AGV path planning method based on deep reinforcement learning

The invention discloses an AGV path planning method based on deep reinforcement learning. The AGV path planning method comprises the following steps: initializing an estimation network, a target network, an environment and an experience pool; starting path planning; the agent interacts with the environment, and at is obtained under the St; st, at, rt and St + 1 are stored in a playback pool, an award rt is obtained, the state is updated, St + 1 is obtained, and St, at, rt and St + 1 are placed in the playback pool; randomly extracting a group of St, at, rt and St + 1 from the playback pool; optimizing a loss function; updating the valuation network Q; setting the current valuation network as a new target network; and ending the path planning. By adopting the above technical scheme and the AGV path planning method based on deep reinforcement learning, the AGV system can search for a better path in a self-optimization manner for a complex and unknown environment, the working efficiency of the AGV system is furt