Evolving Gaits for the Lynxmotion Hexapod II Robot

Gait generation for hexapod robots can be accomplis hed with a two-phase process, whereby leg cycles that e ffectively provide thrust are learned and then the prop e selection and coordination of these cycles needed t o produce a gait are learned. Genetic algorithms can be used to learn both the leg cycles and the coordination. In previous work, this technique was used successfully to generate gaits for the ServoBot. In this paper, we ap ply the techniques that were successful with the ServoBot t o the Lynxmotion Hexapod II Robot. Although physical dif ferences between the two robots required changes to the robot simulations, genetic algorithms and the two-p hase successfully learned effective gaits.