时变系统遗忘因子最小二乘法的有界性收敛性

利用随机过程理论研究了遗忘因子最小二乘法（FFLS）的有界收敛性，给出了参数估计误差的上界，分析表明：i)对于时不变确定性系统；FFLS算法产生的参数估计以指数速度收敛于真参数；ii)对于时不变随机系统，FFLS算法给出界均方估计误差，iii)对于时变随机系统，FFLS算法可以跟踪时变参数，且跟踪误差有界。     

时变系统遗忘因子最小二乘法的有界收敛性

利用随机过程理论研究了遗忘因子最小二乘法 (FFLS)的有界收敛性, 给出了参数估计误差的上界. 分析表明: i)对于时不变确定性系统, FFLS算法产生的参数估计以指数速度收敛于真参数; ii)对于时不变随机系统, FFLS算法给出有界均方估计误差; iii)对于时变随机系统, FFLS算法可以跟踪时变参数, 且跟踪误差有界.     

一类线性时变系统组合自适应迭代学习辨识

针对一类在有限时间区间上可重复运行的既含时变参数又含时不变参数的高阶线性时变系统,提出了一种模型参考组合自适应迭代学习参数辨识算法.应用Lyapunov方法,给出了时不变参数的时域自适应学习律和时变参数的迭代域自适应学习律,分析了参数估计和模型状态跟踪误差的有界性与收敛性.该算法适于时变和时不变参数并存的线性系统的参数辨识,可加快参数估计的收敛速度.仿真例子验证了所提出的辨识算法的有效性.     

一类非线性系统的自适应观测器设计

在故障诊断应用中, 状态方程中的未知参数和输出方程中的未知参数分别表征执行机构故障和传感器故障, 所以研究状态方程和输出方程同时含有未知参数的自适应观测器有着实际的应用意义. 本文基于高增益观测器和自适应估计理论, 针对状态方程和输出方程同时含有未知参数的一类一致可观的非线性系统, 用构造性方法设计了一种联合估计状态和未知参数的自适应观测器. 该自适应观测器的参数估计采用时变增益矩阵, 结构形式及参数设置简单. 给出了使该自适应观测器满足全局指数收敛性的持续激励条件, 并在理论上简洁地证明了该自适应观测器的全局指数收敛性. 数值仿真结果表明该自适应观测器具有良好的快速收敛性、跟踪性等期望性能.     

多新息随机梯度辨识方法

多新息随机梯度辨识方法是系统辨识和参数估计的一种基本方法.该方法由于采用了间断迭代,因此可以克服坏数据对参数估计的影响,且具有较强的鲁棒性,又可以跟踪时变参数.作者从理论上给出了多新息随机梯度辨识方法的推导过程,同时列出多新息随机梯度辨识方法的各种变形.数字仿真实验表明多新息随机梯度辨识方法具有良好的性能.     

基于数据滤波的随机梯度辨识方法

针对有色噪声干扰下的随机系统,利用数据滤波技术,对输入输出数据进行滤波,将具有滑动平均噪声的原始系统转换为白噪声干扰下的系统,提出有限脉冲响应滑动平均系统的滤波增广随机梯度算法,并对该算法进行收敛性分析.此外,为了提高参数估计的精度和加快算法的收敛速度,使用多新息辨识理论提出滤波多新息增广随机梯度算法,并分析其收敛性.与增广随机梯度算法相比,所提出的滤波增广随机梯度算法和滤波多新息增广随机梯度算法可以得到更高精度的参数估计.最后,通过仿真实例表明了所提出算法的有效性.     

神经网络在线投影算法及非线性建模应用

针对神经网络难以在线学习的缺点,把神经网络当作结构已知的非线性系统,权系数的学习看成非线性系统的参数估计,基于新估计准则的非线性系统在线参数估计投影算法,给出前馈神经网络的一种在线运行投影学习算法.理论上证明该算法的全局收敛性,讨论算法参数的物理意义和取值范围.通过2个非线性时变系统的神经网络建模应用的仿真,验证算法的全局收敛性和在线运行能力.     

基于随机牛顿算法的离散系统自适应参数估计

针对一类离散系统,提出一种基于随机牛顿算法的自适应参数估计新框架,相较于已有的参数估计算法,所提出方法仅要求系统满足有限激励条件,而非传统的持续激励条件.所提出算法的核心思想在于通过对原始代价函数的修正,在使用当前时刻误差信息的基础上融入历史误差信息,进而通过对历史信息和历史激励的复用使得持续激励条件转化为有限激励条件;然后,为了解决传统算法收敛速度慢的问题并避免潜在的病态问题,采用随机牛顿算法推导出参数自适应律,并引入含有历史信息的海森矩阵作为时变学习增益,保证参数估计误差指数收敛;最后,基于李雅普诺夫稳定性理论给出不同激励条件下所提出算法的收敛性结论和证明,并通过对比仿真验证所提出算法的有效性和优越性.     

多天线时变信道频偏的估计下界

针对多天线信道频偏参数的随机时变特点,建立参数化的观测模型,根据时变参数估计的克拉美-罗下界求取原理,在非线性的观测模型和序贯的观测值下,利用参数真值推导频偏估计的克拉美-罗下界。由于特征参数的真值不可知,因此将观测模型在估计值处进行二阶泰勒级数展开,通过参数估计值获得频偏参数估计的近似下界。仿真结果表明,基于参数真值和估计值的频偏估计下界具有渐近收敛性,且随信噪比和观测值个数的增加均能达到稳态。     

非线性系统参数估计的一类有效搜索策略

结合模拟退火的随机概率突跳性搜索和单纯形法的凸多面体几何搜索,提出了非线性 系统参数估计的一类有效搜索策略.通过对多种非线性定常系统的多维参数估计和非线性时变 水箱系统的参数与时滞在线联合估计的仿真研究,验证了该方法的可行性、有效性及其一定的 抗噪声能力.     

On the adaptive control of a class of systems with random parameters and disturbances

The adaptive control of discrete time parameter linear stochastic systems with random parameters is investigated. It is shown that systems whose (unknown) autoregressive parameters undergo bounded martingale difference disturbances may be stabilized by the application of the so-called Modified Least Squares adaptive control algorithm. Asymptotically, the sample mean square performance criterion is equal to the one step ahead minimum variance control loss (which equals the prediction error variance when the system parameters are known) plus a term which is bounded by a quantity proportional to the square of the bound on the parameter disturbance. This latter term may be interpreted as the increase in the prediction error variance due to the random parameter variation.     

Random linear-quadratic mathematical models: Computing explicit solutions and applications

This paper deals with the study of linear random population models and with a random logistic model (where parameters are random variables). Assuming appropriate conditions, the stochastic processes solutions are obtained under closed form using mean square calculus. Expectation and variance expressions for the stochastic processes solutions are given and illustrative examples are included.     

Extended stochastic gradient identification algorithms for Hammerstein–Wiener ARMAX systems

An extended stochastic gradient algorithm is developed to estimate the parameters of Hammerstein–Wiener ARMAX models. The basic idea is to replace the unmeasurable noise terms in the information vector of the pseudo-linear regression identification model with the corresponding noise estimates which are computed by the obtained parameter estimates. The obtained parameter estimates of the identification model include the product terms of the parameters of the original systems. Two methods of separating the parameter estimates of the original parameters from the product terms are discussed: the average method and the singular value decomposition method. To improve the identification accuracy, an extended stochastic gradient algorithm with a forgetting factor is presented. The simulation results indicate that the parameter estimation errors become small by introducing the forgetting factor.     

衰减激励条件下最小均方算法的收敛性

给出了衰减激励信号的定义,并在衰减激励条件下,利用随机过程理论,研究了随机系统最小均方算法的收敛速率,阐述了参数估计误差收敛时,衰减指数和算法中设计参变量 (收敛因子或步长 )的选择方法.分析表明:在衰减激励条件下,最小均方算法也具有良好的性能:当衰减指数和设计参变量满足一定条件时,则参数估计误差一致收敛于零.     

Input–output finite-time mean square stabilisation of stochastic systems with Markovian jump

The notion of input–output finite-time mean square (IO-FTMS) stability is introduced for Itô-type stochastic systems with Markovian jump parameters. Concerning a class of random input signals W, sufficient conditions are presented for the IO-FTMS stability and stabilisation of stochastic nonlinear Markov jump systems in terms of coupled Hamilton–Jacobi inequalities. When specialising to the linear case, these criteria are turned into coupled linear matrix inequalities. Moreover, the quadratic IO-FTMS stabilisation is addressed when polytopic uncertainty appears in the transition rate. Finally, a numerical example with simulations is exploited to illustrate the proposed techniques.     

Active disturbance rejection control approach to output‐feedback stabilization of lower triangular nonlinear systems with stochastic uncertainty

In this paper, we apply the active disturbance rejection control approach to output‐feedback stabilization for uncertain lower triangular nonlinear systems with stochastic inverse dynamics and stochastic disturbance. We first design an extended state observer (ESO) to estimate both unmeasured states and stochastic total disturbance that includes unknown system dynamics, unknown stochastic inverse dynamics, external stochastic disturbance, and uncertainty caused by the deviation of control parameter from its nominal value. The stochastic total disturbance is then compensated in the feedback loop. The constant gain and the time‐varying gain are used in ESO design separately. The mean square practical stability for the closed‐loop system with constant gain ESO and the mean square asymptotic stability with time‐varying gain ESO are developed, respectively. Some numerical simulations are presented to demonstrate the effectiveness of the proposed output‐feedback control scheme. Copyright © 2016 John Wiley & Sons, Ltd.     

Performance analysis of stochastic gradient algorithms under weak conditions

By using the stochastic martingale theory, convergence properties of stochastic gradient （SG） identification algorithms are studied under weak conditions. The analysis indicates that the parameter estimates by the SG algorithms consistently converge to the true parameters, as long as the information vector is persistently exciting （i.e., the data product moment matrix has a bounded condition number） and that the process noises are zero mean and uncorrelated. These results remove the strict assumptions, made in existing references, that the noise variances and high-order moments exist, and the processes are stationary and ergodic and the strong persis- tent excitation condition holds. This contribution greatly relaxes the convergence conditions of stochastic gradient algorithms. The simulation results with bounded and unbounded noise variances confirm the convergence conclusions proposed.     

具有BIBO稳定的网络系统随机容错设计

研究了具有均方BIBO稳定的网络化控制系统的随机容错控制及控制器设计问题。针对网络化控制系统的传感器失效故障和执行器失效故障均具有随机性这一现象,将传感器和执行器的故障建模为相互独立的Bernoulli随机变量序列;利用Lyapunov稳定性理论,结合线性矩阵不等式技术,通过对反馈增益矩阵的分解,得到了网络控制系统存在传感器失效故障和执行器失效故障情况下的均方BIBO稳定条件;基于该稳定条件给出了系统随机容错控制器的设计。以数值实例验证了该方法的有效性。     

Mean square exponential input-to-state stability of stochastic memristive complex-valued neural networks with time varying delay

In this paper, mean square exponential input-to-state stability (exp-ISS) of stochastic memristive complex-valued neural networks (SMCVNNs) is investigated. By utilising Lyapunov functional and stochastic analysis theory, a sufficient criterion is derived to assure the mean square exp-ISS of the SMCVNNs. The obtained results not only generalise the previous works in the literature about real-valued neural networks as special cases, but also can be easily checked by parameters of system. Numerical simulations are given to show the effectiveness of our theoretical results.