Evolution of adaptive synapses: Robots with fast adaptive behavior in new environments

This paper is concerned with adaptation capabilities of evolved neural cont rollers. We propose to evolve mechanisms for parameter self-organization in stead of evolving the parameters themselves. The method consists of encodin g a set of local adaptation rules that synapses follow while the robot free ly moves in the environment. In the experiments presented here, the perform ance of the robot is measured in environments that are different in signifi cant ways from those used during evolution. The results show that evolution ary adaptive controllers solve the task much faster and better than evoluti onary standard fixed-weight controllers, that the method scales up well to large architectures, and that evolutionary adaptive controllers can adapt t o environmental changes that involve new sensory characteristics (including transfer from simulation to reality and across different robotic platforms ) and new spatial relationships.