Direct Adaptive Fuzzy Sliding-mode Control Of Visual Ball And Plate System Based On Ant Colony Optimization

Ball and plate system based on visual servoing is an under-actuated, strong coupling and nonlinear system. The position of the ball is gathered by an over-head camera, and gives feedback to the controller to control the plate to let the ball track the expected trajectory. Therefore, the research on the ball and plate system is mainly divided into two parts:image processing and motion control. The good image processing algorithms can get accurate position coordinates of the ball, and the good motion controller can make the ball track the trajectory in real time and accurately, so the system can be used to verify the validity of image and control algorithms. The following work is done based on the two parts.Firstly, the ball and plate system has light interference which makes the system can’t gather the ball position accurately, so a Random Hough Transform circle detection algorithm is designed based on contour tracing to eliminate the "holes" phenomenon caused by reflection, and it also improves the speed of Random Hough Transform circle detection. In addition, in order to make the image processing adaptive to the light intensity change, a histogram equalization algorithm based on fuzzy logic inference is designed.Then, a mathematical model for ball and plate system based on plate’s angle is designed and analyzed. To deal with the uncertainty of the model and the circumstance, a sliding mode controller is proposed. However, the sliding mode control algorithm has "chattering" phenomenon, and a way which combines the direct adaptive fuzzy control algorithm with sliding mode control algorithm is designed to soften the control signal. In order to minimize the control signal, tracking errors and the sliding surface, the ant colony algorithm is used to optimize the factors of sliding surface and fuzzy adaptive laws. Simulation results show that: the algorithm based on ant colony optimization improves the dynamic performance of the control system effectively.Finally, the hardware architecture of the ball and plate system is introduced, and physical control is done by using the image processing and control algorithm proposed in this paper. In conclusion, the algorithm proposed in this paper is effective and efficient.