Design And Implementation Of Autonomous Navigation For Wheeled Robot Based On ROS

With the development of science and technology,mobile robots are widely used in various fields,bringing great convenience to people’s production and life.How to make the robot construct an accurate map and realize autonomous navigation in complex indoor environments has become a research hotspot.This thesis takes the wheeled differential robot as the research object,based on the Robot Operating System(ROS),the two key technologies of Simultaneous Localization And Mapping(SLAM)and path planning in robot autonomous navigation are studied,the main research results of this thesis include:1)Complete the hardware selection and platform construction of the wheeled differential robot,complete the communication between the upper computer,the lower computer and the sensor.The kinematics model,odometer model and sensor observation model of the robot platform are analyzed.2)To solve the problem of inaccurate positioning when using the wheel odometer alone,the extended Kalman filter algorithm is used to fuse the data of the wheel odometer and IMU,and the accuracy of the fused data is verified by the mapping effect of the Gmapping algorithm.The mapping effects of the Gmapping algorithm and the Cartographer algorithm are compared,and the advantages and disadvantages of the two algorithms are expounded.Aiming at the problem of incomplete environmental information in the map constructed by single-line lidar,the data of lidar and depth camera are fused from two aspects.At the map level,Bayesian fusion is performed on the two-dimensional grid maps constructed by the lidar and the depth camera respectively.At the laser level,the laser information and the pseudo-laser information converted from the depth image are fused,threshold processing is added,and the fusion effect is adjusted through the threshold,and the fused data is used for map construction.After experimental comparison,the laser-level fusion scheme proposed in this thesis constructs maps with higher accuracy and enables to obtain richer environmental information.3)Research on path planning algorithms,aiming at the low search efficiency and unsmooth path of the global path planning A* algorithm,the actual cost function and the heuristic function of the A* algorithm are dynamically weighted to improve the search efficiency,and the path is smoothed using the Bezier curve,the search efficiency of the improved A* algorithm is increased by 14.4%,and the path length is reduced by 3.9%.On the basis of global path planning,this thesis designs a PID chassis control algorithm,which can realize the autonomous navigation task of the robot,the success rate of the autonomous navigation of the robot is 76%.The improved global path planning A*algorithm is fused with DWA and TEB local path planning algorithms respectively to conduct autonomous navigation and obstacle avoidance experiments on the robot,the navigation success rates of the two schemes are 88% and 92% respectively,and the robot can avoid unknown obstacles in the navigation process.This thesis verifies the feasibility of the robot platform design scheme through experiments,the experimental results show that the robot can better complete the tasks of positioning,mapping,autonomous navigation and obstacle avoidance.The fusion algorithm and improvement ideas proposed in this thesis have certain application value.