SRL-ORCA: A Socially Aware Multi-Agent Mapless Navigation Algorithm in Complex Dynamic Scenes

摘要

For real-world navigation, it is important to endow robots with the capabilities to navigate safely and efficiently in a complex environment with both dynamic and static obstacles. However, achieving path-finding in non-convex complex environments without maps as well as enabling multiple robots to follow social rules for obstacle avoidance remain challenging problems. In this letter, we propose a socially aware mapless navigation algorithm, namely Safe Reinforcement Learning-Optimal Reciprocal Collision Avoidance (SRL-ORCA). This is a multi-agent safe reinforcement learning algorithm by using ORCA as external knowledge to provide safety guarantees. This algorithm further introduces traffic norms of human society to improve social comfort and achieve cooperative avoidance by following human social customs. The result of experiments shows that SRL-ORCA learns strategies to obey specific traffic rules. Compared to RL, SRL-ORCA shows a significant improvement in navigation success rate in different complex scenarios. SRL-ORCA is able to cope with non-convex obstacle environments without falling into local minima and has a 14.5% improvement in average time to goal compared to ORCA.