一类受输入和参数扰动时滞系统预估及动态更新控制策略的研究

Delays widely exist in various systems and systems with delays exhibit complex dynamic behavior, and in particular, there are many challenges in controlling such systems subject to perturbations. Considering a class of delayed systems with input and parameter perturbations, this research project focuses on three aspects: 1) With the assumption a disturbance input being norm-bounded, the robust stability of delayed systems is analyzed. In order to speed up the disturbance attenuation, an observer-based proportional-integral memory controller is designed which satisfies the prescribed H∞ performance. In tracking control, the problems of error convergence and offset elimination are also studied. 2) A non-fragile H∞ feedback controller and observer-based asynchronous adaptive controller are designed to guarantee the robust stability of the delayed systems with randomly occurring parameter uncertainties. 3) For an unknown nonlinear input perturbation, appropriate neural network-based online prediction compensation mechanism and fuzzy logic adaptive dynamic control tuning law are developed. In this project, some intelligent control approaches are innovatively proposed to deal with the perturbed delayed systems. In the presence of complex perturbations, the dynamic characteristics of the considered systems are revealed. Due to the broad practical background of perturbed systems with delays, the research results of this project will have considerable value in applications.

时滞广泛存在于各类系统中，时滞系统具有非常复杂的动力学特征，受扰时滞系统的控制问题面临诸多挑战。本项目针对一类受输入和参数扰动的时滞系统，主要对三个方面的内容进行研究：1）在范数有界输入扰动条件下，分析时滞系统的鲁棒稳定性，设计满足H∞性能指标的基于观测器的比例积分记忆控制器，提高时滞系统的抗扰能力，并解决在跟踪控制时误差收敛和偏差消除的问题。2）设计非脆弱H∞反馈控制器和基于观测器的异步自适应控制器，确保参数随机扰动下时滞系统的鲁棒稳定性。3）考虑未知非线性输入扰动，采用神经网络在线预估补偿，建立模糊逻辑自适应动态调控准则。本项目力图从新角度提出针对受扰时滞系统的智能控制方法，揭示复杂扰动条件下时滞系统的动力学性质。由于受扰时滞系统具有广泛的实际背景，本项目的研究成果具有较高的应用价值。

在大量的实际系统中，时滞和扰动现象广泛存在。本项目针对时滞系统中的扰动和非线性动力学特性展开研究，并对混杂系统的抗扰性进行分析和鲁棒控制设计。主要研究内容和重要结果包括：1）针对一类受输入和参数扰动的时滞系统，创新地提出了基于等比数列方法的时滞分割技术，并设计满足H∞性能指标的比例积分记忆控制器和观测器，该设计可有效的提高控制系统的抗扰性能。2）以非线性系统为研究对象，通过神经网络和T-S模糊建立系统数学模型，研究系统的状态收敛和鲁棒控制问题，解决了在抗扰控制时误差收敛和偏差消除的问题。3）以混杂系统为研究对象，设计新型脉冲比例加积分控制策略从而降低干扰和未知不确定性对系统性能的影响。设计了具有异步和记忆特性的新型混合比例积分控制算法，有效改进了系统的抗扰性。该控制策略在切换系统、奇异系统的镇定性和抗扰性方面表现出了优良特性。4）以具有混杂特性的正系统为研究对象，考虑时滞、非线性扰动、不确定性以及多模态子系统切换特性，研究了系统的L1-gain控制问题，设计了具有异步和记忆特性的混合比例积分控制器。我们尝试将理论和实践相结合。目前已经通过仿真软件将设计的控制方案成功应用于降压变换电路系统中。