The Application of Fuzzy Neural Networks in Formation Control for Multi-Robot System

The formation problem of multi-robot is a type problem in cooperation and it is one of challenging research directions of multi-robot system. Aiming at the formation control problem, in this paper we propose a kind of hybrid architecture firstly, which combines the reactive architecture based on Motor Schema with the hierarchical architecture. Secondly we design four kinds of basic behaviors based on the reactive architecture, and design each of behaviors using fuzzy neural networks. Lastly we use nerve cell fuse the outputs of each of behaviors qua the final outputs of system, which controls executive appliance to act. This method put forward enhances intelligence and the environmental applicability in this paper. The simulation results validate the feasibility of this method.