As anyone who has walked through a busy office building, retail store, or transit hub knows, navigating through large crowds of people moving in all directions is a complex procedure whose success depends on understanding the explicit and implicit rules that govern the dynamics of large groups and the subtle social cues that people use to predict each others’ future actions. Despite rapid progress in robotics fueled by artificial intelligence, robot navigation through pedestrian crowds is still an open problem that has the potential to transform society once it is solved. This project aims to create an embodied artificial intelligence, fueled by deep learning and informed by psychology, that will enable robots to effectively operate in a wide variety of new settings, including healthcare, retail, and logistics. The interdisciplinary team of mechanical engineers, computer scientists, and psychologists propose a bold, unique, and novel perspective that focuses on complex human factor­driven challenges. More specifically, the project will use methods from psychology to design experiments to carefully study human reactions to robotic technology and group dynamics, use machine learning to quantitatively characterize these dynamics, and synthesize these two perspectives to develop embodied artificial intelligence systems to enable seamless motion for robots.