An Evolved Neural Controller for Bipdedal Walking with Dynamic Balance

TL;DR

This paper presents an evolved neural network controller enabling a simulated bipedal robot with compliant actuators to walk dynamically, introducing novel evolutionary methods and genetic representations to improve control optimization.

Contribution

It introduces a new evolutionary approach with multiple demes and increasing difficulty fitness functions, along with a novel genetic encoding for neural controllers.

Findings

Successful evolution of dynamic walking in simulation

Effective use of multiple demes and increasing difficulty fitness functions

Novel genetic representation improves controller evolution

Abstract

We successfully evolved a neural network controller that produces dynamic walking in a simulated bipedal robot with compliant actuators, a difficult control problem. The evolutionary evaluation uses a detailed software simulation of a physical robot. We describe: 1) a novel theoretical method to encourage populations to evolve "around" local optima, which employs multiple demes and fitness functions of progressively increasing difficulty, and 2) the novel genetic representation of the neural controller.