Design And Implementation Of Robot Navigation System

In this era of increasing development of science and technology,people's lives are more and more abundant,and various intelligent devices are more and more popular.As the crystallization of human technology and wisdom,robots are widely used in various industrial and service industries,providing great help for human production and life.In recent years,intelligent community has attracted more and more attention.People want to improve the quality of life through more intelligent technology.In the interior,smart home is the goal pursued by many enterprises,while outdoors,smart security,intelligent logistics and so on are constantly innovating because of the development of software and hardware technology.Outdoor mobile robots,as an important part of them,have great prospects in community security patrol,community logistics and other aspects.The navigation system is the basis for realizing the upper layer functions of mobile robots.Without a robust navigation system,other functions cannot be discussed,so it is of great value to the research of mobile robot navigation systems.Based on the project of laboratory security robot,this work designs and implements a mobile robot navigation system in community environment.The main work is as follows:Firstly,according to the actual needs of the project,the overall logical architecture of the navigation system is designed.Secondly,in view of the fact that the lidar may not detect the road edge,a road segmentation method based on OTSU using camera image is proposed,which can make the robot avoid the road edge better.Then,We propose a real-time navigation algorithm based on multi-sensor fusion.By using the data of GPS,electronic compass,lidar and camera reasonably,the local path planning is divided into several basic behaviors,such as error correction,obstacle avoidance,direction adaptation and emergency obstacle avoidance,so that the robot can complete navigation smoothly and smoothly.Finally,aiming at the actual situation of our mobile robot,we propose a model-based reinforcement learning local path planning algorithm,which includes state space,action space and reward mechanism.The training model is applied to navigation system.Experiments show that this method can improve the ability of navigation and has great practical value.