Kumar Akash
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kumar Akash has a total of 64 co-authors. Here are the top co-authors by number of shared publications:
- Teruhisa Misu
Honda Research Institute USA, Inc., San Jose, California, USA
14 shared publications - Shashank Mehrotra
Honda Research Institute USA, San Jose, CA, USA
12 shared publications Neera Jain
5 shared publicationsTahira Reid
5 shared publications- Mark Steyvers
University of California, Irvine, Irvine, USA
4 shared publications Kevin Salubre
Honda Research Institute USA, Inc., San Jose, CA, USA
3 shared publications
Latest Publications
- 25 October 2026
- Published by SAGE Publications in Proceedings of the Human Factors and Ergonomics Society Annual Meeting
Human Factors Extended Reality Showcase 2026
Extended reality (XR) technologies are increasingly adopted by human factors researchers and practitioners across domains ranging from manufacturing to workplace safety and ergonomics. This interactive XR Showcase session provides attendees with direct, hands-on engagement with XR devices and applications presented by researchers. This year’s demonstrations reflect growing interest in XR applications for skilled manual work and assembly, workplace safety and ergonomics, visualization, and mobility interaction. While training remains an established theme, an emerging trend is the integration of artificial intelligence within XR. Collectively, these demonstrations illustrate a broader shift in the field toward XR platforms that serve not only as visualization tools, but as integrated systems for performance measurement and human-AI collaboration. The XR Showcase advances the goals of HFES by strengthening community interaction and facilitating knowledge exchange among researchers, practitioners, and students interested in the expanding role of XR in human factors research and practice.
- 25 October 2026
- Published by SAGE Publications in Proceedings of the Human Factors and Ergonomics Society Annual Meeting
Human Factors Extended Reality Showcase 2026
Extended reality (XR) technologies are increasingly adopted by human factors researchers and practitioners across domains ranging from manufacturing to workplace safety and ergonomics. This interactive XR Showcase session provides attendees with direct, hands-on engagement with XR devices and applications presented by researchers. This year’s demonstrations reflect growing interest in XR applications for skilled manual work and assembly, workplace safety and ergonomics, visualization, and mobility interaction. While training remains an established theme, an emerging trend is the integration of artificial intelligence within XR. Collectively, these demonstrations illustrate a broader shift in the field toward XR platforms that serve not only as visualization tools, but as integrated systems for performance measurement and human-AI collaboration. The XR Showcase advances the goals of HFES by strengthening community interaction and facilitating knowledge exchange among researchers, practitioners, and students interested in the expanding role of XR in human factors research and practice.
- Open Access
- 4 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 1134-1142
Human-as-a-Sensor: Inferring Individual Engagement from Teammates Alone via Latent Group State Modeling
Can an individual’s internal state be inferred in the absence of direct observation? This question is increasingly critical in multimodal interaction settings where privacy constraints, occlusion, or sensor failure render self-centric sensing unreliable or unavailable. In this paper, we introduce a novel task formulation for multimodal interaction analysis, termed human-as-a-sensor, in which the state of a target individual is inferred exclusively from the multimodal behavior of their teammates. We investigate this problem in the context of group collaboration, focusing on the estimation of individual engagement and emergent leadership. To address the inherent challenges of this setting, we propose a Latent Group State Model (LGSM), a structured framework that first encodes teammates’ multimodal signals into a shared latent representation capturing group interaction dynamics, and subsequently performs target-conditioned inference. We evaluate the proposed formulation and model on a multimodal dataset of three-person teams. Our results demonstrate that LGSM significantly outperforms conventional machine learning, deep learning, and vision-language model baselines in the others-only setting. Analyses indicate that performance gains are most pronounced in interaction-dense segments, that multimodal fusion plays a critical role in improvements. Rather than substituting direct sensing, the proposed formulation offers a complementary paradigm for privacy-aware and robust multimodal interaction analysis in scenarios where direct observation is infeasible or intentionally restricted.
- 4 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 1134-1142
Open AccessHuman-as-a-Sensor: Inferring Individual Engagement from Teammates Alone via Latent Group State Modeling
Can an individual’s internal state be inferred in the absence of direct observation? This question is increasingly critical in multimodal interaction settings where privacy constraints, occlusion, or sensor failure render self-centric sensing unreliable or unavailable. In this paper, we introduce a novel task formulation for multimodal interaction analysis, termed human-as-a-sensor, in which the state of a target individual is inferred exclusively from the multimodal behavior of their teammates. We investigate this problem in the context of group collaboration, focusing on the estimation of individual engagement and emergent leadership. To address the inherent challenges of this setting, we propose a Latent Group State Model (LGSM), a structured framework that first encodes teammates’ multimodal signals into a shared latent representation capturing group interaction dynamics, and subsequently performs target-conditioned inference. We evaluate the proposed formulation and model on a multimodal dataset of three-person teams. Our results demonstrate that LGSM significantly outperforms conventional machine learning, deep learning, and vision-language model baselines in the others-only setting. Analyses indicate that performance gains are most pronounced in interaction-dense segments, that multimodal fusion plays a critical role in improvements. Rather than substituting direct sensing, the proposed formulation offers a complementary paradigm for privacy-aware and robust multimodal interaction analysis in scenarios where direct observation is infeasible or intentionally restricted.
- Open Access
- 4 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 1095-1104
Speaker Diarization in Static and Egocentric Videos via Speech-Face Alignment
- 4 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 1095-1104
Open AccessSpeaker Diarization in Static and Egocentric Videos via Speech-Face Alignment
- 24 September 2026
- Published by SAGE Publications in Proceedings of the Human Factors and Ergonomics Society Annual Meeting
Mobility Interaction Sim: A Scalable Experimental Research Platform for Mobility Research
This paper introduces Mobility Interaction Sim , a Unity-based experimental platform for human factors research on shared mobility environments where pedestrians, automated vehicles, micromobility riders, and delivery robots negotiate the same physical space. A single project compiles to two deployment targets: a Unity WebGL build for crowdsourced online studies and a native build for an in-laboratory virtual reality (VR) rig that incorporates a physical scooter handlebar, a head-mounted display, and wearable physiology. Scenarios, surveys, attention checks, and cognitive tasks are configured through the Unity Inspector, and a graphical Scene Graph manages condition assignment and counterbalancing. Behavioral and environmental states are logged as a time-stamped JSON stream from which standard kinematic measures such as time-to-collision, gap acceptance, and speed at the decision point can be derived. The platform is described together with three peer-reviewed case studies that employed it (two online, N = 208 and N = 200; one in-person VR, N = 50), as well as the interactive demonstration. The source code will be released under a non-commercial research license, with a browser-based demonstration and a source-access request form linked from the project page.
- 24 September 2026
- Published by SAGE Publications in Proceedings of the Human Factors and Ergonomics Society Annual Meeting
Mobility Interaction Sim: A Scalable Experimental Research Platform for Mobility Research
This paper introduces Mobility Interaction Sim , a Unity-based experimental platform for human factors research on shared mobility environments where pedestrians, automated vehicles, micromobility riders, and delivery robots negotiate the same physical space. A single project compiles to two deployment targets: a Unity WebGL build for crowdsourced online studies and a native build for an in-laboratory virtual reality (VR) rig that incorporates a physical scooter handlebar, a head-mounted display, and wearable physiology. Scenarios, surveys, attention checks, and cognitive tasks are configured through the Unity Inspector, and a graphical Scene Graph manages condition assignment and counterbalancing. Behavioral and environmental states are logged as a time-stamped JSON stream from which standard kinematic measures such as time-to-collision, gap acceptance, and speed at the decision point can be derived. The platform is described together with three peer-reviewed case studies that employed it (two online, N = 208 and N = 200; one in-person VR, N = 50), as well as the interactive demonstration. The source code will be released under a non-commercial research license, with a browser-based demonstration and a source-access request form linked from the project page.
- Open Access
- 19 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 233-237
Experts’ and Community Insights on Measuring Comfort and Perceived Safety in Automated Driving: Synthesis of a Participatory Workshop
- 19 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 233-237
Open AccessExperts’ and Community Insights on Measuring Comfort and Perceived Safety in Automated Driving: Synthesis of a Participatory Workshop
- Open Access
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-25
Exploring The Impact of Proactive Generative AI Agent Roles In Time-Sensitive Collaborative Problem-Solving Tasks
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-25
Open AccessExploring The Impact of Proactive Generative AI Agent Roles In Time-Sensitive Collaborative Problem-Solving Tasks
- 25 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 11 (4) , 4729-4736
Strategic Shaping of Human Prosociality: A Latent-State POMDP Framework
We propose a decision-theoretic framework in which a robot strategically can shape inferred human’s prosocial state during repeated interactions. Modeling the human’s prosociality as a latent state that evolves over time, the robot learns to infer and influence this state through its own actions, including helping and signaling. We formalize this as a latent-state POMDP with limited observations and learn the transition and observation dynamics using expectation maximization. The resulting belief-based policy balances task and social objectives, selecting actions that maximize long-term cooperative outcomes. We evaluate the model using data from user studies and show that the learned policy outperforms baseline strategies in both team performance and increasing observed human cooperative behavior.
- 25 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 11 (4) , 4729-4736
Strategic Shaping of Human Prosociality: A Latent-State POMDP Framework
We propose a decision-theoretic framework in which a robot strategically can shape inferred human’s prosocial state during repeated interactions. Modeling the human’s prosociality as a latent state that evolves over time, the robot learns to infer and influence this state through its own actions, including helping and signaling. We formalize this as a latent-state POMDP with limited observations and learn the transition and observation dynamics using expectation maximization. The resulting belief-based policy balances task and social objectives, selecting actions that maximize long-term cooperative outcomes. We evaluate the model using data from user studies and show that the learned policy outperforms baseline strategies in both team performance and increasing observed human cooperative behavior.
- 15 January 2026
- Published by Taylor & Francis in International Journal of Human–Computer Interaction
- Vol. 42 (16) , 1-28
Shared Control in Micro-Mobility: Effects of Rider Skill and Automation Levels on Subjective Well-Being in E-Scooter Use
- 15 January 2026
- Published by Taylor & Francis in International Journal of Human–Computer Interaction
- Vol. 42 (16) , 1-28
Shared Control in Micro-Mobility: Effects of Rider Skill and Automation Levels on Subjective Well-Being in E-Scooter Use