连续系统的约束方差／H_∞综合设计

考虑线性连续随机系统在稳态方差及Ｈ∞指标约束下的综合设计问题，即寻找状态反馈控制器，使闭环系统每个状态分量的稳态方差满足各自预先给定的约束，同时从噪声输入到系统输出的传递函数的Ｈ∞范数也满足预先给定的约束。本文绘出了期望控制器的存在条件及解析表达式，并用算例说明文中设计方法的直接性与有效性。     

连续系统的约束方差／H∝综合设计

考虑线性连续随机系统在稳态方差及Ｈ∝指标约束下的综合设计问题，即寻找状态反馈控制器，使闭环系统每个状态分量的稳态方差满足各自预先给定的约束，同时从噪声输入到系统输出的传递函数的Ｈ∝范数也满足预先给定的约束。本文给出了期望控制器的存在条件及解析表达式，并用算例说明文中设计方法的直接性与有效性。     

极点与状态方差约束下的动态输出反馈可靠控制

研究了线性不确定系统区域极点与状态方差约束下的动态输出反馈可靠控制器的设计问题．首先在连续型执行器故障模型下，利用线性矩阵不等式方法，建立了容错控制中区域极点指标与方差上界指标间的相容性理论；然后，应用极值原理分析了与区域极点指标相容的方差上界指标的取值范围；最后，给出了同时满足区域极点与状态方差约束下的动态输出反馈可靠控制器的设计步骤．仿真实例验证了结果的有效性．     

圆形区域极点及方差约束下线性离散系统的控制设计

本文考虑线性离散随机系统在圆形区域极点及方差约束睛的控制器设计问题，即设计状态反馈控制器，使闭环系统同时满足预先给定的圆形区域极点约束以及预先给定的各状态分量的方差约束。文中利用代数黎卡提方程方法解决了上述问题。     

离散随机系统的容错约束方差控制设计

本文考虑线性离散随机系统的容错约束方差控制设计问题，即设计反馈控制器，使闭环系统在可能的传感器失效不仅保持渐近稳定，而且满足预先给定的稳态方差约束，文中导出期望了容约束方差控制器存在的充分条件，并进一步给出了其参数化代数表达式。     

全状态约束的时滞系统神经网络输出反馈控制

针对一类严格反馈形式的单输入单输出时滞系统,研究在全状态约束下的输出反馈控制.首先,设计状态观测器估计不可测量的状态;其次,利用RBF神经网络逼近未知的非线性函数,利用障碍Lyapunov函数确保全状态约束及Lyapunov-Krasovskii方法消除时滞对系统的影响;最后,设计输出反馈控制器,并且有更少的更新参数减少了计算负荷.所设计的控制器可以保证闭环系统中所有信号半全局一致最终有界,信号误差收敛到小的领域内.仿真例子进一步验证了所提出方法的有效性.     

统一混沌系统的输出反馈控制器设计

利用输出反馈研究了一类统一混沌系统平衡点渐进稳定问题.基于李亚普诺夫稳定理论,线性矩阵不等式方法及不等式技巧,分别用两输入单输出和单输入单输出反馈设计控制器,解决了状态不可测情况下统一混沌系统的控制问题.即使得混沌系统的状态渐进趋于给定的任一平衡点.与现有文献结果相比,所设计的控制器,反馈增益小,结构简单,保守性小.最后以Lorenz系统为例作了仿真,仿真结果表明了该控制器的有效性.     

一类反馈过程神经元网络模型及其学校算法

提出了一种基于权函数基展开的反馈过程神经元网络模型．该模型为三层结构，由输入层、过程神经元隐层和过程神经元输出层组成．输入层完成系统时变过程信号的输入及隐层过程神经元输出信号向系统的反馈；过程神经元隐层用于完成输入信号的空间加权聚合和激励运算，同时将输出信号传输到输出层并加权反馈到输入层；输出层完成隐层输出信号的空间加权聚集和对时间的聚合运算以及系统输出．文中给出了学习算法，并以旋转机械故障自动诊断问题为例验证了模型和算法的有效性．     

基于参数依赖动态输出反馈鲁棒MPC的混沌系统同步

针对离散时间不确定混沌系统的同步控制问题,提出一种基于参数依赖动态输出反馈鲁棒模型预测控制算法.首先,采用主动控制策略,将具有噪声扰动的主从混沌系统同步问题转化为鲁棒稳定性问题;然后,采用参数依赖动态输出控制器和二次有界概念,在保证闭环系统鲁棒稳定性的同时,降低算法的保守性;最后,通过附加约束条件,能够显式处理混沌系统同步中的输入约束.仿真结果表明了所提出算法的有效性.     

一种推广的组合非线性输出反馈控制

针对多变量饱和线性系统的时变参考输入跟踪问题,研究了一种组合非线性输出反馈控制器的设计方法.基于原始的组合非线性反馈理论,构造了全阶和降阶输出反馈控制器.控制器由线性输出反馈项和非线性反馈项组成,使得闭环系统在包含于吸引域的不变集内渐近稳定.除了能够跟踪时变参考输入外,系统还具有良好的动态性能.仿真结果说明了所开发控制器的有效性.     

Global convergence for adaptive one-step-ahead optimal controllers based on input matching

This paper establishes global convergence for adaptive one-step-ahead optimal controllers applied to a class of linear discrete time single-input single-output systems. The class of systems includes all stable systems whether they are minimum phase or not, all minimum phase systems whether they are stable or not, and some unstable nonminimum phase systems. The key substantive assumption is that the one-step-ahead optimal controller designed using the true system parameters leads to a stable closed-loop system. Subject to this natural restriction, it is shown that a simple adaptive control algorithm based on input matching is globally convergent in the sense that the system inputs and outputs remain bounded for all time and the input converges to the one-step-ahead optimal input. Both deterministic and stochastic cases are treated.     

具有减少辨识参数的随机多变量极点配置自适应控制器

本文提出了一种新的限制输出个数减少随机多变量自适应控制中辨识参数的方法，并给出了减少辨识参数的极点配置自适应算法。虽然采用ｎ个输入１个输出的减少辨识参数的模型来设计控制器，但所提出的控制器能够保证被控系统的几个输出跟踪参考输入信号，仿真结果表明，所提出的方法是成功的。     

Sensor and actuator fault diagnosis of systems with discrete inputs and outputs

The paper describes a method for detecting and identifying faults that occur in the sensors or in the actuators of dynamical systems with discrete-valued inputs and outputs. The model used in the diagnosis is a stochastic automaton. The generalized observer scheme (GOS), which has been proposed for systems with continuous-variable inputs and outputs some years ago, are developed for discrete systems. This scheme solves the diagnostic problem as an observation problem, which is set up here for discrete-event systems. As the system under consideration is described by a stochastic automaton rather than a differential equation, the mathematical background and the diagnostic algorithms obtained are completely different from the well-known observers developed for continuous-variable systems. The GOS is extended here by a fault detection module to cope with plant faults that are different from actuator or sensor faults. The diagnostic algorithm consists of two steps, the first detecting the existence of a fault and the second isolating possible sensor or actuator faults or identifying plant faults. The results are applied to quantized systems whose discrete inputs and outputs result from a quantization of the continuous-variable input and output signals. Experimental results illustrate the proposed diagnostic method.     

Reduced-order models and controllers for continuous-time stochasticsystems: an information theory approach

Reduced-order models and controllers for continuous-time stochastic systems are described. The reduced-order models are chosen to minimize the Kullback-Leibler information distance (KLID) between the outputs of the actual and reduced systems. An LQG controller based on a reduced-order system model is described. A second reduced-order controller is found to minimize the KLID between the closed-loop system outputs with the full and reduced-order controllers     

Unified adaptive control of non-minimum-phase systems Part 1. Weighted control effort and pole placement for stochastic systems

Global convergence of a class of unified adaptive control algorithms for discrete time non-minimum-phase stochastic linear systems is established. It is shown that the algorithms ensure that the system inputs and outputs are sample mean bounded, the minimum variance cost function with weighted control effort achieves its minimum possible values, the weighted factors adjusted on-line eventually converge to expected values, and the closed-loop poles arise at prespecified locations.     

ROBUST Hinfinity CONTROL FOR SYSTEMS WITH TIME-VARYING PARAMETER UNCERTAINTY AND VARIANCE CONSTRAINTS

In this paper, the problem of designing robust Hinfinity controllers for linear continuous-time systems subjected to time-varying parameter uncertainty and steady-state variance constraints is considered. The goal of this problem is to design the state feedback controller, such that for all admissible time-varying parameter perturbations, the steady-state variance of each state is not more than the individual prespecified upper bound and the Hinfinity norm of the transfer function from disturbance inputs to system outputs meets the prespecified upper bound constraint, simultaneously. The parameter uncertainties are allowed to be time-varying and norm-bounded. A purely algebraic matrix equation approach is effectively utilized to solve the problem addressed. The existence conditions as well as the explicit expression of desired controllers are presented, and two illustrative examples are used to demonstrate the applicability of the proposed design procedure.     

Non-approximation-based outputs stabilization for switched large-scale nonlinear systems with quantized inputs and sensor uncertainties

A non-approximation-based output feedback control strategy for a class of switched large-scale nonlinear systems with quantized inputs and sensor uncertainties is proposed. A dynamic gain, which is shared by the state observers and controllers of all the subsystems, is designed so that the effects of sensor uncertainties, quantized inputs, unknown parameters, and external disturbances can be compensated. By constructing some common Lyapunov functions (CLFs) shared by the switched systems, it is proved that with the proposed scheme, the closed-loop system stability can be guaranteed under arbitrary switching, and the outputs of all the subsystems can be steered to within arbitrarily small neighborhoods of the origin.     

Reliable stabilization via factorization methods

Reliable stabilization of multicontroller systems composed of one plant and two controllers are considered. The main objective is to propose a reliability design when controllers use independent inputs and outputs of the plant. The assumption of independence is crucial if one wants to increase the chance that at least one of the controllers survives the sensor and actuator failures, which otherwise could disable both controllers and result in a system breakdown     

Extension of minimum variance estimation for systems with unknown inputs

In this paper, we address the problem of minimum variance estimation for discrete-time time-varying stochastic systems with unknown inputs. The objective is to construct an optimal filter in the general case where the unknown inputs affect both the stochastic model and the outputs. It extends the results of Darouach and Zasadzinski (Automatica 33 (1997) 717) where the unknown inputs are only present in the model. The main difficulty in treating this problem lies in the fact that the estimation error is correlated with the systems noises, this fact leads generally to suboptimal filters. Necessary and sufficient conditions for the unbiasedness of this filter are established. Then conditions under which the estimation error and the system noises are uncorrelated are presented, and an optimal estimator and a predictor filters are derived. Sufficient conditions for the existence of these filters are given and sufficient conditions for their stability are obtained for the time-invariant case. A numerical example is given in order to illustrate the proposed method.     

State-dependent parameter modelling and identification of stochastic non-linear sampled-data systems

State-dependent parameter representations of stochastic non-linear sampled-data systems are studied. Velocity-based linearization is used to construct state-dependent parameter models which have a nominally linear structure but whose parameters can be characterized as functions of past outputs and inputs. For stochastic systems state-dependent parameter ARMAX (quasi-ARMAX) representations are obtained. The models are identified from input–output data using feedforward neural networks to represent the model parameters as functions of past inputs and outputs. Simulated examples are presented to illustrate the usefulness of the proposed approach for the modelling and identification of non-linear stochastic sampled-data systems.