Evolution of Corridor Following Behavior in a Noisy World

Robust behavioral control programs for a simulated 2d vehicle can be constructed by artificial evolution. Corridor following serves here as an example of a behavior to be obtained through evolution. A controller's fitness is judged by its ability to steer its vehicle along a collision free path through a simple corridor environment. The controller's inputs are noisy range sensors and its output is a noisy steering mechanism. Evolution determines the quantity and placement of sensors. Noise in fitness tests discourages brittle strategies and leads to the evolution of robust, noise-tolerant controllers. Genetic Programming is used to model evolution, the controllers are represented as deterministic computer programs. 1 Introduction Designing reactive controllers for autonomous agents can be a challenging task. For increasingly complex behavior, building controllers by hand becomes prohibitively difficult. As suggested in [Cliff 1993b], a promising alternative is to evolve the controllers..