Research On Positioning And Navigation Of Indoor Lidar Mobile Robot

With the development of robot technology and artificial intelligence,robot plays an increasingly important role in many fields.With the emergence of Baidu unmanned car and sweeping robot in recent years,robot has already entered our life.The core technology of robot technology includes three aspects: positioning,mapping and path planning.In this paper,particle filter positioning is selected as the positioning method;gmapping mapping is used as the robot's mapping technology;A* algorithm is used as the robot's path planning algorithm.At the same time,an improved scheme is proposed to solve the problems of redundant points and unsmooth path planning in the path planning process of A* algorithm Experiment.Firstly,the laser radar,which is the sensor used by the robot,is introduced.The principle of the laser radar is analyzed.The model of odometer and robot motion is built for the future experiment.Secondly,the particle filter positioning is studied and analyzed.Through Monte Carlo sampling,sequential importance sampling and importance resampling,the estimation value of effective particles is obtained and the weight is normalized,so as to obtain the estimation of system position and determine the current state of the robot.Because the path planning needs to use grid map,the gmapping algorithm is used to build the map According to the information of sensors,the environment information around the robot is determined and a grid map is built.This paper introduces the principle of A* algorithm and analyzes it.In view of the unnecessary exploration nodes in the path planning of A* algorithm,it puts forward that the distance between the former node and the target point added to the current exploration node as a part of the evaluation formula,so that the A* algorithm can judge the nodes with the same evaluation value which are relatively close to the end point,so as to reduce the calculation resources.In addition,for A* algorithm A* algorithm based on Bezier curve is proposed to smooth the inflexion of the original algorithm.Finally,the paper introduces the ROS and double differential wheel robot platform,and uses the gmapping algorithm to build the map.Firstly,it uses the gazebo platform in ROS to build the map simulation experiment,and then uses the robot platform to build the map experiment for the laboratory and the experimental building environment,and takes the map as the environment map for path planning.The simulation experiment of A* algorithm and improved A* algorithm is carried out in Matlab platform The experimental results show that the improved algorithm can reduce the number of exploration nodes by about 25% for 10 * 10 grid map.The experimental results show that the improved A* algorithm can reduce the number of nodes by about 30% for the experimental building and on the double differential wheel robot The A* algorithm of the line is simulated in Matlab platform,and the path of different turning points is planned by the A* algorithm.The experiment shows that the algorithm can use Bezier curve to smooth the turning point of the path,so as to get a smooth path without turning point.Compared with the original path,the path length can be reduced by 20%-30%.