Research On The Method Of Multi-Robot Collaborative Mapping

Environment map is the key for robot to realize autonomous positioning and navigation and complete complex intelligent tasks.The construction and application of environment map has become one of the core problems in the field of robot research,and the quality of the map directly affects the efficiency of the robot.At present,the mapping technology of single robot is relatively mature.However,a single robot cannot build a global map quickly and accurately when facing a complex,large-scale working environment or tasks with high efficiency requirements.Motivated by this challenge,using multi-robot to quickly collaborate to build an accurate environment map has become an important and critical topic in multi-robot research.Based on the above analysis,the specific research content of this topic is as follows:(1)Study the robot simultaneous localization and mapping(SLAM)algorithm based on graph optimization.According to the front-end constraint information,an optimization objective function is constructed to optimize the sub-map matching relationship,and then the optimized information is used to create a map.The efficiency of building a point cloud map by a single robot is verified by the self-built robot experimental platform.The multi-robot cooperative architecture is studied,and the distributed architecture is used to improve the flexibility,reliability and robustness of the multi-robot system.At the same time,a communication mechanism based on ROS system and Wi-Fi is used to realize the data transmission in the distributed multi-robot system,which ensures the stability and real-time performance of data exchange between multi-robot,and meets the real-time requirements of data transmission in multi-robot cooperative mapping.(2)Research the multi-robot overlapping region search algorithm based on improved Scan-context descriptor.In order to improve the efficiency of the overlapping region search algorithm,an improved Scan-context descriptor is used,which pays more attention to the geometric characteristics of the horizontal and vertical distribution of the point cloud in each circular sector region.At the same time,in order to improve the computational efficiency of the optimal solution problem and improve the success rate of obtaining the global optimal solution,the split iterative closest point algorithm is used to solve the optimal solution problem,which is divided into multiple optimal solution problems according to the distribution of the point cloud.Finally,the experimental results show that the multi-robot search operation time is controlled at about 0.4s,which has higher operation efficiency.At the same time,the mean square error of the trajectory in the absolute error evaluation is about0.5m,indicating that the pose estimation process has higher accuracy.(3)The global optimization algorithm of multi-robot cooperative mapping is studied.In order to improve the optimization accuracy,a more accurate global map is constructed.The constraint conditions of relative pose estimation of multi-robot are introduced into the graph optimization objective function,and the quasi-Newton method is used to solve the optimization objective function,which effectively reduces the amount of computation and ensures the overall high convergence speed of the algorithm.The experimental results show that the distance error between plane and elevation in the global point cloud map constructed by multi-robot cooperation is about 0.05 m,the plane accuracy is 98%,and the elevation accuracy is 96.23%.It shows that the multi-robot can efficiently and accurately construct the global point cloud map,which verifies the feasibility of the multi-robot cooperative mapping scheme.This thesis contains 63 figures,11 tables and 94 references.