Socially Aware Safe Navigation for Heterogeneous Multi-Agent Systems

Risk-aware navigation in uncertain, crowded environments with heterogeneous agents.

Research focus Social navigationRisk-adaptive safetyHeterogeneous agents

This project develops socially aware navigation methods that adapt safety and motion decisions to uncertain interactions among robots, pedestrians, and other moving agents. This work is conducted in collaboration with Toyota Research Institute of North America (TRINA) - AMRD.

Reinforcement Learning for Risk Adaptation via Differentiable CVaR Barrier Functions

Visualization of adaptive risk and safety decisions

Risk-adaptive demo

Real-world robot risk-adaptive navigation using differentiable CVaR barrier functions

Real-world robot demo

Safe Navigation in Uncertain Crowded Environments Using Risk Adaptive CVaR Barrier Functions (IROS 2025)