Visual SLAM And Path Planning For Automatic Road Sweepe

The application of autonomous driving technology to sweepers can not only effectively solve the problems of high labor intensity and high labor costs for sanitation operators,but also better improve the quality and efficiency of cleaning work.Visual Simultaneous Localization and Mapping(SLAM)and path planning are the key technologies for automatic sweepers.Visual SLAM can allow automatic sweepers to understand the surrounding environment information and achieve accurate positioning of vehicles;path planning allows automatic sweepers According to the cleaning task,the vehicle can complete the cleaning operation safely and efficiently.This paper studies the visual SLAM and path planning of automatic sweepers.(1)In order to solve the problem of large positioning error of automatic sweepers in low-texture scenes with lack of point features,an IPL-SLAM algorithm based on information entropy point and line features is proposed.Optimization to filter out features with more task-driven information.At the same time,a new parallel thread is created to extract line features.By merging potential homologous line segments and then introducing geometric constraints,the effective matching of line features is achieved,and a new point-line feature error model is constructed to optimize the pose.Experiments are carried out on the datasets,and the proposed IPL-SLAM algorithm has significant improvements in localization accuracy and robustness.(2)According to the operation characteristics of the automatic sweeper,combined with the cubic B-spline curve,the global welt path planning algorithm of the automatic sweeper was designed;combined with the depth information of the binocular camera,an obstacle detection method based on the SSD network was designed;Aiming at the problems of low search efficiency and slow convergence speed of the progressively optimal Rapidly Expanding Random Tree(RRT*)algorithm,the target bias expansion and ellipse sampling strategy are used to improve the RRT* algorithm,which is used as a local path planning algorithm for automatic sweepers.The simulation results show that the proposed planning algorithm is more effective.(3)The improved visual SLAM algorithm and path planning algorithm are tested through real vehicle experiments.The experiments show that the system accuracy has been greatly improved.In terms of time efficiency,it can meet the real-time requirements of the system;the automatic cleaning vehicle is realized.The global cleaning path tracking and local path planning prove that the improved algorithm proposed in this paper is feasible in practical scenarios.