Laurent Mevel

Learn More
We consider a hidden Markov model with multidimen-sional observations, and with misspecification, i.e. the assumed coefficients (transition probability matrix, and observation conditional densities) are possibly different from the true coefficients. Under mild assumptions on the coefficients of both the true and the assumed models, we prove that : (i) the(More)
We consider a hidden Markov model (HMM) with mul-tidimensional observations, and where the coeecients (transition probability matrix, and observation conditional densities) depend on some unknown parameter. We study the asymptotic behaviour of two recursive es-timators, the recursive maximum likelihood estimator (RMLE), and the recursive conditional least(More)
— In this paper we study " nonstationary consistency " of subspace methods for eigenstructure identification, i.e., the ability of subspace algorithms to converge to the true eigenstruc-ture despite nonstationarities in the excitation and measurement noises. Note that such nonstationarities may result in having time-varying zeros for the underlying system,(More)
In this paper a new recursive output-only identification algorithm is proposed based on stochastic realization, a classical covariance driven subspace identification technique. The recursive algorithm yields results that are in close agreement with those of its non-recursive counterpart. Furthermore, the relatively low complexity of the algorithm makes it(More)
Recommended by George Moustakides This paper reports on the theory and practice of covariance-driven output-only and input/output subspace-based identification and detection algorithms. The motivating and investigated application domain is vibration-based structural analysis and health monitoring of mechanical, civil, and aeronautic structures.
The problem of structural model identification under both known and unknown input excitations, is addressed. In-flight data analysis is an important instance of that problem. Input/output and output-only eigenstructure identification methods are described and compared, within two classes of methods: subspace-based and prediction error. In particular,(More)
Subspace identiication algorithms have proven eecient for performing output-only identiication of the eigenstructure of a linear MIMO system subject to uncontrolled, unmeasured, and nonstationary excitation. Such a problem arises in mechanical engineering, for modal analysis of vibrating structures and machines. A common practice there is to collect data(More)