A DQN-based AGV global path planning method and system

Abstract

The invention relates to a DQN-based AGV global path planning method and system, belonging to the technical field of path planning, and solves the problems of long time consumption and inaccurate path planning structure in the existing path planning method. The method includes: constructing a simulated environment of a two-dimensional grid map with a grid side length p based on the perceived storage environment information, wherein the storage environment information includes obstacles, the starting point of the AGV, the current position of the AGV, and the end position of the AGV ; Expand the action set that AGV takes in the next step in the two-dimensional grid map; set a continuous and reasonable reward function to distinguish the reward value obtained by AGV at each grid in the two-dimensional grid map; and design the optimal path discrimination condition And according to the optimal path discrimination condition, after the AGV runs the specified cycle, the optimal path is output. Combined with the increased AGV movement direction, modifying the optimal path discrimination conditions can effectively improve the accuracy of path planning. The efficiency of path planning can be improved by setting a continuous and reasonable reward function.

Description

technical field [0001] The invention relates to the technical field of path planning, in particular to a DQN-based AGV global path planning method and system. Background technique [0002] With the development of intelligent warehousing technology, daily tasks such as lifting, transporting, unpacking, and packing of goods in the warehouse can be completed by AGV (mobile robot), which can improve the efficiency of access to goods in the warehouse. Reasonable path planning can improve the working efficiency of AGV and reduce the failure rate of AGV. Existing path planning algorithms mainly include: Dijkstra algorithm and A* algorithm based on traditional graph theory; bionic algorithms such as ant colony, particle swarm, and genetics. The above algorithm has defects such as easy to fall into local optimum, and it is difficult to deal with complex storage environment. [0003] In the known storage environment, using DQN (that is, deep Q network) for global path planning is a ...

AI Technical Summary

Problems solved by technology

[0004] In view of the above analysis, the embodiment of the present invention aims to provide a DQN-based AGV global path planning method and system to solve the problems that the existing path planning method takes a long time and the path planning structure is inaccurate

Images