The Research On Multi-Robot SLAM In The Complicated Unknown Environment

With the development of robot technology,the robot has gradually entered our production and life.In an unknown environment,the smart robots need to obtain their own position,attitude and environmental information for autonomous navigation,formation and cooperative missions.With the environmental information,the robots can realize obstacle avoidance,path planning and formation.As one of the robotic key technologies,the simultaneous localization and mapping is still an important content in the field of robots.This paper is based on the theory of multi robot,combined with the related technologies of multi robot SLAM to study and realize the multi robot system for mapping fast and high quality map.The main research results are as follows:(1)Combined with the current experimental environment and robots,a robot pose model and environment model of feature extraction were built.To prevent the absence of environmental data,an environment data addition strategy of path planning was designed.The feature points of the local map were analyzed and processed,a triangulation graph of the local map can be obtained.The path points were chose for the robot to collecting invisible region data.The objective and starting points has been dynamic programming to get the shortest path,reduce the energy loss.In the Matlab the experimental simulation verified the correctness of the model.(2)In the indoor environment,the similar structure of the indoor will lead to error matching.This paper combined the Keenan-Motley with the clustering algorithm,and constructed the map correlation function.According to the value of the function to determine the sequence of the map fusion,to prevent duplication of the map and splicing errors.The experiments showed that the algorithm can realize the fusion of the sub map,build a global map,and improve the accuracy of the map,the fusion effect of the map was expected.