Evolutionary Acquisition for Moving Performance of Reduced D.O.F's Quadruped Robot

Abstract

In this paper, an application of genetic algorithm for evolutionary generation of initial poses of a quadruped robot reduced degrees of freedom is described. Each leg of the robot has a slider-crank mechanism to reduce the degrees of freedom and is driven by an actuator. To generate the suitable initial pose, the initial angles of four legs are coded by real number as gene and tuned by genetic algorithm under the estimation functions for forward movement and right and left rotating movement. As a result of tuning, the adequate phases shift among the legs are generated. The experimental results demonstrate that the proposed scheme is effective for generation of the suitable initial poses to achieve the successful forward and rotating movements and the robot can walk smoothly for omni-directional movement with the generated initial poses.