Kalman filtering in position control using a vision sensor

In recent years owing to the development of computer hardware and image processing technology, the researches to build control system using a vision sensor are stimulated. Vision sensors provide additional information about the environment of the workplace as well as afford position information of an object. However, a vision sensor is sensitive to surrounding environment such as intensity, diffusion and reflection of light and has measurement noise due to changing conditions of light. The presence of measurement noise of vision sensor is an obstacle factor to precise control systems. For these reasons, the estimation theory is used to control accurately with reducing the impact of noise. It is possible to control a system accurately by using the estimated data. In this paper as measurement noise of vision sensor exists in the control system, Kalman filter is used to overcome the effect of noise and to more accurately control the position of the object using the estimated data. Experimental results show that the Kalman filter reduces the influence of measurement noise and improves the performance of position control of an object.