Evolution and Development of Modular Control Architectures for 1D Locomotion in Six-legged Animats

Abstract

An evolutionar y approach is used to design neural control architectures for virtual sixlegged anim ats.Using a geom etry-or iented variation of the cellular encoding scheme and syntactic constraints that reduce the size of the genetic search space, the developmental prog rams of straight locom otion controllers are rst evolved.One such controller is then included as the rst module in a larger architecture, in which a second neural module is evolved and develops connections to the rst one, so as to set locom otion on or o according to sustained or instantaneous exter nal control signals.Such an incremental approach should prove useful to the autom atic design of relatively complex control architectures that might, in particular, implement some cognitive abilities over and above mere stimulus response mechanisms.