多无人车协同规划及突发情况下应急响应与恢复方法

In the face of unexpected situations such as sudden changes in the environment and vehicle failures, traditional centralized collaborative planning methods are unable to perform rapid emergency response, system conflict resolution and orderly recovery. Aiming at the above problems, a cooperative path planning and emergency response and recovery method for multiple unmanned vehicles under kinematic constraints is proposed. Firstly, a collaborative path planning model under kinematic constraints is constructed using a centralized search architecture, and a body constraint tree is introduced to eliminate multi-unmanned vehicle collisions in continuous space. Secondly, in order to ensure that the unmanned vehicles affected by the emergency situation can stop safely and quickly and reduce the interference to the system, a method for selecting the starting point of emergency planning based on the warning area is proposed, and a random search tree is generated based on the state of the unmanned vehicles at the time of the emergency event to perform a hierarchical search, which combines the random search tree, the internal space-time constraints of the system, and the demand for stability in the movement of the unmanned vehicles to ensure that the unmanned vehicles can smoothly transit to the emergency stopping point. Finally, during the system recovery process, for the time difference of different unmanned vehicles arriving at the emergency response stopping point, an asynchronous planning strategy is adopted to incorporate the vehicle trajectories not affected by the emergency event into the spatio-temporal constraints of collaborative planning, and finally generate the recovery trajectory clusters after the emergency response. In the simulation test, the average duration of generating the optimal emergency response trajectory is 0.324 s. The emergency response module improves the planning success rate from 28.7% to 87.6%, and the asynchronous module increases the operating efficiency by nearly 93% when there are 15 sudden obstacles.