Research On Multi-robot Autonomous Mapping Based On Rapid-exploration Random Tree

Efficient robot navigation requires obtaining maps in advance.In order to obtain the map automatically,the robot needs to explore the unknown environment with the minimum cost and time under the premise of ensuring the map is completely covered.Aiming at the problems of slow detection speed,incomplete perception of environmental information,and poor system robustness of single robot system,the rapid exploration random tree algorithm is applied to the rapid detection of multi-robot system boundary points.The improved autonomous map of multirobot systems based on market mechanism is used to build a task allocation algorithm,assign target points to each robot,which effectively coordinates the behavior between the robots,and improve the efficiency of map construction.The main research contents are as follows:(1)Based on the Robot Operating System,through the improvement of Hassan's rrt_exploration platform,a multi-robot_exploration_Map based on rapid exploration of random tree,multi-robot autonomous mapping platform is established,to improve the task allocation module of multi robot autonomous map building and map fusion method.The Voronio polygons are used to divide the environment,and the rapid exploration random tree algorithm is used to explore all Voronoi polygons one by one,and the boundary points of the known and the unknown area are detected,which lays the foundation for the task allocation method of the next multi-robot autonomous map construction.(2)Using the obtained boundary division between the explored and the unexplored areas,a task allocation method is designed for autonomous mapping of multi-robot system based on the market mechanism method.Simulation results show that the proposed method can allocate boundary points reasonably,and complete the construction of the whole map in a reasonable time.Compared with traditional algorithms,this method can complete the construction of the map with fewer paths and can avoid Repeated exploration of the same area by robots to a certain extent.So as to reduces the cost of map construction and improves the efficiency of multi-robot map construction.(3)After the robot completed the exploration mapping task,it is necessary to fuse the local map created by each robot into a global map.In view of the problem of transforming the reference coordinate system of each robot in the map fusion problem,based on the map fusion algorithm proposed by Horner,the initial position and direction of each robot are calculated by using the multi-sensor data of the robot,and the global coordinate system of multiple robots is established.At the same time,based on the transformation method,the overlapping area between local maps is found,and a set of transformation and similarity heuristic algorithms are used to fuse the local map into a global map.