带参数辨识的 H-∞鲁棒控制方法

H ∞ robust control approach with parameter identification

从控制方法的实用性出发，综合鲁棒控制和自适应控制方法，处理存在各种非结构不确定性及参数不确定性的线性对象。运用状态空间 Ｈ∞方法设计鲁棒控制器，抑制外部干扰，未建模动态及参数误差。进而引入鲁棒辨识器以确定对象的标称参数，减小鲁棒控制的保守性。并证明了闭环系统的稳定性和鲁棒性。结合 Ｒｏｈｒｓ 反例，验证了算法有效性。该方法实现了自适应律与控制方法的分离，从而避免了传统‘确定性等价原理’带来的系统分析和实际应用的复杂性。

This paper tackles linear systems with both unstructured uncertainties and parametric uncertainties combining robust control and adaptive control methods in order to make advanced control theory more practical. State space H ∞ approach is employed to design a robust controller so as to deal with external disturbances, unmodelled dynamics and parameter errors. A robust identification approach is also presented to obtain nominal parameters and reduce the conservativeness of conventional robust control methods. The stability and robustness of the closed system are proved. An illustrative example based on Rohrs' contradiction shows the effectiveness of the proposed method. This approach separates the adaptive law and the control approach such that it avoids traditional ‘certainty equivalence principle’ which makes system analysis and practical application complicated.