基于KRLS的非均匀采样非线性系统辨识

利用提升技术可将非均匀采样非线性系统离散化为一个多输入单输出传递函数模型,从而将系统输出表示为非均匀刷新非线性输入和输出回归项的线性参数模型,进一步基于非线性输入的估计或过参数化方法进行辨识.然而,当非线性环节结构未知或不能被可测非均匀输入参数化表示时,上述辨识方法将不再适用.为了解决这个问题,利用核方法将原始非线性数据投影到高维特征空间中使其线性可分,再对投影后的数据应用递推最小二乘算法进行辨识,提出基于核递推最小二乘的非均匀采样非线性系统辨识方法.此外,针对系统含有有色噪声干扰的情况,参考递推增广最小二乘算法的思想,利用估计残差代替不可测噪声,提出核递推增广最小二乘算法.最后,通过仿真例子验证所提算法的有效性.     

基于扩张状态观测器的可调系统参数辨识与实验

在测量噪声条件下,研究基于扩张状态观测器(ESO)的参数辨识方法和实际应用.结合仿真与实验,分析测量噪声对ESO观测状态输出的影响.为了提高实际应用中的辨识效果,通过引入积分型滤波器对观测状态中的噪声进行抑制.所提出方法可以很好地处理非均匀周期采样数据,适用于不规则采样系统的参数辨识.最后,在一类参数可调实验系统上构造两类实际对象,通过实验验证所提出辨识方法的有效性和实用性.     

非均匀采样数据系统的一种辨识方法

阐述了非均匀采样方案,推导了非均匀多率采样系统的状态空间模型,进一步获得了对应的传递函数模型.为解决辨识模型信息向量中存在未知变量的问题,使用辅助模型技术,用辅助模型的输出代替系统的未知变量,进而提出了非均匀采样数据系统的辅助模型随机梯度辨识算法.为了提高算法收敛速度和改善参数估计精度,在算法中引入遗忘因子,给出了相应的辅助模型带遗忘因子随机梯度算法.仿真结果表明,引入遗忘因子后,算法的收敛速度加快,参数估计精度提高.     

非均匀多采样率非线性系统的模糊辨识

针对非均匀多采样率非线性系统辨识问题, 提出一种基于模糊模型的辨识方法. 首先, 分析了非线性系统在输入信号非均匀周期刷新, 输出信号周期采样的情况下, 非线性系统可以通过提升技术, 利用多个局部的线性模型加权组合来描述; 然后, 提出一个基于GK模糊聚类和递推最小二乘的模糊辨识算法; 最后, 针对化工pH 中和过程非线性系统, 采用非均匀采样数据建立其模糊模型, 以验证所提出方法的有效性.

     

多采样率系统的辨识问题综述

在多率采样系统中, 采样间隔不均匀. 本文综述了文献中有关多率采样系统的数学模型, 如线性周期时变模型、频域模型和连续状态空间模型等. 同时对相应的辨识方法, 如提升、频域方法、子空间辨识方法等, 也进行了全面的综述. 对多率采样系统辨识中存在的一些问题, 包括辨识模型的选择、一致性问题、带约束条件的辨识方法和辨识收敛性等, 也作了深入的讨论.     

非均匀周期采样系统的递阶最小二乘辨识方法

针对非均匀周期采样系统,通过状态空间模型离散化方法得到其输入输出表达形式.鉴于参数化后得到的辨识模型同时包含1个参数向量和1个参数矩阵,利用递阶辨识原理,将辨识模型分解为分别含有参数向量和参数矩阵的2个虚拟子系统;考虑到系统的因果约束问题,将包含参数矩阵的子系统分解为子子系统进行辨识,从而提出这类非均匀采样系统的递阶最小二乘辨识方法.仿真例子表明该算法是有效的.     

基于子空间方法的非均匀周期刷新和采样系统辨识

针对非均匀周期刷新和采样系统的建模问题, 对于含有提升变量的状态空间模型, 提出基于子空间技术的辨识方法. 首先, 通过系统的采样数据建立由Hankel 矩阵组成的扩展状态空间方程; 然后, 利用斜交投影的原理、方法和奇异值分解, 通过子空间辨识算法确定增广观测矩阵和状态向量, 通过最小二乘方法确定模型的参数矩阵; 最后, 通过仿真实例表明了所提出算法的有效性.

     

基于高斯混合模型聚类的非均匀采样系统的多模型切换辨识

从概率统计方法出发,提出一种基于高斯混合模型聚类与递推最小二乘算法的非均匀采样系统的多模型建模方法.首先,采用高斯混合模型作为调度函数,使用最大期望(EM)算法迭代更新估计高斯混合模型中参数,从而通过每个子系统的高斯概率密度函数计算和比较来确定子系统的激活情况; 其次,采用递推最小二乘算法估计局部子系统参数;然后,使用鞅收敛定理对所提出的算法性能进行分析; 最后,通过非均匀采样系统的多模型建模来证明所提出方法的有效性.     

线性时变反馈系统的状态空间模型和模态参数递推辨识

针对闭环系统中时变状态空间模型和模态参数的辨识问题, 提出一种递推辨识格式, 将这种格式与递推子空间方法结合, 得到一种辨识方法. 该方法通过重建输入输出数据之间的关系, 递推辨识得到闭环系统的时变状态空间模型和模态参数. 算例研究了系统在模态参数突变和周期变化两种情况下的辨识问题, 仿真结果表明, 所提出算法能有效辨识线性时变反馈系统的状态空间模型和模态参数.

     

非均匀采样数据系统的新型模型描述方法

提升技术是处理非均匀采样数据（Non-uniformly sampled-data,NUSD）系统的标准工具.然而,提升状态空间模型存在因果约束问题,相应的提升传递函数模型结构复杂,且参数数目过多.因此,它们不便于非均匀采样数据系统的辨识与控制.通过引入时变后移算子,本文提出了一种输入输出表达的新型模型描述方法.该模型能够克服提升系统模型的缺点,使得传统单率系统的辨识和控制方法能够推广到非均匀采样数据系统中.仿真结果表明了新模型的优越性和有效性.     

Continuous-time system modelling using the weighted-parabola-overlapping numerical integration method

Numerical integration is a widely used method in scientific areas where the uniform integral interval can be selected optionally. Simpson integration and Newton-Cotes formula, for example, are available in many textbooks and software. However, in system science, the data may come from practical sampled continuous systems. Sometimes the sampling interval is non-uniform and the integration values at each sampled point are needed. For example, when the parameters of a continuous system model need to be estimated by least squares, we want the integration value at each sampling point. In that case, trapezoidal numerical integration is mostly adopted, which has only first-order algebraic accuracy (the straight line). When sampled points are fewer, the errors of numerical integration such as noise can interfere with the modelling. We propose a weighted-parabola-overlapping (WPO) numerical integration method, which possesses third-order algebraic accuracy and needs nearly the same computation of second-order power function curve fitting (the parabola). The proposed method is applicable to non-uniformly sampled systems. As a specific case, it is as adequate as uniformly sampled systems and the expression becomes very simple. Two examples are given: one is the identification of a non-uniformly sampled system by least-squares (LS), the other is local cerebral blood flow (LCBF) measurement with positron emission tomography (PET) by the integrated projection (IP) technique. Both are non-uniformly sampled systems and have very few data.     

基于奇异值分解的非均匀采样非线性系统的模糊模型辨识

针对复杂、不确定、非均匀采样数据的非线性系统,提出一种基于矩阵奇异值分解(SVD)的模型结构辨识和参数估计的建模方法.首先,利用矩阵奇异值(SVD)分解算法分析各局部模型与奇异值、积累贡献率的关系,确定模糊模型的规则数,从而实现模型的结构优化;然后,为了克服递推最小二乘出现的误差积累、传递现象,采用奇异值分解的递推最小二乘估计模型的结论参数;最后,通过仿真实例验证所提出算法的有效性.     

Subspace-based continuous-time identification of fractional order systems from non-uniformly sampled data

This work focuses on the identification of fractional commensurate order systems from non-uniformly sampled data. A novel scheme is proposed to solve such problem. In this scheme, the non-uniformly sampled data are first complemented by using fractional Laguerre generating functions. Then, the multivariable output error state space method is employed to identify the relevant system parameters. Moreover, an in-depth property analysis of the proposed scheme is provided. A numerical example is investigated to illustrate the effectiveness of the proposed method.     

Recursive Bayesian Algorithm for Identification of Systems with Non-uniformly Sampled Input Data

To identify systems with non-uniformly sampled input data, a recursive Bayesian identification algorithm with covariance resetting is proposed. Using estimated noise transfer function as a dynamic filter, the system with colored noise is transformed into the system with white noise. In order to improve estimates, the estimated noise variance is employed as a weighting factor in the algorithm. Meanwhile, a modified covariance resetting method is also integrated in the proposed algorithm to increase the convergence rate. A numerical example and an industrial example validate the proposed algorithm.     

A recursive algorithm for estimating multiple models continuous transfer function with non-uniform sampling

A multiple model recursive least squares algorithm combined with a first-order low-pass filter transformation method, named λ-transform, is proposed for the simultaneous identification of multiple model orders continuous transfer functions from non-uniformly sampled input–output data. The λ-transformation is shown to be equivalent to a canonical transformation between discrete z-domain and δ-domain using the negative value of the λ-transform filter time-constant instead of the sampling interval parameter. The proposed algorithm deals with oversampling, sampling jitter or non-uniform sample intervals without the need for extra digital anti-aliasing pre-filtering, downsampling or interpolation algorithms, producing multiple models with a cost function that facilitates automatic selection of best-fitted models. Besides, measurement noise is noted as beneficial, bringing up an inherent bias toward low-order models. Simulated examples and a drum-boiler level experimental result exhibiting non-minimum phase behaviour illustrate the application of the proposed method.