Research On Multi-robot Collaborative SLAM Algorithm In Indoor Environment

In recent years,multi-robot collaborative SLAM(simultaneous localization and mapping)technology has become one of the research hotspots in the field of robotics.Collaboration between multiple robots can greatly improve work efficiency.In many cases,the robot works in an unknown environment,such as clearing obstacles,navigating,searching,etc.in the fire scene,but before performing these tasks,it is necessary to obtain an environmental map.The traditional indoor multi-robot SLAM algorithm has many shortcomings: it is not flexible enough in the assignment of exploration tasks,which often causes problems such as low exploration coverage and high exploration path repetition rate;in terms of map fusion,it is greatly affected by environmental factors and the mapping accuracy is not high..Therefore,this article will carry out in-depth research from exploration task allocation and map fusion to improve the efficiency and robustness of collaborative mapping of multi-robot indoor environments.The work done is as follows:First of all,to address the problems of islands(ie,small areas that have not been explored),low exploration efficiency,and non-optimal target selection in the auction collaboration algorithm,improvements have been made.Active state variables are added to the auction algorithm.When the robot explores unknown areas or other robots during the mapping process,it performs corresponding behaviors by switching its own active state,which solves the problem of islands in the auction algorithm and low exploration efficiency.Problem: On this basis,combined with the virtual potential field algorithm,calculate the resultant force of each task point on the robot,and select the task point with the highest resultant force as the exploration target,which solves the problem of non-optimal robot target selection.Through MATLAB simulation experiments,it is verified that the improved algorithm has a certain improvement in the efficiency of collaborative mapping and exploration coverage.Secondly,based on the Gmapping algorithm,a hybrid map fusion algorithm is studied according to whether the robots meet during the mapping process.When the robots meet during the exploration process,the relative observation model of the system is constructed,and the relative transformation matrix between the robot coordinate systems is calculated to realize map fusion.If the two robots do not meet,but the built map contains part of the same area,the map fusion problem is transformed into a problem of finding the optimal solution,and the PSO(particle swarm algorithm)is used to find the conversion matrix that maximizes the overlap of the two maps.Realize map integration.The two algorithms cooperate with each other to integrate the sub-maps to realize the collaborative mapping of multiple robots.Through the single-robot SLAM experiment and the dual-robot map fusion experiment,the effectiveness of the algorithm is verified.Finally,under the ROS system,the algorithm of this paper is verified experimentally based on the Turtlebot3 robot platform.Completed the multi-robot collaborative SLAM experiment in the real environment.Through the comparative analysis of the experimental results,the improved algorithm has a certain improvement in saving time and cost and exploring coverage,and can coordinate the accurate mapping of multi-robots in the indoor environment.