Omni-Directional Vision-Based Control Strategy for Humanoid Soccer Robot

Abstract

The humanoid soccer robot system can be considered as a visual feedback system that includes the design and implementation of the hardware architecture, visual servo system, control strategy system and sensory system. This paper is mainly to integrate the two major parts of the humanoid soccer robot, the visual system and control strategy system. At first, the architecture of the omni-directional vision system is setup. The advanced label algorithm (ALA) is then employed for object recognition in the vision processing system. The corresponding offending and defending control strategies are developed. Two special data transformation mechanisms are also examined. Finally the experiments of the humanoid soccer robot are performed to verify the benefit and the feasibility of the proposed control strategies.