Speed-command-independent parameters self-tuning and mismatch compensation for servo speed control

Abstract

Usually the servo system using the traditional generalized predictive control is limited by the reference command, the expression of which is only a step signal or a ramp signal. Meanwhile, during the identification convergence, the model mismatch composed of the modeling errors and the nonlinear disturbance, often leads to the obvious speed spike and even damping oscillation. To solve the above problems, a speed-command-independent 2 degree-of-freedom (2DOF) N-type PI (NPI) controller based on an improved generalized predictive control (IGPC) is designed, as well as an extended sliding-mode compensator (ESMC) is proposed in this paper. First, to enhance the steady-state performance for tracking speed command, a 2DOF NPI speed controller is utilized to track the arbitrary-order polynomial command. Then, based on the forward N-order difference operator, an extended prediction model, the output of which only contains the speed error, is established to free the IGPC from the speed command. Next, the optimal control law is obtained by the rolling optimization of a reconstructed cost function, and further is simplified to match with the 2DOF NPI controller online. Finally, the ESMC is designed to effectively weaken the speed spike caused by the model mismatch, and a special positive definite function is utilized to ensure its stability. Simulation and experimental results verify the advantages of the proposed scheme.