Integrated Adaptive Repetitive Learning Control of Linear Motor Servo Systems With Periodic Tasks

Periodic repetitive tasks are common in industrial production, especially in intelligent equipment for batch processing. This paper studies the disturbance suppression problem of linear motor servo system (LMSS) in periodic trajectory tracking, and proposes an integrated adaptive repetitive learning control (IARLC) algorithm. Switching dynamics involving linear and nonlinear exponentials are constructed to achieve finite-time convergence or boundedness of the tracking error. For periodic uncertainty, a feedforward compensation term that does not depend on model prior information is designed, and a fully saturated learning law is used to limit the estimation error to a preset interval. Specifically, parameter estimation based on co-resonance reset and a projection-type adaptive algorithm are used to achieve compensation of physical model parameters. Non-periodic uncertainties and unknown disturbances are optimally attenuated by a robust feedback control law. The stability of the closed-loop system is analyzed based on the Lyapunov theory, and comparative experimental results verify the advantages of the proposed method in the presence of parametric and non-parametric uncertainties. Note to Practitioners—Modern industrial production usually uses linear motors as power components to benefit from their high precision and conversion efficiency. In addition, batch assembly and processing require manufacturing equipment to perform repetitive tasks over and over again. Based on this background, this paper studies the periodic trajectory tracking problem of LMSS and proposes a new finite time IARLC method. The controller is designed based on model compensation and robust stabilization, and the periodic dynamic characteristics are effectively processed using historical information, which reduces the complexity of system design. The legacy non-periodic disturbances and parameter time-varying effects are weakened by the designed feedback control law and parameter estimation algorithm. In particular, the linear and nonlinear error switching surface is constructed to accelerate the stability of the system. In the presence of parameter and non-parameter uncertainties, the stability and precise tracking of the closed-loop system are achieved, which has practical significance for practical engineering applications.