自校正控制的虚拟等价系统理论

Virtual equivalent system theory for self-tuning control

摘要:

用虚拟等价系统理论统一分析基于任意控制策略和任意参数估计算法的自校正控制系统的稳定性和收敛性．对于各种被控对象(确定、随机、最小相位、非最小相位),只要控制策略满足常规条件,即控制策略对参数估计得到的模型构成稳定闭环系统,则自校正控制系统当采用适当低阶模型,参数估计不收敛时也有可能稳定和收敛,参数估计的取代条件为有界、慢时变以及参数估计误差为相对无穷小量(相对于回归向量范数)．

Abstract:

Based on virtual equivalent system (VES) theory, this paper presents the viewpoint that the structure information of plant to be controlled and the convergence of parameter estimates are both unnecessary for the stability and convergence of various types of self-tuning control (STC) systems. A general self-tuning control system comprised of arbitrary control strategy, arbitrary parameter estimation algorithm, and arbitrary plant (deterministic or stochastic, minimum or non-minimum phase), is stable and convergent if only the control strategy stabilizes the parameter-known model; the parameter estimates are bounded and slow time varying with relative infinitesimal estimation error.