Human-Centered Physical AI and Human-Autonomy Teaming
Project Desciption
Future Physical AI systems must work effectively with humans in dynamic, uncertain, and safety-critical environments. This project will develop general methods for adaptive human-autonomy teaming by enabling autonomous systems to model human intentions, capabilities, preferences, and evolving states, and to use this understanding in collaborative decision making. The research will investigate Human Digital Twins, multimodal human modeling, continual adaptation, and transparent interaction as foundations for trustworthy collaboration between humans and embodied AI. The resulting methods will be designed to generalize across autonomous systems, including robots, intelligent manufacturing systems, healthcare technologies, and autonomous vehicles, with the broader goal of advancing Physical AI systems that can understand, anticipate, and collaborate with people in real-world environments.
Start Date
January 1, 2027
Postdoc Qualifications
Candidates should hold a Ph.D. in Computer Science, Human-Computer Interaction, Robotics, Electrical Engineering, Mechanical Engineering, Civil Engineering, Cognitive Science, or a closely related field by the start of the appointment. The ideal candidate will have demonstrated expertise in one or more of human-centered AI, computational human modeling, human-robot interaction, human-autonomy teaming, machine learning, embodied AI, multimodal sensing, behavioral modeling, or autonomous systems. Experience with user studies, human-subject experiments, statistical learning, explainable AI, or deploying AI systems in real-world environments is highly desirable. Applicants should be excited to conduct interdisciplinary research at the intersection of Physical AI, artificial intelligence, robotics, and human-centered computing while collaborating across multiple research groups. This closely matches Kexin's background in multimodal human sensing, adaptive autonomy, and explainable machine learning.
Co-advisors
Ziran Wang, ziran@purdue.edu, Lyles School of Civil and Construction Engineering, https://ziranw.github.io/;
Brandon Pitts, bjpitts@purdue.edu, Edwardson School of Industrial Engineering, https://engineering.purdue.edu/IE/people/ptProfile?resource_id=158978
Bibliography
Ma, Y., et al. LaMPilot: An Open Benchmark Dataset for Autonomous Driving with Language Model Programs. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024.
Cui, C., Ma, Y., Ye, W., & Wang, Z. Drive as You Speak: Enabling Human-Like Interaction with Large Language Models in Autonomous Vehicles. IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW), 2024.
Cui, C., et al. LLM4AD: A Survey of Large Language Models for Autonomous Driving. IEEE Intelligent Transportation Systems Magazine, 2025.
Pitts, B. J., Wang, Q., & Duerstock, B. S. AI-Enabled Accessible Travel in Autonomous Vehicles: Promises, Perceptions, and Prototypes. In Advances in Human-AI Collaboration. John Wiley & Sons, 2026.
Luster, M. S., & Pitts, B. J. Investigating the Effects of Driver-Vehicle Availability in Human-Autonomous Vehicle Teaming (HAVT). Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 69(1), 320-323, 2025.