Home Home
João Xavier, IST - ISR (Portugal)
Victor Barroso, IST - ISR (Portugal)
In this paper we address the problem of the blind reconstruction of binary sequences transmitted by $P$ sources, sharing simultaneously the same frequency channel. This is the situation in Space Division Multiple Access (SDMA) systems, where an array receiver is used to discriminate the spatial signatures of the sources. We exploit the geometry induced by the finite nature of the sources alphabet in order to derive a blind multichannel inversion algorithm. This algorithm minimizes a generalized constant modulus (GCM) cost function, subject to the constraints imposed by the signal space geometry (geometric priors). We define the GCM cost function for the case of additive Gaussian noise channels. We show that, except for a sign ambiguity, the minimization of the GCM criterion has a unique solution. Therefore, convergence to the global minimum is guaranteed. Computer simulations illustrate the performance achieved with the proposed method.
ic973437.pdf ic973437.pdf
TOP
Daniel Iglesia, Universidad de La Coruña (Spain)
Adriana Dapena, Universidad de La Coruña (Spain)
Luis Castedo, Universidad de La Coruña (Spain)
This paper explains how fractional cyclic moments of CPFSK signals can be used in adaptive beamforming. It is shown that CPFSK signals have non-zero fractional cyclic moments because they generate spectral lines when they are raised to the inverse of its modulation index, which is generally a fractional number. To exploit this property, an optimization criterion is proposed to compute the antenna coefficients that maximize the output SINR. The resulting technique is blind because it does not need to know the transmitted symbols: only the carrier frequency, the symbol rate and the modulation index is required.
ic973441.pdf ic973441.pdf TOP
Alle-Jan van der Veen, Delft University of Technology (The Netherlands)
Jeroen Tol, Delft University of Technology (The Netherlands)
Zero/constant modulus (ZCM) signals are complex signals for which every sample is either zero or has modulus 1. Such signals arise after imprecise carrier-to-baseband conversion of binary (0,1)-modulated signals, or with intermittent phase-modulated signals. We consider the separation of linear superpositions of such signals using analytic techniques. An application is the separation of multiple partly coinciding aircraft transponder signals (SSR reply signals).
ic973445.pdf ic973445.pdf TOPJean-François Cardoso, CNRS/ENST (France)
This paper proposes a unifying view of source separation via the concepts of `estimating function' and `estimating equation'. We exhibit the estimating functions corresponding to various known techniques like ICA, JADE, infomax, maximum likelihood, cumulant matching, etcldots We also show how equivariant batch and adaptive algorithms stem from each particular estimating function.
ic973449.pdf ic973449.pdf TOP
Pierre Comon, Thomson Marconi (France)
Eric Moreau, ISITV (France)
Contrast-based separation of sources have a number of advantages. Among others, they are optimal (in a precise sense) in presence of noise of unknown statistics. Here a new contrast is proposed that allows not only to obtain the optimal solution analytically, but also yields better performances in terms of variance of the estimated mixing matrix. This contrast needs source kurtosis to have the same sign, and is thus appropriate to multichannel blind equalization in communications.
ic973453.pdf ic973453.pdf TOP
Santiago Zazo, University of Alfonso X, Madrid (Spain)
José M. Páez-Borrallo, Univ. Politécnica de Madrid (Spain)
Recent works have presented novel techniques that exploit cyclostationarity for channel identification (equalization) in data communication systems using only second order statistics. In particular, it has been shown the feasibility of blind identification based on the 'forward shift' structure of the correlation matrices of the source. In this paper we propose an alternative algorithm based on this property but with an improved choice of the autocorrelation matrices to be considered. This new representation of the equalization problem provide a cost function formulated as a generalized Rayleigh quotient, which may be efficiently implemented by using conjugate gradient techniques. Several simulations over different data transmission constellations support our theoretical analysis.
ic973457.pdf ic973457.pdf TOP
Lars K. Hansen, University of Texas at Austin (U.S.A.)
Guanghan Xu, University of Texas at Austin (U.S.A.)
When sequentially separating a linear combination of co-channel digital signals, it is necessary at each step to test the validity of the currently estimated signal prior to proceeding to extract the next one. We describe a procedure for use with sequential algorithms which uses a deflation-based approach combined with a simple test statistic. The deflation step removes the contributions of the currently identified signals. The simple test statistic takes into account the error terms introduced into the data by the deflation. The method has been successfully applied in an existing sequential estimation algorithm.
ic973461.pdf ic973461.pdf TOP
V.U. Reddy, Stanford University (U.S.A.)
Constantinos B. Papadias, Stanford University (U.S.A.)
Arogyaswami Paulraj, Stanford University (U.S.A.)
Recently, a number of classes of multipath channels which are not blindly identifiable from fractionally spa- ced samples and second-order cyclic spectra have been presented. In this paper, we consider the blind identification problem of these channels using multiple antennas and show that they will not in general give rise to any common roots among the sub-channels formed from the antennas, and hence, they can be identified from second-order statistics.
ic973465.pdf TOP
Adel Belouchrani, Villanova University (U.S.A.)
Moeness G. Amin, Villanova University (U.S.A.)
This paper addresses the problem of the blind source separation which consists of recovering a set of signals of which only instantaneous linear mixtures are observed. A blind source separation approach exploiting the difference in the time-frequency (t-f) signatures of the sources is considered. The approach is based on the diagonalization of a combined set of `spatial time-frequency distributions'. Asymptotic performance analysis of the proposed method is performed. Numerical simulations are provided to demonstrate the effectiveness of our approach and to validate the theoretical expression of the asymptotic performance.
ic973469.pdf ic973469.pdf TOP
Giacinto Gelli, II Univ. Napoli (Italy)
Luigi Paura, II Univ. Napoli (Italy)
The paper deals with signal extraction performed by processing data received by an array of sensors. The proposed method is blind, i.e., it does not require any a priori knowledge of directional information associated with the signals of interest (SOI's). Such information is obtained directly from the received data by exploiting the higher-order cyclostationarity (HOCS) properties exhibited by most communication signals. The proposed method is inherently tolerant to both possibly non-stationary Gaussian disturbances as well as non-Gaussian interferences not exhibiting the same HOCS properties presented by the SOI's. Simulation results confirm the effectiveness of the method when it operates in severely degraded disturbance environments.
ic973473.pdf ic973473.pdf TOP
Mati Wax, RAFAEL (Israel)
Yosef Anu, RAFAEL (Israel)
We present a new two-step approach to blind beamforming based on the least squares criterion. The first step consists of "whitening" the array received vector, i.e., transforming its response matrix to some unknown unitary matrix. The second step consists of estimating the unitary matrix from the fourth order cumulants by a least squares criterion. In contrast to the corresponding "joint diagonalization" step of the JADE algorithm, our second step exploits all the structural information in the problem and consequently yields better performance. Simulation results demonstrating the improved performance over the JADE algorithm are included.
ic973477.pdf TOP