Research On Multi-Robot Task Assignment And Path Planning Algorithms In Indoor Enviroments

In recent years,scientific and technological advancements,as well as socioeconomic growth,have considerably aided in the development and use of indoor mobile robots,and various types of indoor mobile robots have made human life and work more convenient.Because indoor mobile robots are becoming more common,they must not only collaborate but also go quickly and safely to task locations in demanding interior conditions.As a result,in this paper,it is required to begin research on multi-robot task assignment and path planning algorithms for indoor situations.Task assignment,indoor localization,and path planning are critical components of indoor robotics,and research in these areas is growing in popularity.Nevertheless,existing multi-robot task assignment,indoor localization,and path planning algorithms have some flaws.The following are the major flaws:(1)When a large number of task demands are generated,the optimization-based task assignment method is slow in solving the MRTA problem and falls into local optimum,resulting in a poor solution;(2)indoor environments are difficult to satisfy LOS conditions,making precise indoor positioning with UWB only difficult;and(3)in indoor environments,some moveable areas of the robot will inevitably be congested and blocked.To tackle these issues,this study provides a cooperative job assignment strategy,a fused indoor localization system,and an improved sampling heuristic optimum path planning algorithm,as follows:(1)A cooperative task assignment strategy based on the K-Means++and the IEO algorithms is proposed in this paper.Because the EO algorithm can still be improved,this paper introduces circle chaos mapping,Levy flight,and the iterative cosine operator,resulting in the IEO algorithm.The collaborative task assignment strategy converts the MRTA problem to the MTSP,clusters the task objective points using the K-Means++algorithm,divides the MTSP into multiple TSPs,solves the multiple TSPs separately using the IEO algorithm,and then integrates the optimal solutions to obtain the MTSP’s optimal solution.(2)In this paper,an indoor positioning system based on IMU and UWB fusion is proposed.Due to the higher number of mistakes in UWB measurement data,a two-step calibration approach based on LS is proposed to estimate the error parameters in this paper.Second,in order to mitigate the effects of NLOS,in this paper,an IMU and UWB fusion localization algorithm based on EKF is proposed.(3)In this paper,a potential function-based sampling heuristic optimal path planning(CCPF-RRT*)algorithm considering congested regions is proposed.First,the CCPF-RRT*algorithm incorporates an artificial potential field into RRT*,which minimizes the total number of iterations and hence accelerates convergence.Second,a movement cost function that includes congestion severity and path length is constructed to analyze the link between two path nodes thoroughly.And the function is used to direct the expansion of new nodes and the updating of parent nodes so that the algorithm takes congestion intensity and path length into account when generating paths.Third,the algorithm combines the benefits of F-RRT*to reduce the cost of producing parent nodes for random nodes.Furthermore,the DWA algorithm is utilized to design the robot’s route in real time based on the global path planned along the CCPF-RRT*,enabling safe avoidance of unforeseen obstacles.Experiments were conducted on each of the algorithms described in this research to validate their usefulness for the specific study,and the experimental results are as follows:(1)The IEO algorithm is faster and more accurate than the EO algorithm,and the collaborative task assignment strategy solution is reasonable.(2)The two-step LS-based calibration method not only accurately estimates the error parameters,but the localization accuracy of the EKF-based IMU and UWB fusion localization algorithm is about 50%better than the WLS method,and the algorithm has good robustness,with an average localization error of about 0.05 m.The CCPFRRT*algorithm outperforms RRT*,Q-RRT*,PQ-RRT*,and F-RRT*in terms of initial solution and fast convergence speed.Furthermore,the CCPF-RRT*and DWA fusion algorithms can plan safe and efficient courses for robots in real-time in crowded regions with dynamic impediments.In conclusion,the approaches proposed in this paper are suitable for indoor mobile robots.