A Class Of Multi-robot System Task Allocation Method

With the continuous expansion of robot's applications, the tasks become increasingly diverse, and the environment becomes increasingly complex. As a result, a single robot can'l meet people's needs in many cases. Compared to a single robot, a multi-robot system has many advantages. It is able to complete more complex tasks. So the study of the robot has been turned to the multi-robot system gradually.Multi-robot task allocation is a critical problem of the multi-robot system. The quality and efficiency of the task completion can be improved through the study of multi-robot task allocation. In this paper, the study of the multi-robot system is based on the background of the battlefield assault, and different attack missions should be assigned to the different robots according to a certain method.First, a system model based on the actual multi-robot system is established. In theory, multi-robot task allocation is actually a multi-objective optimization problem. Some optimization functions and constraints should be considered. Task allocation is separated into pre-allocation and re-allocation according to the dynamic characteristic of the system.In the stage of the task pre-allocation, an improved multi-objective evolutionary algorithm is used. There are several parts in the calculation, such as the initialization of the population, non-dominated sorting, selection, crossover, mutation operations and so on. Further more, the practical constraints and the method of aggregation distance calculation are researched. Through the simulation, it can be found that the improved multi-objective evolutionary algorithm can effectively solve the problem of the multi-robot task pre-allocation, and the results have the good diversity.In the stage of task re-allocation, a quantum particle swarm optimization is studied. Then, the diversity of the population can be improved by the chaos theory and the mutation operator. The simulation results show that this QPSO algorithm has the good real-time to solve the problem of the multi-robot task re-allocation.At last, a multi-robot task allocation simulation system is designed. A large number of the simulation results verify the validity of the model and the algorithms. It can realize the multi-robot task allocation.