Evolving multiple sensory-motor controllers based on cellular neural network

There has been extensive work done to construct an optimal neural network for controlling a mobile robot by evolutionary approaches such as genetic algorithms, genetic programming, and so on. However, evolutionary approaches have difficulty in obtaining a controller that conducts complex and general behaviors. In order to overcome this shortcoming, we propose a method of combining several evolved modules by a rule-based approach. The multi-module integration method can make complex and general behaviors by combining several modules that are evolved or programmed to perform simple behaviors. Experimental results show the potential of the multi-module integration method as a sophisticated technique to make an evolved neural network carry out complex and general behaviors.