Iterative Tuning of Notch Filter for Motion Control with TLBO Algorithm

In the realm of ultra-precision motion control, notch filters are commonly used to mitigate the resonance peaks inherent in the motion system, which is pivotal in diminishing residual vibration, curtailing settling time, and enhancing servo accuracy. Ideally, a notch filter should embody the inverse model of the resonant peak. However, the conventional design of notch filters requires an advanced meticulous identification of the motion system's resonance peak, a process that is both time-consuming and arduous. Inspired by the iterative tuning of feedforward control parameters, this paper introduces a preliminary notch filter parameter tuning method based on the teaching-learning-based optimization (TLBO) algorithm. This method leverages intelligent optimization algorithms to finetune the numerator and denominator parameters of the transfer function, obviating the need for the linear-in-the-parameters model required in traditional iterative tuning. By utilizing an iterative learning mechanism, the proposed approach more effectively satisfies the persistent excitation condition and dynamically adjusts the optimization range of the TLBO algorithm according to prior experimental data, thus attaining superior tuning accuracy. The efficacy of the proposed method has been comprehensively validated through simulation.