时变大时滞系统参数辨识的仿真研究

研究了时变大时滞系统的参数辨识问题.大时滞系统大多采用补偿控制方法,但是补偿控制方法需要系统的精确数学模型,因而获得大时滞系统的数学模型成为了补偿控制的关键,时变特性使问题复杂化,从而影响了大时滞系统的控制精度.为解决上述问题,提出了一种神经网络的参数辨识策略,利用一个神经元对系统的时滞参数进行辨识,从而可以将时滞从系统模型中分离出来,可利用一个RBF神经网络模型辨识系统的其它参数,使神经元的输出作为RBF神经网络的一个输入,从而实现了串-并联结构的双神经网络拓扑.拓扑结构可以比串级的神经网络提高训练速度,因而也就更适合于实时控制.针对工业锅炉回水温度控制系统的仿真结果验证了所提辨识算法的正确性.     

多输入不确定离散时滞系统的灰色滑模控制

针对存在时变参数不确定性和随机干扰的多输入不确定离散时滞系统,设计了基于LMI的滑模控制器以消除时滞的影响,并使系统状态在有限时间内收敛到零,然后对系统的不确定部分建立灰色估计模型,并进一步设计了灰色补偿器。仿真结果表明,采用所设计的灰色滑模控制器,不仅有效地消除了时滞的影响,抑制了时变参数不确定因素和随机干扰,而且保证多输入不确定离散时滞系统具有良好的鲁棒稳定性。     

执行器失效不确定时滞系统的指定衰减度鲁棒可靠控制

针对一类含有时变时滞的不确定参数线性系统,研究了在执行器发生故障情况下的鲁棒可靠控制器设计问题.系统中的参数不确定性满足广义匹配条件,时变时滞的大小及其变化率有界,并假设故障执行器元件的输出为零.经过适当的模型变换,将原系统的鲁棒指数镇定问题转化为另一个等价系统的鲁棒镇定问题.根据Lyapunov稳定性理论和线性矩阵不等式(LMI)方法,分别给出了鲁棒可靠控制器存在的时滞无关和时滞相关充分条件.仿真结果表明了该控制器设计方法的有效性.     

基于GA参数整定的时变时滞系统灰色预测控制

针对一类数学模型未知且存在时变时滞的复杂系统,提出一种基于遗传算法参数整定的灰色预测控制方法;该方法采用BP神经网络对系统的时变时滞进行辨识,利用灰色预测算法对系统的输出进行预测,进而使用基于遗传算法整定PID控制器对系统进行输出反馈控制;该方法将灰色预测算法与遗传算法相结合,有效提高了控制器的自适应性;通过仿真实例,结果表明该方法能够对具有大时滞、大惯性、模型不确定等特点的复杂系统进行有效地控制;该方法是可行的、有效的.     

时变时滞不确定系统的鲁棒输出反馈控制

研究了时变时滞不确定系统基于状态观测器的动态输出反馈实现鲁棒镇定的分 析和综合问题.所研究的系统不仅同时包含时变状态时滞和时变控制时滞,而且包含时变未 知且有界不确定参数.提出了确保该系统可通过输出反馈鲁棒镇定的充分条件,并将该充分 条件转化为线性矩阵不等式(LMI)问题,最终通过求解两个LMI来构造输出反馈控制律.     

基于神经网络的时滞系统自适应内模控制

将基于人工神经网络的时变时滞系统参数辨识算法与内模控制相结合,提出了时变时滞系统自适应内模控制算法.理论分析及仿真结果表明,该算法能克服时滞及参数的变化,具有鲁棒性好、抗干扰能力强的特点.     

直流微电网分布式控制的时滞稳定化设计

直流微电网中,大部分控制器在设计时未考虑通信延时对系统性能的影响,这可能会导致控制器在实际应用中失效甚至影响系统的稳定性.本文基于一种分布式控制策略,建立了直流微电网系统的时滞模型,重点研究了通信延时对系统稳定性的影响.结合Razumikhin稳定性理论,提出了一种针对时变时滞系统的全时滞稳定性判据.为了更加适用于实际系统,在另一种全时滞稳定性判据的基础上,对系统时滞相关稳定性进行分析,得到保证系统稳定的延时上界,进而给出了一种确定控制参数范围的方法.与传统的将通信延时处理为一阶惯性环节的分析方法相比,基于时滞系统的分析方法更切合实际,为系统的稳定运行提供了一个更宽的时滞范围,提高了系统的可靠性.仿真和实验结果表明本文提出的控制参数设计方法能保证系统在最大延时下的稳定运行.     

时变时滞非线性系统的间歇控制

时变时滞广泛存在于各种非线性系统中,研究了时变时滞非线性系统的间歇控制及其在保密通信中的应用问题,提出了一种间歇控制策略,理论上分析了其正确性,并且给出一个定理来确定控制器的相关参数。根据提出的定理,设计出间歇控制器使得两个含有时变时滞的Chua电路指数达到同步。将该方法应用到混沌保密通信中,在两个系统达到同步的基础上,发送端的信号能够在接收端很好地恢复出来,表明了该方法的可行性。     

周期时变时滞非线性参数化系统的自适应学习控制

针对一阶未知非线性参数化周期时变时滞系统, 设计了一种自适应学习控制方案. 假设未知时变参数, 时变时滞和参考信号的共同周期是已知的, 通过重构系统方程, 将包含时变时滞在内的所有未知时变项合并成为一个周期时变向量, 采用周期自适应律估计该向量. 通过构造一个Lyapunov-Krasovskii型复合能量函数证明了所有信号有界并且跟踪误差收敛. 结果被推广到一类含有混合参数的高阶非线性系统. 通过两个仿真例子说明本文所提出的控制算法的有效性.     

随机时变时滞非线性马尔可夫跳跃系统故障检测问题研究

本文研究一类具有随机时变时滞及随机发生非线性(RONs)的离散时间马尔可夫跳跃系统的故障检测问题.针对多重时变时滞发生时所具有的随机特性,使用相互独立的伯努利随机过程来刻画这一现象.进一步地,在故障检测研究中考虑一类由伯努利分布白序列描述的随机非线性干扰.本文的目的是设计一个故障检测滤波器,使得整个误差动态系统满足均方指数稳定及相应的H∞性能指标.通过Lyapunov稳定性理论及随机分析技术,建立满足均方指数稳定和H∞干扰抑制的充分条件,进而使用半定程序方法求解一个凸优化问题,获取所设计的故障检测滤波器增益特性.最后,通过仿真算例证实了该设计方法的实用性和有效性.     

具有非线性扰动多时滞系统鲁棒稳定性分析与分散鲁棒控制

研究一类非线性参数扰动满足范数有界条件时多重时滞系统的鲁棒稳定性及其分散鲁棒控制。首先利用Lyapunov函数方法分析系统的鲁棒稳定性,获得一种新的稳定性条件,然后利用标量Lyapunov方程方法讨论系统指数稳定的条件,通过对系统采取分散反馈控制,进一步得到系统可鲁棒镇定和可指数镇定的条件。     

区间化时变时延的网络化切换系统建模与控制

研究了具有从传感器到控制器和控制器到执行器存在双边时变时延的网络控制系统指数稳定性的问题.首先将时延变化范围划分为多个等分区间,然后采用增广矩阵的方法建立了参数不确定的离散时间切换闭环系统模型.同时基于平均驻留时间分析方法,给出了系统满足指数稳定的条件,接着进一步的建立了时延区间划分个数与系统状态指数衰减率的定量关系.该方法有效降低了系统设计的保守性,一定程度上减少了系统状态收敛的时间.最后通过数值仿真验证了所提方法的有效性.     

随机时滞神经网络的全局指数稳定性

首先对一般随机系统的渐近特性进行了讨论.然后结合神经网络的特点,应用李雅普诺夫第二方法对一类随机时滞神经网络系统的全局指数稳定性进行了分析,给出了易于判定随机时滞神经网络几乎必然指数稳定性新的代数判据,并给出实例进行仿真实验.     

Exponential stability of nonlinear delay equation with constant decay rate via perturbed system method

This paper studies the exponential stability of nonlinear differential equations with constant decay rate under the assumption that the corresponding crisp equation (without delay, simply, nondelay equation) is exponentially stable. Different from most publications dealing with delay systems by applying Lyapunov-type methods, the perturbed system method is used in this paper. It shall be shown that the considered equations will remain exponentially stable provided the time lag is small enough. Moreover, we formulate and estimate the threshold of delay ensuring exponential stability when a constant decay rate appears explicitly in system model, which is better than the existing results.     

Robust Exponential Stability and Stabilization of a Class of Nonlinear Stochastic Time‐Delay Systems

This paper presents new exponential stability and delayed‐state‐feedback stabilization criteria for a class of nonlinear uncertain stochastic time‐delay systems. By choosing the delay fraction number as two, applying the Jensen inequality to every sub‐interval of the time delay interval and avoiding using any free weighting matrix, the method proposed can reduce the computational complexity and conservativeness of results. Based on Lyapunov stability theory, exponential stability and delayed‐state‐feedback stabilization conditions of nonlinear uncertain stochastic systems with the state delay are obtained. In the sequence, the delayed‐state‐feedback stabilization problem for a nonlinear uncertain stochastic time‐delay system is investigated and some sufficient conditions are given in the form of nonlinear inequalities. In order to solve the nonlinear problem, a cone complementarity linearization algorithm is offered. Mathematical and/or numerical comparisons between the proposed method and existing ones are demonstrated, which show the effectiveness and less conservativeness of the proposed method.     

On exponential stability of integral delay systems

This paper is concerned with exponential stability of a class of integral delay systems with a prescribed decay rate. First, by carefully exploring the literature on this topic, a delay decomposition approach is established to reduce the conservatism in the existing sufficient conditions by constructing new Lyapunov–Krasovskii (LK) functionals. It is proven that the proposed sufficient conditions are less conservative than a recently established set of sufficient conditions. Second, by analyzing the characteristic equation of the considered integral delay system, necessary and sufficient conditions for the stability are obtained by computing the right-most zeros of a certain auxiliary point-delay linear system, for which stability criteria that are easy to test are also established based on this method. Numerical examples illustrate the effectiveness of the obtained results.     

Exponential stability of time-delay systems with nonlinear uncertainties

The global exponential stability and the exponential convergence rate for time-delay systems with nonlinear uncertainties are investigated. A novel exponential stability criterion for the system is derived using the Lyapunov method. These stability conditions are formulated as linear matrix inequalities (LMIs), which can be easily solved by various convex optimization algorithms. The issue of exponential stability for time-delay systems with nonlinear uncertainties using generalized eigen value problem (GEVP) approach remains open, which motivates this paper. Numerical examples are given to illustrate the usefulness of our proposed method.     

Stability and control of uncertain ICPT system considering time‐varying delay and stochastic disturbance

This paper studies the problem of robust exponential stability for uncertain inductively coupled power transfer (ICPT) system considering time‐varying delay and stochastic disturbance. Firstly, the model of the system is set up via a time domain method. Secondly, based on the Newton‐Leibnitz formula and stability theory, a new stability analysis of the ICPT system with uncertain parameter and time‐varying delay or stochastic disturbance is presented respectively. The proposed approach can be applied for analyzing other similar systems, and the obtained results can be also used for estimating convergence rate and region of stability. Thirdly, based on the Lyapunov‐Krasovskii functional (LKF) approach and the stochastic stability theory, robust exponential stability criteria in the mean square are derived and the relevant controller is designed. The proposed method can further reduce conservatism. Finally, the correctness and effectiveness of the obtained results are verified by an example and simulations indicate that the larger time delay, the slower attenuation. Simultaneously, the uncertain parameter and the stochastic disturbance have a significant influence on the region of stability. In addition, in respect of reducing conservatism, the effectiveness of the proposed method is demonstrated by comparing with other papers. In addition, the designed controller shows better performance for the ICPT system with the parameter uncertainty, the time‐varying delay, and the stochastic disturbance.     

Modelling and control of networked control systems with both network-induced delay and packet-dropout

A new switched linear system model is proposed to describe the networked control system (NCS) with both network-induced delay and packet-dropout. A sufficient condition is derived for the exponential stability of the closed-loop NCS, and the obtained condition establishes the quantitative relation between the packet-dropout rate and the stability of the NCS. Design procedures for the state feedback stabilising controllers are also presented by using the augmenting technique. An illustrative example is provided to demonstrate the effectiveness of the proposed method.     

String stability of infinite-dimension stochastic interconnected large-scale systems with time-varying delay

The exponential string stability for a class of nonlinear interconnected large-scale systems with time-varying delay is analysed by using the box theory and constructing a vector Lyapunov function. Under the assumption that the time delay is bounded and continuous, a criterion for exponential string stability of the systems is obtained by analysing the stability of differential inequalities with time-varying delay. The large-scale system is exponential string stable when the conditions associating with the coefficient matrices of the system and the solutions of the Lyapunov equations, interconnected with the system, are satisfied. Since it is independent of the delays and simplifies the calculation, the criterion is easy to apply.