Multi-AGV Path Planning Algorithm And Task Allocation Scheduling Strategy

With the proposal of "Made in China 2025",traditional manufacturing factories are gradually transforming into automation and intelligence,using automated equipment to replace humans to complete production and transportation.As an important part of the smart factory,the Automated Guided Vehicle(AGV)can efficiently complete the task of cargo transportation and has broad application prospects and practical value.Based on the application of AGV-related technologies in manufacturing factories,this paper studies the key issues of multi-AGV path planning and task allocation scheduling,including optimizing single-source path planning,proposing multi-AGV collision-free planning strategies,proposing multi-task assignment strategies,and realizing optimization Targets,etc.,and establish a factory map model,AGV body kinematics model,develop a simulation system.The following is a brief summary of the research content of this paper:(1)For the single-source path planning problem,an improved A~* algorithm based on the heuristic function is proposed,which uses the AGV state combined with the fuzzy logic rules as the heuristic information,and dynamically adjusts the weight of the heuristic function during the search process,so that the path expands faster toward the target.At the same time,it hardly falls into the local optimal solution.Compared with the classical algorithm,the shortest path is also solved,but the algorithm in this paper consumes less time and space and improves the search efficiency.Combining the advantages of the grid method and the topology method,a gridtopology double-layer map model is established,and the factory map is modeled;Mecanum wheeled mobile AGV is used,and kinematics analysis is performed.(2)For the collision conflict problem in multi-AGV motion,a collision-free path planning strategy based on the Petri net model is proposed.First,the common intersection conflicts are analyzed and classified according to their characteristics,and the Petri net is used to simulate the conflict situation.Constraints are established for actual needs,and various conflict resolution models are proposed.The solution model is used as a path planning algorithm for a collision-free path planning strategy for local re-path planning.After planning,conflict avoidance is realized,and the solution is verified in the reachability analysis.correctness of the model.(3)For the multi-AGV multi-task allocation problem,a multi-AGV multi-task allocation scheduling optimization strategy based on the improved particle swarm genetic algorithm is proposed.Firstly,the constraint equation is established for this problem,and the optimal target model of task assignment is established based on the particle swarm algorithm,the improved particle swarm genetic algorithm is used to optimize its optimization space,which speeds up the convergence speed of the objective function,and can find the approximate optimal solution faster than the classical algorithm,and the optimization result is also better.Based on this algorithm,a periodic task allocation strategy is further proposed.Considering factors such as priority,existing path planning,and cost minimization,tasks are added to the queue to wait,which also achieves the optimal total cost.(4)Design a simulation system to simulate the real factory map,order tasks and vehicle configuration,and test the multi-AGV path planning,obstacle avoidance planning,task allocation scheduling and other functions.In the experiment,the path planning algorithm can plan a reasonable plan according to the starting point and destination The obstacle avoidance planning algorithm can re-plan a collision-free route in the conflict area according to the original route,and the task assignment scheduling algorithm can assign multiple tasks to the AGV under the condition of satisfying the constraints,and minimize the total cost.The results are in line with expectations,which verifies the feasibility of the algorithm and system.