A design scheme of neural oscillatory networks by hierarchical evolutionary calculation for generation of humanoid biped walking patterns

Abstract

Humanoid biped walking patterns attracts more and more attention both in the fields of robotics and prosthetics. This paper presents a new scheme for the design of a neural oscillatory network which can be used to generate the desired walking pattern trajectories. The scheme is based on the evolutionary computation concept, and has the advantages of simplicity, flexibility, and high calculation speed. The scheme proved to be effective from the results of simulations. This research also established the foundation for research on expressing various human walking patterns in mathematical models.