The research content of this article is aimed at the intelligent manufacturing workshop of new energy vehicle batteries. Regarding the connection of production processes in the workshop, that is, AGV can achieve path planning and obstacle avoidance strategies in the face of dynamic and complex scenes during material handling, firstly, based on the production layout in the workshop, a neighborhood weighted grid modeling method is used to reduce the path offset error caused by the large speed difference between adjacent grid units in the grid model, and construct a more reasonable grid map. Then, on the basis of the grid map, the deep Q learning algorithm is used to achieve AGV path planning. The traditional algorithm introduces the method of updating the data in the experience pool and improving the direction reward function to solve the dimensional disaster and low learning efficiency of the deep Q learning algorithm in unknown environments. The, Due to issues such as poor robustness of planning strategies, a reasonable AGV path was obtained through simulation experiments, while improving the calculation speed and accuracy of the algorithm.
本站僅提供期刊文獻檢索。
