The co-evolution of model parameters and control programs in evolutionary robotics

Evolutionary robotics is a research area that makes use of the various forms of evolutionary computation to provide a means of designing robot control systems. In this paper, we introduce a new way of integrating the actual robot and its model during evolutionary computation. This method, which involves the co-evolution of model parameters, is applied to the problem of learning gaits for hexapod robots. The form of evolutionary computation used is the cyclic genetic algorithm (CGA), which was introduced in previous work (Parker et al. (1996)) to deal with the issue of evolving controllers for cyclic behaviors. Tests done in simulation show that the CGA operating on the co-evolving model of the robot can adapt to changes in the robot's capabilities to provide a system of any-time learning.