Safe Path Planning in Unmanned Warehousing Scenarios

In order to solve the safety problem during vehicle traveling in the context of unmanned warehousing, this paper proposes an improved RRT algorithm based on deep reinforcement learning (SAC-RRT). In this paper, a safety constraint module is added to the algorithm, and the deep reinforcement learning SAC algorithm is used to guide the direction of RRT expansion and dynamically adjust the sampling step size through the obstacle area ratio to improve the efficiency of vehicle transport. Experiments have proved that with this algorithm, the unmanned forklift can dynamically adjust its path, avoid areas such as cargo falling, and quickly arrive at the designated location while ensuring safety.