Research And Implementation Of SLAM Algorithm For Mobile Robot On ROS System

With the extensive application of mobile robots in daily life,the autonomous technology of mobile robots widely attracted attention in the field of scientific research.At the same time,SLAM(Simultaneous Location and Mapping)is key technologies for mobile robot to locate and navigate in unfamiliar environment.It has important theoretical research value and application value.Based on mobile robot motion model and observation model,the EKF-SLAM,UKF-SLAM,IEKF-SLAM algorithm are studied and Simulated in this paper.As reported,the extended Kalman Filter(EKF)based SLAM has one serious drawback which is the linear approximation of non-linear functions.Hence,this paper presents an improved EKF-SLAM solution,based on conditional iterations with detection of loop closing(iEKF-SLAM).Combining an effective map management strategy,when the loop is closed,the condition to perform iterated update is satisfied and then the iterative update procedure employed in the iterated extended Kalman Filter(IEKF)is implemented to update the states of the robot and the map for achieving the maximum a posterior probability estimation.This approach improves the accuracy and consistency of the estimate on the basis of a smaller increase in computational complexity.By using the public simulation software package and Vitoria Park data sets and comparing the experimental results of the several SLAM algorithms,that the iterative update strategy is adopted when the path of a mobile robot forms a closed loop can greatly improve the estimation accuracy of the robot,and the algorithm proposed in this paper is superior.In order to test the algorithms in SLAM,a mobile robot platform based on the Robot Operating System is built in this paper.With ROS as the development platform,the communication framework of the built robot platform is set up,and the base control scheme is planned.We debug and improve the mobile platform with the ROS tools,build the map and location navigation experiment with the mobile platform,and analyze the communication between the nodes of the mobile platform.In order to advance the work,making the built mobile platform with multisensor complementarity,the frontier technology of visual SLAM is analyzed.Two popular algorithms(ORB-SLAM2 and Direct Sparse Odometry)of visual SLAM are studied in this paper.We implements and analyzes two kinds of visual SLAM algorithms under ROS environment,and implements the handheld kinect2 mapping under the laboratory environment based on DSO algorithm.