Research on inspection path planning method based on DQN algorithm
YANG Yaojun
ZHANG Qianjin
WANG Wei
Abstract：The traditional DQN algorithm has a slow network convergence speed,and it is difficult to obtain the optimal path in a short training round when it is applied to path planning research.In this paper,an improved method of DQN algorithm was proposed.Firstly,the state information required for path planning of inspection robot was analyzed and determined,and a reasonable reward function was designed,so that the mobile robot can obtain the shortest inspection path without collision through algorithm learning.Secondly,the DQN algorithm was improved to reduce the training times of mobile robots and shorten the machine learning time.The experimental verification was carried out by building a simulation environment.The experimental results showed that the improved algorithm can complete the inspection path planning task in the complex obstacle map,and significantly improve the learning efficiency.
Keywords：inspection robotreinforcement learningpath planning
Publication Date：2024-11-30
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：8( 55-62 )