HomeChair: Lang Tong, University of Connecticut, USA
Home
Channarong Tontiruttananon, Auburn University (U.S.A.)
Jitendra K. Tugnait, Auburn University (U.S.A.)
The problem of closed-loop system identification given noisy input-output measurements is considered. The closed-loop system operates under an external cyclostationary input which is not measured. Noisy measurements of the (direct) input and output of the plant are assumed to be available. The various disturbances affecting the system are either stationary or cyclostationary with cycle frequencies different from the input cycle frequencies. The closed-loop system must be stable but it is allowed to be unstable in open-loop. A frequency-domain parametric solution is proposed and analyzed using an equation error formulation, and cyclic spectrum and cross-spectrum of the input-output measurements. The parameter estimator is shown to be consistent. A simulation example using an unstable open-loop system is presented to illustrate the proposed approach.
ic981108.pdf (From Postscript)
TOPXiang-Gen Xia, University of Delaware (U.S.A.)
In this paper, we present a method to identify channels with both Doppler and time shifts using mixed training signals. The training signals we use consist of two parts, where one part is a constant and the other part is a conventional training signal, such as a pseudo-random signal. These two parts may be separated in either the time or the frequency domain. We provide a necessary and sufficient condition on the channel identifiability in terms of the time and Doppler shifts when the mixed training signals are used.It can be shown that the condition holds almost surely in most cases of interests in practice. Some numerical examples are also presented.
ic981394.pdf (From Postscript) TOPChristine Serviere, CEPHAG-ENSIEG (France)
We focus on the feasibility of the source separation in the frequency domain. First, it is linked with the convergence speed towards gaussianity of signals after L-point discrete Fourier Transform. We test here a distance to gaussianity thanks to the spectral kurtosis. We analyse the influence of L, of the duration of the source tricorrelations and of a non linear filtering. We mainly develop the case of QARMA processes. The second point consists in the reconstruction of the spectra of the estimated sources from the signals identified at each frequency bin. Indeed, the source associated to the ith identified signal is not necessarily the same from one frequency bin to another. The algorithm efficiency is then illustrated on QARMA processes, including the procedures of separation and reconstruction.
ic981567.pdf (Scanned) TOP
Anisse Taleb, INPG-LIS (France)
Christian Jutten, INPG-LIS (France)
Serge Olympieff, INPG-LIS (France)
This paper proposes a new approach for sources separation in special nonlinear mixtures, called post nonlinear mixtures (PNL). We first explain the nice separability properties of these mixtures: solutions have almost the same indeterminacies than separation of sources in instantaneous linear mixtures. The method proposed in this paper is based on the minimization of the mutual information, which needs the knowledge of source distributions or more exactly of log-derivative of source distributions (the so-called score functions). The algorithm consists of three adaptive blocks: one nonlinear block, devoted to adaptive estimation of source score functions, drives the adaptation of the two other blocks corresponding to estimation of the linear and nonlinear parts of the mixtures. The paper finishes with a few experimental results which prove the efficacy of the algorithm.
ic981770.pdf (From Postscript) TOP
Sergio A. Cruces-Alvarez, University of Seville (Spain)
Luis Castedo-Ribas, University of A Coruna (Spain)
In this paper we present several Gauss-Newton algorithms for blind source separation of convolutive mixtures. The algorithms can be interpreted as generalizations of two previous approaches due to Gerven-Compernolle and Nguyen-Jutten. Since they are of the Gauss-Newton type, they exhibit a fast rate of convergence. Also, we present a stability analysis for two sources and instantaneous mixtures where we show that the algorithms cannot converge to non-separating solutions.
ic981914.pdf (From Postscript) TOP