In this paper, the phase lead by the general phase-lead compensator is analysed, and a criterion of its parameters selection is proposed. In order to get the relatively large time-interval prediction for stochastic signal, two types of differentiator-predictors are presented to estimate future signal and its derivatives: linear high-gain differentiator-predictor and nonlinear differentiator-predictor. Furthermore, a nonlinear extended differentiator-predictor is designed for future estimation and chattering rejection. The stability analysis of differentiator-predictors is described in time domain. The iterative transfer function method is proposed to analyse the robustness in frequency domain, and the rules of differentiator-predictor parameters selection are presented. The analysis shows that the nonlinear differentiator-predictor has stronger adaptability and robustness than the linear high-gain differentiator-predictor. Simulations demonstrate the effectiveness of the proposed methods.

本文对一般相位超前补偿器的相位超前进行了分析，提出了其参数选择的准则。为了对随机信号进行较大的时间间隔预测，提出了两类微分预测器来估计未来信号及其导数：线性高增益微分预测器和非线性微分预测器。此外，还设计了一个非线性扩展微分预测器，用于未来的估计和抖动抑制。在时域中描述了微分预测器的稳定性分析。提出了迭代传递函数的方法来分析频域的鲁棒性，并给出了微分-预测参数的选​​择规则。分析表明，非线性微分预测器比线性高增益微分预测器具有更强的适应性和鲁棒性。仿真证明了所提出方法的有效性。