Map Construction And Path Planning Of Mobile Robots Based On Multi-sensor Fusion

In recent years,mobile robots have been widely used in all walks of life,and the scope of activities is also expanding.Autonomous navigation in outdoor environments is an indispensable basic function of mobile robots.This paper studies the key technologies of autonomous navigation of mobile robots: multi-sensor fusion,map construction and path planning.The main research contents are as follows:Since GPS is a global positioning,there is no drift error,and the visual odometry of the stereo camera has accurate local positioning and long-term drift,and the two have complementary characteristics.Therefore,this paper uses the stereo camera and GPS as the main sensors to build a pose graph optimization model,and obtain a more accurate global optimized pose of the mobile robot through multi-sensor fusion positioning.The experimental results show that the global optimized pose reduces the drift error of the visual odometry,and improves the positioning accuracy of the mobile robot.Aiming at the problem that the sparse point cloud map generated by the ORB-SLAM2 algorithm cannot be used for path planning,a global map construction technology based on feature points is proposed.The global optimized pose is used as the mobile robot localization,and the matched feature points extracted by the stereo camera are converted to the global coordinate system,and then the two are uniformly converted to the grid coordinate system,and the ray is sent from the grid coordinates of the mobile robot to the grid coordinates where the feature points are located,and the corresponding access counter and occupancy counter are added.The state value of the grid is set according to the calculated idle probability,and a two-dimensional occupancy grid map is incrementally established as the mobile robot moves.Experiments verify the effectiveness of the constructed grid map.Aiming at the problem that the traditional D* algorithm does not consider the shape and size of the mobile robot and regards it as a mass point,this paper adopts the method of setting up an obstacle expansion layer on the occupied grid map to keep the mobile robot away from obstacles when driving;for the problem that the path planned by the traditional D* algorithm does not conform to the kinematic characteristics of the mobile robot,the uniform B-spline is used to smooth the planned path and detect obstacles on the path to ensure that the smooth path can avoid obstacles.Planning experiments verify that the proposed path planning algorithm can establish a safe and smooth path from the starting point to the target point,which improves the safety and fluency of the mobile robot in actual driving.