In recent years,the demand for electronic products represented by smart phones has been increased.Consumers have correspondingly increased the appearance requirements of the products.Therefore,the demand of the metal frame and the outline of the glass processing will become more and more accordingly,which has spawned a huge market.For example,in the processing and manufacturing process of the mobile phone metal frame and the outline of the glass,the same batch of workpieces are often repeatedly processed in a long time.The design of original intention of the current processing equipment and machine is mostly to solve different types of complex parts processing problems,without considering the same workpiece duplication of processing problems.Most of the current processing equipment and machine will regard mass production as a simply repeat of the last processing and thus a lot of previous processing data is wasted.Therefore,how to make the processing equipment use the previous processing data in the repeated processing to guide the next processing to improve the processing quality and yield of the production becomes a worthy topic.In view of the above problems,this thesis firstly analyzes the model of servo system,and uses the model of the system to make a detailed theoretical analysis of the first-order and high-order parameter optimization iterative learning algorithms.The algorithms are applied to single-axis position controller in the simulations.The simulation results show that the algorithms can effectively improve the convergence rate and tracking accuracy of tracking error.Contour error is a more important measure of the quality of the process.The reduction of the uniaxial tracki ng error does not necessarily reduce the contour error.Therefore,this thesis studies and improves the cross-coupling iterative learning algorithm.We combine the natural estimation method with the cross-coupled iterative learning algorithm to propose a cross-coupled iterative learning algorithm based on natural estimation with higher control accuracy,which makes the contour error decrease gradually in iterative process.The cross-coupling iterative learning algorithm has some limitations on the ratio of the error of one axis to the radius of curvature.However,the parameters of uniaxial control parameters are not ideal in the actual machining process,which results in the increase of tracking error.In the condition of large curvature,this condition is not satisfied,which limits the application of the algorithm.Therefore,this thesis improves the algorithm to solve this problem.The single-axis controller incorporates parameter optimization iterative learning algorithm to speed up the convergence of tracking error.At the same time,two adjustment parameters are added to balance single-axis control and contour control,thereby expanding the scope of the algorithm.Through the simulation in MATLAB,the effectiveness of the proposed algorithms is verified.Then we take the XY platform as an experimental object to verify that the parameter optimization iterative learning algorithm can effectively improve the tracking error convergence speed and the cross-coupling iterative learning algorithm can reduce the contour error during the iterative process. |