A Modified Recursive Probabilistic Velocity Obstacles Model for Multi-agent Navigation

An Modified Recursive Probabilistic Velocity Obstacles approach for real-time multi-agent navigation is proposed in this paper. It is an extension of the Recursive Probabilistic Velocity Obstacles, which takes into account the perception and motion goals of other objects. The theoretical proof of the effectiveness is carried out, which is validated in several virtual collision avoidance experiments. The experimental results fully show that the proposed method can guarantee collision-free and oscillation-free motions for each of the agents.