An analysis of the parametrization by data driven local coordinates for multivariable linear systems

In this paper, we study a novel parametrization for state-space systems, namely data driven local coordinates (DDLC) which have recently been introduced and applied. Even though DDLC has meanwhile become the default parametrization used in the system identification toolbox of the software package MATLAB, an analysis of properties of DDLC, which are relevant to identification, has not been performed up to now. In this paper, we provide insights into the geometry and topology of the DDLC construction and show a number of results which are important for actual identification such as maximum likelihood-type estimation.     

An analysis of separable least squares data driven local coordinates for maximum likelihood estimation of linear systems

In this paper we study a novel parametrization for state-space systems, namely separable least squares data driven local coordinates (slsDDLC). The parametrization by slsDDLC has recently been successfully applied to maximum likelihood estimation of linear dynamic systems. In a simulation study, the use of slsDDLC has led to numerical advantages in comparison to the use of more conventional parametrizations, including data driven local coordinates (DDLC). However, an analysis of properties of slsDDLC, which are relevant to identification, has not been performed up to now. In this paper, we provide insights into the geometry and topology of the slsDDLC construction and show a number of results which are important for actual identification, in particular for maximum likelihood estimation. We also prove that the separable least squares methodology is indeed guaranteed to be applicable to maximum likelihood estimation of linear dynamic systems in typical situations.     

Subspace identification of multivariable linear parameter-varying systems

A subspace identification method is discussed that deals with multivariable linear parameter-varying state-space systems with affine parameter dependence. It is shown that a major problem with subspace methods for this kind of system is the enormous dimension of the data matrices involved. To overcome the curse of dimensionality, we suggest using only the most dominant rows of the data matrices in estimating the model. An efficient selection algorithm is discussed that does not require the formation of the complete data matrices, but processes them row by row.     

Kernel methods for subspace identification of multivariable LPV and bilinear systems

Subspace identification methods for multivariable linear parameter-varying (LPV) and bilinear state-space systems perform computations with data matrices of which the number of rows grows exponentially with the order of the system. Even for relatively low-order systems with only a few inputs and outputs, the amount of memory required to store these data matrices exceeds the limits of what is currently available on the average desktop computer. This severely limits the applicability of the methods. In this paper, we present kernel methods for subspace identification performing computations with kernel matrices that have much smaller dimensions than the data matrices used in the original LPV and bilinear subspace identification methods. We also describe the integration of regularization in these kernel methods and show the relation with least-squares support vector machines. Regularization is an important tool to balance the bias and variance errors. We compare different regularization strategies in a simulation study.     

Constrained state-space system identification with application to structural dynamics

Constrained identification of state-space models representing structural dynamic systems is addressed. Based on physical insight, transfer function constraints are formulated in terms of the state-space parametrization. A simple example shows that a method tailored for this application, which utilizes the non-uniqueness of a state-space model, outperforms the classic sequential quadratic programming method in terms of robustness and convergence properties. The method is also successfully applied to real experimental data of a plane frame structure.     

Subspace state space system identification for industrial processes

We give a general overview of the state-of-the-art in subspace system identification methods. We have restricted ourselves to the most important ideas and developments since the methods appeared in the late eighties. First, the basics of linear subspace identification are summarized. Different algorithms one finds in literature (such as N4SID, IV-4SID, MOESP, CVA) are discussed and put into a unifying framework. Further, a comparison between subspace identification and prediction error methods is made on the basis of computational complexity and precision of the methods by applying them on 10 industrial data sets.     

Optimal instrumental-variable methods for identification of multivariable linear systems

The identification of multivariable linear systems using instrumental-variable (IV) methods is discussed. The emphasis is on accuracy properties. The IV estimates of multi-input-multi-output system parameters are proved to be asymptotically Gaussian distributed under weak conditions. An explicit expression for the covariance matrix of the parameter estimates is given. It is then shown how this matrix can be optimized by appropriately choosing the IV variant. The optimal accuracy so obtained is for some model structures equal to that corresponding to a prediction error method.     

On-line structured subspace identification with application to switched linear systems

This article is concerned with the identification of switched linear multiple-inputs–multiple-outputs state-space systems in a recursive way. First, a structured subspace identification scheme for linear systems is presented which turns out to have many attractive features. More precisely, it does not require any singular value decomposition but is derived using orthogonal projection techniques; it allows a computationally appealing implementation and it is closely related to input–output models identification. Second, it is shown that this method can be implemented on-line to track both the range space of the extended observability matrix and its dimension and thereby, the system matrices. Third, by making use of an on-line switching times detection strategy, this method is applied to blindly identify switched systems and to label the obtained submodels. Simulation results on noisy data illustrate the abilities and the benefits of the proposed approach.     

State-space analysis and identification for a class of hysteretic systems

In this paper we present results on the twin subjects of system analysis and system identification for a class of state-space realizable dynamic systems under the influence of hysteresis. The class of systems in question consists of models in the form of a linear time-invariant dynamic system in series with a differential model of hysteresis. It will be demonstrated that under fairly light constraints on the differential model of hysteresis, it is possible to design a series of experiments leading towards the identification of the full state-space realization. The approach is tested successfully on a high-precision mechanical translation system affected by hysteresis.     

An approach to multivariable system identification

Two prototype identifiable structures are presented which make possible the identification via an equation-error model reference adaptive system of linear plants with rational transfer function matrices. The structures include as specialisations many of the particular structures presented hitherto in the literature. Convergence properties are also discussed, and several modes of convergence are distinguished: model output to plant output, model transfer function matrix to plant transfer function matrix, and model parameters to plant parameters. Conditions are presented for exponentially fast convergence in the absence of noise.     

Optimal semistable control for continuous-time linear systems

In this paper, we develop a new H₂ semistability theory for linear dynamical systems. Specifically, necessary and sufficient conditions based on the new notion of weak semiobservability for the existence of solutions to the semistable Lyapunov equation are derived. Unlike the standard H₂ optimal control problem, a complicating feature of the H₂ optimal semistable control problem is that the semistable Lyapunov equation can admit multiple solutions. We characterize all the solutions using matrix analysis tools. With this theory, we present a new framework to design H₂ optimal semistable controllers for linear coupled systems by converting the original optimal control problem into a convex optimization problem.     

Continuity of dynamical systems: A system theoretic approach

This paper studies continuity of linear time-invariant dynamical systems, defined in terms of the system’s behavior. This concept is related to parameter continuity of associated system representations. For the case at hand, these will be autoregressive (AR) representations. The main result states that a family of linear time-invariant systems, with uniformly bounded dimension of the state space, converges if and only if the systems admit a convergent full rank (AR) representation.     

Stability and stabilization of linear sampled-data systems with multi-rate samplers and time driven zero order holds

In multi-rate sampled-data systems, a continuous-time plant is controlled by a discrete-time controller which is located in the feedback loop between sensors with different sampling rates and actuators with different refresh rates. The main contribution of this paper is to propose sufficient Krasovskii-based stability and stabilization criteria for linear sampled-data systems, with multi-rate samplers and time driven zero order holds. For stability analysis, it is assumed that an exponentially stabilizing controller is already designed in continuous-time and is implemented as a discrete-time controller. For each sensor (or actuator), the problem of finding an upper bound on the lowest sampling frequency (or refresh rate) that guarantees exponential stability is cast as an optimization problem in terms of linear matrix inequalities (LMIs). Furthermore, sufficient conditions for controller synthesis are formulated as LMIs. It is shown through examples that choosing the right sensors (or actuators) with adequate sampling frequencies (or refresh rates) has a considerable impact on stability of the closed-loop system.     

Output feedback pole placement for linear time-varying systems with application to the control of nonlinear systems

The output feedback pole placement problem is solved in an input-output algebraic formalism for linear time-varying (LTV) systems. The recent extensions of the notions of transfer matrices and poles of the system to the case of LTV systems are exploited here to provide constructive solutions based, as in the linear time-invariant (LTI) case, on the solutions of diophantine equations. Also, differences with the results known in the LTI case are pointed out, especially concerning the possibilities to assign specific dynamics to the closed-loop system and the conditions for tracking and disturbance rejection. This approach is applied to the control of nonlinear systems by linearization around a given trajectory. Several examples are treated in detail to show the computation and implementation issues.     

Structural identification and software package for linear multivariable systems

This paper is concerned with the structural identification of linear multivariable systems and an interactive identification package. The structural identification is done by taking the time-invariant subsystem from the realizations of the input-output relations identified using data of disjoint time intervals, and the statistical hypothesis test is employed to determine the order, where the input-output relation is identified based on the generalized least squares method using the possibly larger model for the plant. The identification package is for the identification of the input-output relation of a linear multivariable system, for the structural identification based on the realization and for data management.     

基于状态观测与教学优化算法的多变量系统历史数据驱动辨识

常规智能算法与历史数据结合进行多变量系统辨识的方法,选取表征系统由稳态过渡到动态过程的数据作建模数据,当该过程含有未知扰动时,无法准确建立对象模型.本文提出一种基于状态观测与教学优化算法的多变量系统历史数据驱动辨识方法.该方法选取系统由动态回归稳态的历史数据,并根据其稳态终值进行去稳态分量处理.再将其分为两段,应用状态观测器与预估模型对第1段数据末端的系统状态进行估计,并将估计值作为第2段数据对应的系统初态;应用第2段数据的输入对预估模型进行仿真,采用教学优化算法寻优预估模型参数,使仿真输出接近实际输出.仿真实验表明该方法可以克服扰动对模型辨识精度的影响.最后对某火电机组协调控制系统进行建模,结果表明了该方法的有效性.     

Global structure of families of multivariable linear systems with an application to identification

The identification problem for linear stochastic systems may be stated roughly as follows: given observations on two stochastic processes which are the input and output of some unknown linear system, determine some estimate of the parameters of the system. A set of candidate linear systems which contains the “true” system is introduced, and probabilistic assumptions on the two stochastic processes turn the identification problem into the deterministic problem of minimizing some objective function over this candidate model set. If this set is a manifold, the existence of globally convergent identification algorithms hinges on the critical point behavior of the objective functions which it carries. By way of Morse Theory, the critical point behavior of objective functions on a manifold has implications with regard to the topology of the manifold. This paper analyzes the topology and critical point behavior of objective functions on a specific manifold of linear systems which appears frequently as the candidate model set in identification problems. This manifold is the set Σ of allm-input,p-output linear systems of fixed McMillan degree with real or complex coefficients. Over this manifold sits the principal bundle \(\tilde \Sigma\) of minimal realizations of systems in Σ It is shown that there exist three natural analytic metrics on the associated vector bundle. It is also shown that, in the real case, the first Stiefel-Whitney class of the bundle \(\tilde \Sigma\) has min(m, p)-1 nonvanishing powers; the same conclusion is drawn about the first Chern class of \(\tilde \Sigma\) in the complex case. These results, which follow from Morse Theory and some elementary homotopy and homology theory, imply that the category of the bundle \(\tilde \Sigma\) is at least min(m, p), and hence that the Lusternik-Schnirelmann category of Σ is at least min(m, p). It follows that canonical forms (i.e. sections of \(\tilde \Sigma\) ) may exist only when min(m, p) = 1 and that any objective function on Σ with compact sublevel sets has at least min(m, p) critical points. In particular, there exist on Σ no globally convergent gradient algorithms when min(m, p) > 1.     

On linear models for nonlinear systems

Best linear time-invariant (LTI) approximations are analysed for several interesting classes of discrete nonlinear time-invariant systems. These include nonlinear finite impulse response systems and a class of nonsmooth systems called bi-gain systems. The Fréchet derivative of a smooth nonlinear system is studied as a potential good LTI model candidate. The Fréchet derivative is determined for nonlinear finite memory systems and for a class of Wiener systems. Most of the concrete results are derived in an ?_∞ signal setting. Applications to linear controller design, to identification of linear models and to estimation of the size of the unmodelled dynamics are discussed.