entertainment were associated with lower cognitive load, which predicted stronger collaboration willingness. Anthropomorphism increased willingness through an affective pathway but did not reduce cognitive load, indicating that a socially appealing interface is not automatically easier to use in demanding driving contexts.
Why it matters
It argues for intelligent cockpit design grounded in cognitive resources rather than surface-level anthropomorphism.
These summaries are concise research orientations, not substitutes for the original papers. Follow the source links for methods, evidence, limitations, and complete author information.
