Carlos Pantaleón

Blind equalization of constant modulus signals using support vector machines

In this paper, the problem of blind equalization of constant modulus (CM) signals is formulated within the support vector regression (SVR) framework. The quadratic inequalities derived from the CM property are transformed into linear ones, thus yielding a quadratic programming (QP) problem. Then, an iterative reweighted procedure is proposed to blindly restore the CM property. The technique is suitable for real and complex modulations, and it can also be generalized to nonlinear blind equalization using kernel functions. We present simulation examples showing that linear and nonlinear blind SV equalizers offer better performance than cumulant-based techniques, mainly in applications when only a small number of data samples is available, such as in packet-based transmission over fast fading channels.

Stochastic blind equalization based on PDF fitting using Parzen estimator

This work presents a new blind equalization approach that aims to force the probability density function (pdf) at the equalizer output to match the known constellation pdf. Quadratic distance between pdfs is used as the cost function to be minimized. The proposed method relies on the Parzen window method to estimate the data pdf and is implemented by a stochastic gradient descent algorithm. The kernel size of the Parzen estimator allows a dual mode switch or a soft switch between blind and decision-directed equalization. The proposed method converges faster than the constant modulus algorithm (CMA) working at the symbol rate, with a similar computational burden, and reduces the residual error of the CMA in multilevel modulations at the same time. A comparison with the most common blind techniques is presented.

A new EM-based training algorithm for RBF networks

In this paper, we propose a new Expectation-Maximization (EM) algorithm which speeds up the training of feedforward networks with local activation functions such as the Radial Basis Function (RBF) network. In previously proposed approaches, at each E-step the residual is decomposed equally among the units or proportionally to the weights of the output layer. However, these approaches tend to slow down the training of networks with local activation units. To overcome this drawback in this paper we use a new E-step which applies a soft decomposition of the residual among the units. In particular, the decoupling variables are estimated as the posterior probability of a component given an input-output pattern. This adaptive decomposition takes into account the local nature of the activation function and, by allowing the RBF units to focus on different subregions of the input space, the convergence is improved. The proposed EM training algorithm has been applied to the nonlinear modeling of a MESFET transistor.

Underdetermined blind source separation in a time-varying environment

The problem of estimating n source signals from m measurements that are an unknown mixture of the sources is known as blind source separation. In the underdetermined —less measurements than sources— linear case, the solution process can be conveniently divided in three stages: represent the signals in a sparse domain, find the mixing matrix, and estimate the sources. In this paper we adhere to that approach and parametrize the performance of these stages as a function of the sparsity of the signals. To find the mixing matrix and track its variations in the dynamic case a nonparametric maximum-likelihood approach based on Parzen windowing is presented. To invert the underdetermined linear problem we present an estimator that chooses the “best” demixing matrix in a sample by sample basis by using some previous knowledge of the statistics of the sources. The results are validated by Montecarlo simulations.

Bayesian estimation of chaotic signals generated by piecewise-linear maps

Chaotic signals are potentially attractive in a wide range of signal processing applications. This paper deals with Bayesian estimation of chaotic sequences generated by piecewise-linear (PWL) maps and observed in white Gaussian noise. The existence of invariant distributions associated with these sequences makes the development of Bayesian estimators quite natural, Both maximum a posteriori (MAP) and minimum mean square error (MS) estimators are derived. Computer simulations confirm the expected performance of both approaches, and show how the inclusion of a priori information produces in most cases an increase in performance over the maximum likelihood (ML) case.

Modeling nonlinear power amplifiers in OFDM systems from subsampled data: a comparative study using real measurements

A comparative study among several nonlinear high-power amplifier (HPA) models using real measurements is carried out. The analysis is focused on specific models for wideband OFDM signals, which are known to be very sensitive to nonlinear distortion. Moreover, unlike conventional techniques, which typically use a single-tone test signal and power measurements, in this study the models are fitted using subsampled time-domain data. The in-band and out-of-band (spectral regrowth) performances of the following models are evaluated and compared: Salehs model, envelope polynomial model (EPM), Volterra model, the multilayer perceptron (MLP) model, and the smoothed piecewise-linear (SPWL) model. The study shows that the SPWL model provides the best in-band characterization of the HPA. On the other hand, the Volterra model provides a good trade-off between model complexity (number of parameters) and performance.

Nonlinearity estimation in power amplifiers based on undersampled temporal data

In this paper we apply undersampling techniques to capture the temporal input-output relationship of RF power amplifiers. This approach avoids the distortion introduced by the upconverter and downconverter stages. We develop polynomial models with memory from the available data and evaluate its performance estimating device parameters like Adjacent power Ratio (ACPR) and AM-AM curves. The estimated parameters show good agreement with the empirical ones.

Neural networks for large- and small-signal modeling of MESFET/HEMT transistors

In this paper, we present a comparative study of three neural networks-based solutions for large- and small-signal modeling of MESFET and HEMT transistors. The first two neural architectures are specific for this modeling problem: the generalized radial basis function (GRBF) network, and the smoothed piecewise linear (SPWL) model. These models are compared with the well-known multilayer perceptron (MLP) network. Results are presented for both the large- and small-signal regimes separately. Finally, a global model is proposed that is able to accurately characterize the whole behavior of the transistors. This model is based on a simple combination of the best models obtained for the two kinds of regimes.

Deconvolution of seismic data using adaptive Gaussian mixtures

Based on a Gaussian mixture model for the reflectivity sequence, the authors present a new technique for blind deconvolution of seismic data. The method obtains a deconvolution filter that maximizes at its output a measure of the relative entropy between the proposed Gaussian mixture and a pure Gaussian distribution. A new updating procedure for the mixture parameters is included in the algorithm: it allows one to apply the algorithm without any prior knowledge about the signal and noise. A simulation example illustrates the performance of the proposed method.

A fast blind SIMO channel identification algorithm for sparse sources

We address the blind identification of single-input-multiple output (SIMO) finite impulse response systems when the input signal is sparse. The problem is equivalent to underdetermined blind source separation (BSS), but with temporal correlation among the sources. Exploiting the sparse character of the input signal, the algorithm solves three different problems: first, to estimate the directions of the columns of the channel matrix; second, to estimate the L/sub 2/-norm of the columns; and finally, to find the correct ordering of the columns of the mixing matrix. The last step is not required for the blind source separation (BSS) problem, since any permutation of the columns is admissible for BSS. The performance and computational cost of the algorithm in a noiseless situation is compared against subspace-based techniques.