Passivity-based discrete visual motion observer taking account of camera frame rates

In this paper, we investigate a vision-based 3D rigid body motion estimation problem. In one of our previous works, we have tackled the problem under the situation that the visual measurement of a camera is assumed to be continuously available. On the other hand, this paper proposes a new estimation mechanism, called discrete visual motion observer, explicitly taking account of camera frame rates. We first show that the present observer works successfully for a static target object. We then give estimation performance analysis for a moving target by employing the notion of ultimate boundedness. The present convergence and performance analysis gives a guideline to get the maximum allowable observer gain. Moreover, the effectiveness of the present estimation mechanism is demonstrated through simulation and experiments.