Formation Control Strategy of Multi-agent System with Improved Probabilistic Roadmap Method Applied in Restricted Environment

In order to solve the obstacle avoidance adaptability problem of multi-agent swarm, a formation control strategy of dynamic barrier avoidance and static collision-free path was proposed. We introduce Probabilistic Roadmap Method (PRM) to solve the problems of track planning and formation strategy in restricted environment. First of all, a multi-agent formation model with fixed undirected communicative topology was established, the virtual leader was set up as the virtual control center of the swarm. Formation and trajectory planning were established around the virtual control center, then we verified the consistency. Secondly, the improved PRM sampling method was designed and optimized to ensure a reasonable sampling density in the restricted environment, so as to improve the quality of trajectory planning. Finally, through the dynamic-static collaborative barrier avoidance method, combined with the lack of dynamic artificial potential field barrier avoidance, the formation control strategy of dynamic adjustment was constructed. In order to solve the multiagent to avoid other agents and obstacles, the problem may occur that the established formation can not adapt to the surrounding environment, resulting in swarm deadlock or agent out of formation. The simulation results show that the multi-agent swarm can adjust the formation strategy in time and avoid obstacles effectively, and then pass through the waypoint to reach the target point.