Cyclic gait learning based on the Ant Colony Optimization

This paper focuses on a development of a feasible gait pattern that can be used to control hexapod robot locomotion. The main objective of the presented paper is to develop suboptimal gait pattern to control walking robot. We are considering the construction of a six-legged walking robot (Hexapod). Hexapod robot locomotion should be controlled via controller. Controller should be capable of producing in open loops coordinated walking patterns. Gait generation is an optimization problem with constrains which can change during the time. Usage of Swarm Intelligence methods to obtain feasible solution of this complex optimization task is presented in this paper. Ant Colony Optimization methods were found as an appropriate learning algorithms to accomplish this task. The control of a hexapod locomotion would be based on the model of artificial central pattern generators. This paper proposes to invent proper leg co-ordination control using a controller learned through the Ant Colony Optimization methods.