In this thesis, we developed a stereo vision system using CCD cameras to coordinate the motions of two arms of a humanoid robot with 27 degree of freedom(d.o.f). To do this, a closed-form solution of the inverse kinematics for the end-point of two arms with 12 d.o.f. was obtained and coordinates of the specific object were computed for the vision system. Using mapping coordinates computed by the relation between the vision system and the end-point of the robot arm, a visual-servoing system for the humanoid robot was developed.
In addition to this, a 3D graphical simulator was developed using the physical software-engine to simulate the two-arm motion coordination system before performing experiments for the system.
Some experiments to verify the motion coordination of the visual servoing system were performed. Through the experiments, a successful result that the gripper of the arm catches the object was achieved.