A self-adaptive learning sliding mode control method for multi-degree-of-freedom magnetic levitation planar motor

Abstract

The invention proposes a multi-degree-of-freedom magnetic levitation planar motor self-adaptive learning sliding mode control method. The present invention uses the dynamic decoupling method to transform the multi-freedom control model into mutually independent single-degree-of-freedom models, and establishes a general system parameter model including uncertain items and external disturbances for the single-degree-of-freedom models; according to the model, an adaptive sliding mode is constructed The controller model suppresses external disturbances and iteratively learns compensation terms for uncertain terms. The iterative learning compensation item and the sliding mode control item are combined in parallel to obtain an adaptive iterative learning sliding mode controller; the stability analysis and error convergence analysis of the control algorithm are carried out through the Lyapunov theory; the control algorithm is applied to the actual magnetic levitation Planar motor system to verify the effectiveness of the control algorithm. The control method of the invention solves the problem of tracking control under external interference and uncertain items in the magnetic levitation planar motor system, and the algorithm has strong robustness, good self-adaptive ability and high tracking precision.

Description

technical field [0001] The invention belongs to the field of magnetic levitation planar motor control, in particular to a multi-degree-of-freedom magnetic levitation planar motor self-adaptive learning sliding mode control method. [0002] technical background [0003] Magnetic levitation planar motor, as a new type of drive element, has been extensively researched and developed in the past few decades. The magnetic levitation planar motor does not require mechanical guide rail support, and can directly realize a two-dimensional planar drive with a large stroke, which greatly simplifies the mechanical movement structure, and is small in size and light in weight, and can realize high-speed movement. Additionally, precision motion can be achieved under vacuum since no mechanical or air bearing support is required. These advantages make it have broad application prospects in semiconductor lithography systems and other high-precision industrial fields. [0004] Adaptive sliding...

AI Technical Summary

Problems solved by technology

[0004] Adaptive sliding mode control has good robustness and adaptive ability to system parameters, but it only suppresses the uncertain items related to the state in the system through simple robust items, resulting in its tracking effect being too conservative and stable. large state error

[0005] The iterative learning algorithm is a data-driven technology that does not rely on an accurate system model. It can optimize the control input in the current iteration according to the tracking error information in the previous iteration cycle, thereby improving the control accuracy of the system. However, iterative learning It belongs to feed-forward compensation control, which cannot guarantee the stability of the system. At the same time, iterative learning control has certain limitations in dealing with non-repetitive errors of the system, which affects the transient tracking effect of the system.

Images