Research On Multi-mobile Robot Scheduling System

Multi-robot scheduling technology is increasingly used in the current intelligent manufacturing production and warehousing and logistics industries.Based on the Key Project of Zhejiang Provincial Science and Technology Department,this paper decomposes the scheduling problem of multi-mobile robots in discrete manufacturing workshops into two problems of Multi-robot task allocation(MRTA)and Multi-Agent Path Finding(MAPF),aiming to improve the scheduling system’s work efficiency and coordination capability of the scheduling system.The main work of this paper is as follows:(1)Combining the scheduling characteristics of multi-mobile robots in a discrete manufacturing plant,a hybrid control system is used to control the scheduling system,and the raster map method is used to model the environment and determine the MRTA problem type of ID[ST-SR-TA].(2)Establish the MRTA problem model,construct the cost function,and propose the Negotiation and Task Reordering Auction algorithm(NTRA)based on the negotiation and task reordering mechanism for the problem that the single-task auction algorithm tends to ignore the new tasks generated near the robot,and the algorithm is improved as follows:(i)improve the objective function,and propose the optimization objectives of minimizing the robot load balancing index and minimizing the total duration of task transportation to evaluate the operation effect of the algorithm;(ii)optimize the task "bidding" process,and propose a "virtual auctioneer-robot" negotiation mechanism to determine the task release order based on the real-time distance between the task and the robot.(iii)Optimize the task "bidding" process,and propose the robot task queue reordering mechanism to optimize the task execution order and improve the robot load performance.The experimental results compared with similar algorithms show that the NTRA algorithm has the smoothest estimated path and performs better than other algorithms in terms of task execution cost,completion time and load balancing performance;when the ratio of robots to tasks is less than 25%,the load balancing index of the improved algorithm is s Tablewithin 2.69-4.21,which is conducive to the balanced and s Tableoperation of the system.(3)The MAPF problem model is established,the path conflict types are defined,and the enhanced two-way A* conflict search algorithm is proposed to address the problems of "blind" search of the conflict search algorithm(CBS)and poor quality of feasible solutions of the enhanced conflict search algorithm(ECBS),as follows:(i)Improve the low-level search algorithm of ECBS,and adopt The improved bidirectional A* algorithm is used for low-level path search to improve the path search accuracy on the basis of maintaining the path search efficiency;(ii)the second-order Bessel curve is used to optimize the path turning point to avoid the robot’s abrupt change of position and enhance the system working stability.The experimental results show that compared with the ECBS algorithm,the path cost of the improved algorithm decreases about 8.50% and the planning time increases about 5.95%,which has certain optimization effect.(4)Build the server software of the multi-mobile robot scheduling system and the hardware platform of the mobile robot,and design experiments to verify the overall scheduling scheme integrating the improved task assignment algorithm and the path planning algorithm.The experimental results show that the improved scheme is feasible,and it can allocate tasks to mobile robots in a balanced and efficient manner,improve the quality of feasible path solutions without affecting the efficiency of path planning,and achieve orderly collision avoidance among mobile robots.