Vision guided homing for humanoid service robot

Human beings can easily use their eyes to guide arms and hands to reach any particular pose. We have been trying to make robot to imitate such intelligent hand-eye coordination behavior. In this paper we report our research work on vision guided homing for humanoid service robot HARO-1 which has active stereo vision and two modular arms. The work is based on the 3D vision reconstruction and geometric kinematics analysis. This intelligent behavior of HARO-1 significantly improves its performance by automatically rectifying its original home position and implementing self-homing guidance by stereo vision system. The implementation and experimental studies are also presented.