Control characteristics of cooperative conveyance system for multiple mobile robots using complex-valued neural network

This paper describes control characteristics of the cooperative conveyance system for multiple mobile robots. In our previous research, we proposed the cooperative pattern generator using the real-valued neural network (RVNN) and the complex-valued neural network (CVNN) to generate the cooperative motion for a pair of robots. Connection weights of the networks are tuned by the genetic algorithm. In this paper, we confirmed the advantages of the CVNN from a comparison of learning results and experimental results using a pair of robots with the CVNN and the RVNN.