搬运机器人在无人仓的“货到人”拣选系统中得到了广泛应用，然而多个搬运机器人协同作业还存在着路径寻优规划不合理、路径冲突等问题。考虑路径的负载平衡及搬运机器人转弯耗费的时间，建立了带有时间窗的多搬运机器人路径规划模型；结合调整优先级的避障策略，提出了一种改进的“离线–在线”两阶段动态路径规划算法，实现全局离线路径规划和在线冲突规避，提升了无人仓系统的运行效率。最后，设计三组仿真实验，通过对比搬运机器人的总运行时间，验证了改进算法的有效性。
Handling robots have been widely used in the picking system of “part-to-picker” in un-manned warehouse. However, there are still problems such as unreasonable path seeking planning and path conflict in the collaborative operation of multi-handling robot. Considering the load balance of the path and the turning time of the handling robot, the path planning model of the multi-handling robot with time window is established. Combined with the obstacle avoidance strategy of adjusting the priority, an improved “offline-online” two-stage dynamic path planning algorithm was proposed to realize the combination of global off-line path planning and online conflict avoidance, thus improving the operation efficiency of the un-manned warehouse system. Finally, three sets of simulation experiments are designed to verify the effectiveness of the improved algorithm by comparing the total running time of the handling robots.