Evolving bipedal locomotion with genetic programming - a preliminary report

Abstract

Shows how genetic programming can be applied to the task of evolving the neural oscillators that produce the coordinated movements of human-like bipedal locomotion. In biomechanical engineering research, robotics and neurophysiology, it is of major interest to clarify the mechanism of human bipedal walking. This serves as the basis for developing several applications, such as rehabilitation tools and humanoid robots. Nevertheless, because of the complexity of the neuronal system that interacts with the body dynamics system to make walking movements, much is left unknown about the details of the locomotion mechanism. Researchers have previously been looking for the optimal model of the neuronal system by trial and error. In this paper, we apply genetic programming to induce the model of the nervous system automatically and show its effectiveness by simulating a human bipedal gait with the obtained model. Our experimental results are preliminary but they show some promising evidence for further improvements.