Research On The Navigation Method Of Mobile Robot In Unknown Environment

Industry is developing rapidly in recent years,robot,especially mobile robot application is more and more widely,its core navigation technology development is growing by leaps and bounds,the navigation algorithm based on vision sensors emerge in endlessly.In particular,the depth of the reinforcement learning to some extent to imitate the human progress through continuous learning and motivate yourself,in trying to learn in each state should take action.This feature has a lot to do with the robot's navigation behavior.The purpose of this research is to use this idea to study the navigation method of mobile robot in unknown environment,and to study the robot navigation method based on deep reinforcement learning.This method through the depth camera feature extraction with the environmental information of and make a decision,the realization of ultimate in unknown and complex environment,mobile robot through deep learning mechanism of reinforcement learning algorithm can independently complete the starting point to the target point navigation task,and it is concluded that the optimal path.In the course of the research,the robot simulation environment was set up,and the DDPG algorithm was trained,evaluated and compared with the traditional Slam navigation method in the test environment.Experimental results show that the navigation algorithm based on deep reinforcement learning avoids the tedious task of building a map,which is superior to the navigation algorithm based on Slam in both navigation accuracy and computation time.In this paper,deep reinforcement learning algorithm is applied to the robot navigation task,and the task of fast navigation is realized by mobile robot in the face of unknown environment.Compared with the traditional map-based navigation algorithm,it can meet the navigation requirements in different unfamiliar environments,which is more in line with the development direction of the future navigation algorithm.