Research On Control System Of Autonomous Mobile Robot

SLAM includes two meanings:localization and mapping.In this paper,the robot motion model is designed firstly,and the map model is determined to be topological diagram and raster diagram.Then EKF and Gmapping algorithms based on particle filter are analyzed and verified by simulation respectively.The analysis results show that EKF-SLAM has larger limitations and higher algorithm complexity,but it has a guiding significance for other SLAM algorithms.Gmapping algorithm is based on particle filtering algorithm,and the error of simulation results is small,so field experiments are needed to verify the operation of the specific algorithm.Then,the Hector-SLAM algorithm was analyzed and optimized for its shortcomings.The bicubic interpolation method was used to improve data continuity,and the data of IMU and odometer were fused.Finally,the location was carried out based on AMCL algorithm.The overall optimization results reduce the dependence of traditional HECTSLAM on the performance of the lidar,and can solve the boundary tilt and other problems,and improve the positioning accuracy.Path planning algorithm is the core algorithm of autonomous mobile robot operation,which is mainly divided into two parts: single robot path planning algorithm and multi-robot path planning algorithm.The single-robot path planning algorithm is improved on the basis of the pass-through A* algorithm to reduce the number of turns and path consumption.In this project,it is mainly applied to SLAM navigation.The multi-robot path planning algorithm improves the A* algorithm based on the principle of time window,and sets the Bundle area in the map,which needs to apply and release resources.Finally,the multi-robot simultaneous task can run normally,and greatly reduces the occurrence of conflicts and deadlocks.Finally,the ground control system is designed and developed,which is mainly responsible for the creation of tasks and the monitoring of the robot state.In the main view,the real-time position of the robot can be observed.Firstly,the mapping accuracy of Gmapping algorithm is analyzed through experiments.There are some problems of boundary deviation and unclear in the field mapping experiment of experimental data.Then the experimental data of Hector-SLAM show that the accuracy of the improved SLAM algorithm is significantly improved compared with the original algorithm,and the positioning accuracy can be significantly improved.Finally,the project decided to adopt the improved HECTOR-SLAM as the mapping algorithm of the project.The result of SLAM mapping is used as the background diagram of the operation interface of the scheduling system,so that the real-time position of the robot can be seen more clearly.The test proves that the robot runs normally under the condition of single robot running.In the multi-robot simulation operation environment,the multi-robots run smoothly without collision,conflict and deadlock.