Research And Implementation Of Autonomous Exploration And Mapping System For Mobile Robot Based On Multi-sensor

As technology advances,mobile robots have been widely used in household service industry,security industry,warehousing logistics as well as other fields.With respect to conventional mobile robots,an a priori environment map needs to be artificially provided.In order to further enhance the intelligence and autonomy of mobile robots,this study adopts sensors including two-dimensional light detection and ranging(2D LIDAR),Kinect V2 depth camera,photoelectric encoder and inertial measurement unit(IMU)to design an autonomous exploration and mapping system of indoor mobile robots.The thesis mainly conducts a study on mobile robots in terms of SLAM(simultaneous localization and mapping)algorithm,autonomous exploration algorithm and path planning algorithm.In doing so,the SLAM algorithm is optimized;the autonomous exploration algorithm based on frontier is improved;the autonomous exploration and mapping of mobile robots in unknown environments is thus achieved.First of all,the thesis introduces the overall design scheme of the system and illustrates hardware selection,hardware connection and the like.In the second place,the data obtained through 2D LIDAR is fused with point cloud data calibrated by Kinect,which offsets the limitation that LIDAR sensors can only detect a single plane;the extended Kalman filter is used to fuse the data obtained through photoelectric encoder with the data collected by IMU,thus avoiding the errors caused by the robot wheel slipping and bumping so as to realize more accurate robot localization;the use of Cartographer mapping algorithm eliminates the accumulative errors of robot localization through closed-loop detection and graph-based optimization,therefore making the environment map built by mobile robots more accurate.Furthermore,the exploration algorithm based on frontier offers an effective strategy of autonomous exploration which enables robots to quickly find the best exploration target site in the known local map.Then,the improved A* algorithm and artificial potential field are used to complete the global path planning and local path planning,respectively.Finally,many experiments have been carried out in the laboratory,thus verifying the reasonableness and reliability of the autonomous exploration and mapping system of mobile robots based on multi-sensor design.