Measuring the complexity of the real environment with evolutionary robot: evolution of a real mobile robot Khepera to have a minimal structure

A real mobile robot was evolved in several kinds of real environment with various complexities. The fitness function of GA operations included the complexity measure of the control structure, i.e., individuals with a simpler structure obtained a higher score. Evolution lead to developing a robot with a minimal structure sufficient to live and perform tasks in the given environment. The minimal structure itself lets one perceive easily the control scheme (or skeleton) of the robot for the given environment, and suggests a design scheme for more economical robots. Moreover, the value of the complexity measure of the control structure after evolution in a certain environment could be used as an index of complexity of the environment. The authors used a neural network with variable numbers of connections for a control structure of a real mobile robot Khepera. The network summed up signals from eight proximity sensors to generate outputs to two motors. The robot was evolved in four different kinds of environment with various complexities to perform the task of navigation with obstacle avoidance. The number of connections was used for the complexity measure of the control structure, which was included in the fitness function. After evolution, robots with a minimal number of connections for a given environment were indeed developed. The number of connections obtained was lower in a simpler environment, showing the feasibility to use the complexity measure as a complexity index of a given environment