Research On Key Problems Of Multi-robot Uniform Deployment

With the rapid development of robot technology,multi-robot system is applied in various fields.Compared with single robot,multi-robot has great advantages in space search,environment monitoring and random environment network deployment.This paper studies the uniform deployment of boundary robots in any environment,mainly from the generation of target task points,global path planning,complete allocation of multi-target points and multi-robot local path planning.Firstly,according to the research objectives,a robot experimental system suitable for uniform deployment is constructed.The system includes the selection of the robot motor,the selection of the camera and the design of the upper main program.Secondly,aiming at the generation of robot target task points,namely region uniform segmentation,a centroid Voronoi region uniform segmentation method integrating pixel coordinates is proposed.The method can effectively solve the uniform region segmentation of irregular graphics by collecting pixel coordinates of color graphics as centroid Voronoi partition data set.Each subregion divided by centroid Voronoi is almost equal in size,and the centroid of each subregion is the target position of the robot.Experimental analysis shows the feasibility of this method for uniform segmentation of irregular regions.Thirdly,aiming at the problem of global path planning of robot in known environment,this paper adopts an improved A* path planning algorithm,which improves the search neighborhood and heuristic function of traditional A* algorithm.Experimental comparison with traditional A*algorithm shows that this algorithm is superior to traditional A* algorithm in both path length and time efficiency.Then,for the complete allocation problem of multi-robots and multi-objective points,that is,multiple robots are assigned to multiple target points one by one to make the total path shortest.In this paper,an improved simulated annealing algorithm is used to solve the complete assignment of target task points,and the solution generation and objective function of the traditional simulated annealing algorithm are improved.Experimental analysis shows that the algorithm improves the time efficiency and the total path length,and the furthest distance robot path becomes shorter.Finally,for the local path planning problem of multi-robot,the dynamic window method is adopted in this paper.The algorithm is combined with the global path planning algorithm to ensure that the robot is close to the global path in the actual movement process,and avoid the collision between robots.Experimental analysis shows the feasibility of the algorithm.