Observer based fine motion control of autonomous visual alignment systems

In the flat panel display manufacturing, the visual alignment system functions as an important part to get high quality products. In this paper, an observer based control method is suggested to enhance the performance of the visual alignment control system, where the error compensation trajectory is to be changed according to intermittent updates of the observer output which tracks the alignment marks in mask and panel. It can be considered a visual servo control with coarse-fine control strategy, where the fine but slow vision loop includes a coarse vision loop but with a faster sampling rate and the Kalman filter is to estimate the centroids of alignment marks in the inner coarse vision loop. As a result, this enables a vision-based fine motion control comparing with any standard method of using single vision loop. In constructing a Kalman filtering algorithm, the reduced coarse images from the original fine ones are used as the measurements and the moving equation of alignment mark is obtained by solving the forward kinematics of the parallel alignment stage. Finally, the performance of the proposed control is verified through experimental results.