Cooperative object tracking by high-speed binocular head

In this study, we propose a novel cooperative control of a robot head and binocular eyes for tracking a moving object. Humans turn their heads and eyes toward a static object with a partial responsibility ratio modeled by a sigmoid function related to the angle of the head to the object. Also, when a human tracks a slowly moving object inside his field of view, he does not move his head. That is, how a human moves his head to track an object depends on the position and motion of the object. Similarly, we propose a new cooperative method of reducing head motion by estimating the target motion. This method reduces the head movement while tracking error is about the same as conventional methods. In this paper, we describe the proposed method and its experimental results.