VAN OVERSCHEE,P.,AND DE MOOR, B. 1996. Subspace Identification for Linear Systems: Theory, Implementation, Applications.Kluwer Academic Publishers, Dordrecht, Netherlands.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....One problem is the initialisation of the EM algorithm. Standard initialisation schemes can lead to poor formant trajectories. These trajectories are however important for vowel intelligibility. This paper investigates the suitability of subspace state space system identification (4SID) techniques [ 10,11], which are popular in system identification, for EM initialisation. Speech is split into fixed length, overlapping frames. Overlap encourages temporally smoother parameter transitions between frames. Due to the slow non stationary behaviour of speech, each frame of speech is assumed ....

....from stage two can be considered as outputs from a parallel bank of Kalman filters, each one estimating a state from the previous i observations, and initialised using zero conditions. The particular subspace algorithm and software used in this paper is the sto pos algorithm as detailed in [10]. Although this algorithm introduces a small bias into some of the parameter estimates, it guarantees positive realness of the covariance sequence, which in turn guarantees the definition of a forward innovations model. 3 Experiments Experiments are conducted on the phrase in arithmetic , ....

[Article contains additional citation context not shown here]

Van Overschee, P. & De Moor, B. (1996) Subspace Identification for Linear Systems: Theory, Implementation, Applications. Dordrecht, Netherlands: Kluwer Academic Publishers.

.... requiring knowledge of certain structure indices which are difficult to determine in practice [152] An interesting solution to MIMO model identification (including ARMA parameter estimation) via a subspace based realization approach using state space formulation has recently been proposed [439]. It is applicable to both inputoutput ( deterministic ) models with noisy output measurements as well as to multivariate time series ( stochastic models) Systems and techniques not captured by the above formulations (stationary linear time series, linear systems with noisy output measurements ....

P. van Overschee and B. de Moor, Subspace Identification for Linear Systems: Theory - Implementation - Methods, Kluwer Academic, Boston, MA, 1996.

....IN SUBSPACE IDENTIFICATION N. Mastronardi, P. Van Dooren, S. Van Huffel Universit a della Basilicata, Potenza, Italy, Katholieke Universiteit Leuven, Leuven, Belgium Catholic University of Louvain, Louvain la Neuve, Belgium Katholieke Universiteit Leuven, Leuven, Belgium Two recent approaches [4, 14] in subspace identification problems require the computation of the # factor of the ## factorization of a block Hankel matrix #,which, in general has a huge number of rows. Since the data are perturbed by noise, the involved matrix # is, in general, full rank. It is well known that, from a ....

P. Van Overschee and B. De Moor, Subspace Identification for Linear Systems: Theory, Implementation, Applications, Kluwer Academic Publishers, Dordrecht, 1996.

....system eigenstructure identification. 2 System theoretic issues and eigenstructure identification In this section, we present avariantofsubspace algorithms for output only, eigenstructure identification for MIMO systems. For general references on subspace algorithms, the reader is referred to [10,11,12,13].The variant we consider here is not using raw data as input, but rather covariance matrices from data, but it has been recognized that this is a minor modification. Also, compared with [10] wedonot consider weighting matrices. The reason is that is has been shown in [5] that weighting matrices ....

....for MIMO systems. For general references on subspace algorithms, the reader is referred to [10,11,12,13] The variant we consider here is not using raw data as input, but rather covariance matrices from data, but it has been recognized that this is a minor modification. Also, compared with [10], wedonot consider weighting matrices. The reason is that is has been shown in [5] that weighting matrices play no role when exact model order is known, whichwe assume here. The arguments presented here are well known [12] we just state them here for the sake of clarity and completeness. We are ....

Van Overschee, P. and B. De Moor (1996). Subspace Identification for Linear Systems : Theory -- Implementation -- Methods. Kluwer.

....the lack of accurate damping estimates and the determination of operational deflection shapes in stead of mode shapes, since no modal model is fitted to the data. Therefore we are looking for more advanced methods. The stochastic subspace identification method (SSI) is such a method (Van Overschee and De Moor, 1996). It identifies a stochastic state space model from output only measurements. The state space model is a very general model that is also suitable for our purposes: it can describe a linear vibrating structure excited by white M U(t) C 2 H U(t) KU(t) F(t) 1) x k 1 = Ax k w k y k = ....

....(Ljung, 1987) that tries to identify an ARMAV model. This paper presents a novel implementation of SSI, the so called reference based stochastic subspace identification (SSI ref) The key step of SSI is the projection of the row space of the future outputs into the row space of the past outputs (Van Overschee and De Moor, 1996). The idea is now to take instead of all past outputs only a set of well chosen reference sensors. This reduces the dimensions of the involved matrices and also the computation time, without having a negative influence on the quality of the estimated modal parameters. 2. VIBRATING STRUCTURES AND ....

[Article contains additional citation context not shown here]

Van Overschee, P. and De Moor, B., 1996, Subspace identification for linear systems: theory - implementation - applications. Kluwer Academic Publishers, Dordrecht, Netherlands.

....stochastic and unmeasurable ambient excitation, the traditional FRF based modal parameter estimation methods are excluded. Time domain system identification techniques that deal with output only data are more suitable. One of the most recent of these techniques is the stochastic subspace method [2,3]. This system identification method assumes that the dynamic behaviour of the system can be described by a state space model. The subspace method identifies the state space matrices. From these, the modal parameters can be extracted. However since the force acting on the structure remains unknown, ....

VAN OVERSCHEE, P. AND DE MOOR, B. Subspace identification for linear systems: theory - implementation - applications. Kluwer AcademicPublishers, Dordrecht, The Netherlands, 1996.

....the lack of accurate damping estimates and the determination of operational deflection shapes in stead of mode shapes, since no modal model is fitted to the data. Therefore we are looking for more advanced methods. The stochastic subspace identification method (SSI) is such a method (Van Overschee De Moor 1996). It identifies a stochastic state space model directly from the output only measurements without the need to convert the time signals to correlations or spectra. The state space model is a very general model that is also suitable for our purposes: it can describe a linear vibrating structure ....

....mass of the antennae. 3 SYSTEM IDENTIFICATION 3.1 Reference based stochastic subspace identification It is beyond the scope of this paper to explain in full detail the stochastic subspace identification method or the reference based variant. The interested reader is referred to literature: Van Overschee De Moor (1996) and Peeters De Roeck (1999) Here only the main ideas behind the method are given. The method assumes that the dynamic behaviour of a structure excited by white noise can be described by a stochastic state space model (1) A justification of this statement can for instance be found in Peeters ....

Van Overschee, P. & B. De Moor 1996. Subspace identification for linear systems: theory - implementation - applications. Dordrecht: Kluwer Academic Publishers.

.... 7 Bu k 7 w k y k 9 Cx k 7 Du k 7 v k , E[ w p v p w T q v T q ] 9 QS S T R pq (1) SYSTEM IDENTIFICATION Recently a lot of research effort in the system identification community was spent to subspace identification as evidenced by the book of Van Overschee De Moor [7] and the second edition of Ljung s book [8] Subspace methods identify a combined deterministic stochastic state space model from input output data by applying robust numerical techniques such as QR factorization, Singular Value Decomposition and Least Squares. Such a state space model is written ....

VANOVERSCHEE, P. ANDDEMOOR, B. Subspace identification for linear systems: theory - implementation - applications. Kluwer Academic Publishers: Dordrecht, The Netherlands, 1996.

No context found.

P. Van Overschee and B. De Moor, Subspace Identification for Linear Systems: Theory, Implementation, Applications. Kluwer Academic Publishers, 1996.

No context found.

P. Van Overschee, B. De Moor, Subspace Identification for Linear Systems: Theory - Implementation - Applications, Kluwer Academic Publishers, Boston/London/Dordrecht, (1996).

No context found.

P. Van Overschee and B. De Moor, Subspace Identification for Linear Systems: Theory -- Implementation -- Applications, Kluwer Academic Publishers, Boston (1996).

No context found.

Van Overschee P., De Moor B. Subspace identification for linear systems: Theory, Implementation, Applications. Kluwer Academic Publishers, 1996, 254 pp. Downloadable from http://www.esat.kuleuven.ac.be/ ~ sistawww/cgibin /pub.pl

No context found.

P. Van Overschee and B. De Moor, Subspace Identification for Linear Systems: Theory---Implementation---Applications. Boston, MA: Kluwer, 1996.

No context found.

P. Van Overschee and B. De Moor. Subspace Identification for Linear Systems: Theory, Implementation, Applications. Kluwer Academic Publishers, 1996.

....theory that is closely related to CCA and that was introduced by Shannon [18] in 1948. A slightly di#erent interpretation in terms of channel capacity and information rate is given in [17] Another area where CCA is applied, is stochastic realization and identification of dynamical models [1, 3, 5, 11, 12, 15, 16, 21, 22]. The order of the model and a state sequence can be derived from the canonical correlations and the canonical variates of the past and the future output data. Katrien De Cock is a research assistant at the K.U.Leuven. Dr. Bart De Moor is a full professor at the K.U.Leuven. Our research is ....

.... (CP 40: Sustainibility e#ects of Tra#c Management Systems) Direct contract research: Verhaert, Electrabel, Elia, Data4s, IPCOS; In this paper we will work with the geometric interpretation of canonical correlation analysis, as is usually done in the subspace identification literature, see e.g. [22]. The canonical correlations and the canonical variates are respectively equal to the cosines of the principal angles between and the principal vectors in two linear subspaces. These subspaces are the row spaces of block Hankel matrices obtained by stacking the measured input and output sequences. ....

[Article contains additional citation context not shown here]

P. Van Overschee and B. De Moor, Subspace Identification for Linear Systems: Theory -- Implementation -- Applications, Kluwer Academic Publishers, Boston, 1996. 15

.... a similarity transfor mation together with the estimated noise co variances matrices 8, so that the second order statistics of the output of the model and of the given output are equal (or equivalent in the sense of Faurre [3] In the last decade, so called subspace iden tification methods [7] have been developed. 1E denotes the expected value operator and 5pq the Kronecker delta. It is assumed that the process is stationary and ergodic: E[akb] limj o [ xJ a b T1 i=0 i J Typically, in a first step, Kalman filter state sequences i xJ and i 1 xJ of the system are estimated ....

....0]i I denotes the time indices of the outputs in the first column of Y01i 1 . The input data block Hankel matrix Uo]i 1 is constructed in a similar way. The calculations are performed in a numerically reliable way, based on the singular value decomposition (SVD) and QR decomposition (see [7] for details) After the estimation of the Kalman filter states sequences 7i and i ) ifi he first step, the system matrices (A, B, C, D) of the combined stochastic deterministic identifica tion problem are identified in the second step: min [i 1 (3) Motivated by consistency results [7] for j ....

[Article contains additional citation context not shown here]

P. Van Overschee and B. De Moor, Subspace Identification for Linear Systems: Theory, Implementation, Applications, Kluwer Academic Publishers, Boston/London /Dordrecht, 1996.

....and measurement noise w k 2 R n and v k 2 R l are assumed to be white, zero mean, Gaussian with unknown covariance matrices as given in (3) The model matrices A; C and the covariance matrices Q; S; R have appropriate dimensions. In the last decade, so called subspace identification methods [10] have been developed. Typically, in a first step, Kalman filter state sequences X i 2 R n Thetaj and X i 1 2 R n Thetaj of the system are estimated directly from inputoutput data using geometric operations of subspaces spanned by the column or row vectors of block Hankel matrices formed by ....

....denotes the time indices of the outputs in the first column of Y 0ji Gamma1 . The input data block Hankel matrix U 0ji Gamma1 is constructed in a similar way. The calculations are performed in a numerically reliable way, based on the singular value decomposition (SVD) and QR decomposition (see [10] for details) After the estimation of the Kalman filter states sequences X i and X i 1 in the first step, the system matrices ( A; C) of the stochastic identification problem are identified in the second step: min A; C fi fi fi fi fi fi fi fi X i 1 Y iji Gamma A ....

P. Van Overschee and B. De Moor, Subspace Identification for Linear Systems: Theory, Implementation, Applications, Kluwer Academic Publishers, Boston/London/Dordrecht, 1996.

No context found.

VAN OVERSCHEE,P.,AND DE MOOR, B. 1996. Subspace Identification for Linear Systems: Theory, Implementation, Applications.Kluwer Academic Publishers, Dordrecht, Netherlands.

No context found.

van Overschee, Peter and Bart de Moor (1996). Subspace identification for linear systems : theory, implementation, applications. Boston: Kluwer.

No context found.

P. Van Overschee and B. De Moor. Subspace Identification for Linear Systems: Theory, Implementation, Applications. Kluwer Academic Publishers, Dordrecht, 1996.

No context found.

P. Van Overschee and B. De Moor. Subspace Identification for Linear Systems: Theory, Implementation, Applications. Kluwer Academic Publishers, Dordrecht, 1996.

No context found.

P. Van Overschee and B. De Moor. Subspace identification for linear systems: theory, implementation, applications. Kluwer Academic Publishers, Dordrecht, 1996. 31

No context found.

P. Van Overschee and B. De Moor. Subspace Identification for Linear Systems: Theory, Implementation, Applications. Kluwer Academic Publishers, Dordrecht, 1996.

No context found.

P. Van Overschee and B. De Moor, Subspace Identification for Linear Systems: Theory, Implementation, Applications. Kluwer Academic Publishers, 1996.

No context found.

P. Van Overschee and B. De Moor, Subspace Identification for Linear Systems: Theory, Implementation, Applications. Kluwer Academic Publishers, 1996.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC