Realization of robust controllers in evolutionary robotics: a dynamically-rearranging neural network approach

The evolutionary robotics approach has been attracting a lot of attention in the field of robotics and artificial life. In this approach, neural networks are widely used to construct controllers for autonomous mobile agents, since they intrinsically have generalization, noise-tolerant abilities and so on. However, there are still open questions: (1) the gap between simulated and real environments, (2) the evolutionary and learning phase are completely separated, and (3) the conflict between stability and evolvability/adaptability. In this paper, we try to overcome these problems by incorporating the concept of dynamic rearrangement function of biological neural networks with the use of neuromodulators. Simulation results show that the proposed approach is highly promising.