Research Of Task Cooperation For Multi-Robot Based On The Agent Technology

With the development of robot technology, the capability and the application field of robots are becoming greater and greater. On the one hand, when a complex task is too sophisticated to complete by a single robot, the cooperation and coordination of more than two robots are needed to be adopted; on the other hand, people hope to enhance the working efficiency through the cooperation and coordination of the multi-robot. In this paper, agent theory and its application in robot system are analyzed. On the one hand, agent theory can be applied in the design of robot. The design starts with the open and free angle, and processes at the systematical point of all the multi-robot system, to ensure the multi-robot system open. On the other hand, multi-agent can be applied in the design of the communication of multi-robot system. All the communications can be designed with Agent cooperation mechanism, at the unite theory frame, so as to ensure it consistent and natural.In the second place, to perform tasks, robots must form coalition when a single robot cannot perform tasks and tasks cannot be partitioned, or the partition is computationally too complex. Coalition formation in MAS (Multi-Agent System) is addressed to solve robots coalition formation in Muti-Robots. The paper describes two type robot formation problems (single task coalition and sequence tasks coalition structure), and gives them corresponding mathematic models. Just like other optimization problems, coalition formation need search the optimal solution or quasi-optimal solution in a complex and huge space. If algorithm traverses the whole search space, search process will lead to combination explosion that make it impossible for algorithm to complete search in polynomial time. We use Genetic Algorithm (GA) to realize agents' coalition formation. GA can reduce the cost of searching the optimal solution or quasi-optimal solution, and enhance the efficiency of the whole system.