Evolving gaits for hexapod robots using cyclic genetic algorithms

Abstract

A major facet of multi-legged robot control is locomotion. Each leg must move in such a manner that it efficiently produces thrust and provides maximum support. The motion of all the legs must be coordinated so that they are working together to provide constant stability while propelling the robot forward. In this paper, we discuss the use of a cyclic genetic algorithm (CGA) to evolve control programs that produce gaits for actual hexapod robots. Tests done in simulation and verified on the actual robot show that the CGA successfully produces gaits for both fully capable and disabled robots.

Keywords:

• Evolutionary robotics
• Legged robots
• Gait generation
• Learning
• Cyclic control
• Evolutionary computation

Notes

Gary B. Parker is the Jean C. Tempel '65 Assistant Professor of Computer Science and the Director of the Computer Science Program at Connecticut College. He has a B.A. in Zoology from University of Washington, an M.S. in Computer Science from Naval Postgraduate School, and a Ph.D. in Computer Science and Cognitive Science from Indiana University.

His research focuses on evolutionary methodologies for learning in autonomous robots. In addition to developing the cyclic genetic algorithm, he developed punctuated anytime learning (PAL), a means of integrating the actual robot with its simulation during evolutionary computation, which allows the system to make adjustments to the on board controller while learning is carried out off line, giving the robot the capability to adapt to changes in real time. He also does research in colony robotics, co-evolving cooperative teams, emergent systems, and neural networks.