The Swirling Brain points out a NewScientist.com article about a new control system for walking robots. Yasuo Kuniyoshi and Shinsuke Suzuki of the Intelligent Systems Lab at Tokyo University have created an efficient walking algorithm for a 12-legged robot based on chaos. Their method results in emergent walking behaviours from random control inputs much faster than traditional evolutionary computing methods based on genetic algorithm and seems to mirror what happens in baby animals who are able to learn to walk very quickly.