Open-source software

Human–robot interactions recorded at Clemson’s Zucker Family Graduate Education Center in May 2022

Physics-based balance recovery and contact interactions in dense crowds

Composing control policies for long-horizon basketball moves

Adapting physics-based character controllers to new tasks and environments

Composite motion learning for synthesizing coordinated body-part motions in physically simulated characters

Interactive character control through physics-based imitation learning

Particle-based exploration for learning physics-based character controllers

Vehicle trajectory prediction using social and environmental context

Human trajectory prediction with social attention and timewise latent variables

Distributed coordination for agents navigating shared environments

Collision-free crowd simulation using optimization-based implicit integration

Human-inspired multi-agent navigation through knowledge distillation

Anticipatory collision avoidance for robots with general motion dynamics

Human-like collision avoidance based on pedestrian interaction data