By Topic

Signal Processing, IEEE Transactions on

Issue 12 • Date Dec. 2003

Filter Results

Displaying Results 1 - 25 of 32
  • List of reviewers

    Page(s): 3295 - 3301
    Save to Project icon | Request Permissions | PDF file iconPDF (161 KB)  
    Freely Available from IEEE
  • Author index

    Page(s): 3302 - 3310
    Save to Project icon | Request Permissions | PDF file iconPDF (240 KB)  
    Freely Available from IEEE
  • Subject index

    Page(s): 3310 - 3338
    Save to Project icon | Request Permissions | PDF file iconPDF (383 KB)  
    Freely Available from IEEE
  • An efficient low-complexity technique for MLSE equalizers for linear and nonlinear channels

    Page(s): 3236 - 3248
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (700 KB) |  | HTML iconHTML  

    In this paper, a novel sequence equalizer, which belongs to the family of cluster-based sequence equalizers, is presented. The proposed algorithm achieves the maximum likelihood solution to the equalization problem in a fraction of computational load, compared with the classic maximum likelihood sequence estimation (MLSE) equalizers. The new method does not require the estimation of the channel impulse response. Instead, it utilizes the estimates of the cluster centers formed by the received observations. Furthermore, a new cluster center estimation scheme, which exploits the intrinsic dependencies among the cluster centers, is proposed. The new center estimation method exhibits enhanced performance with respect to convergence speed, compared with an LMS-based channel estimator. Moreover, this gain in performance is obtained at substantially lower computational load. The method is also extended in order to cope with nonlinear channels. The performance of the new equalizer is tested with several simulation examples, using both the quadrature phase shift keying (QPSK) and the 16-quadrature amplitude modulated (QAM) signaling schemes for linear and nonlinear communication channels. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive minimum symbol-error rate equalization for quadrature-amplitude modulation

    Page(s): 3263 - 3269
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (841 KB) |  | HTML iconHTML  

    We propose the adaptive minimum symbol-error rate algorithm, which is a low-complexity technique for adapting the coefficients of a linear equalizer in systems using pulse-amplitude or quadrature-amplitude modulation. The proposed algorithm very nearly minimizes error probability in white Gaussian noise and can significantly outperform the minimum-mean-squared error equalizer (by as much as 16 dB) when the number of equalizer coefficients is small relative to the severity of the intersymbol interference. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Filter design for MIMO sampling and reconstruction

    Page(s): 3164 - 3176
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (925 KB) |  | HTML iconHTML  

    We address the problem of finite impulse response (FIR) filter design for uniform multiple-input multiple-output (MIMO) sampling. This scheme encompasses Papoulis' generalized sampling and several nonuniform sampling schemes as special cases. The input signals are modeled as either continuous-time or discrete-time multiband input signals, with different band structures. We present conditions on the channel and the sampling rate that allow perfect inversion of the channel. Additionally, we provide a stronger set of conditions under which the reconstruction filters can be chosen to have frequency responses that are continuous. We also provide conditions for the existence of FIR perfect reconstruction filters, and when such do not exist, we address the optimal approximation of the ideal filters using FIR filters and a minmax l2 end-to-end distortion criterion. The design problem is then reduced to a standard semi-infinite linear program. An example design of FIR reconstruction filters is given. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Quantization of multiaspect scattering data: target classification and pose estimation

    Page(s): 3105 - 3114
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (799 KB) |  | HTML iconHTML  

    In many sensing scenarios, the observed scattered waveforms must be quantized for subsequent transmission over a communication channel. Rate-distortion theory plays an important role in defining the bit rate required to achieve a desired distortion. The distortion is typically defined in the context of signal reconstruction, with the goal of achieving high-fidelity synthesis of the compressed data. For sensing applications, however, the objective is often not simply signal reconstruction but classification performance as well. Other related metrics include target-pose estimation. We consider multiaspect wave scattering, in which classification and pose estimation are performed based on the quantized scattering data. Moreover, rate-distortion theory is employed to place bounds on pose-estimation performance when both the target identity and pose are unknown a priori. It is demonstrated that block-coding with Bayes-VQ may yield performance approaching the bound. Example results are presented for measured acoustic waveforms scattered from underwater elastic targets. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Channel estimation for multicarrier multiple input single output systems using the EM algorithm

    Page(s): 3280 - 3292
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (896 KB) |  | HTML iconHTML  

    This paper investigates the problem of blindly and semi-blindly acquiring the channel gains for an underdetermined synchronous multiuser multicarrier system. The special case of a multiple-input single-output (MISO) channel is considered where the different users transmit at the same time and in the same bandwidth. In order to separate the different users blindly, techniques exploiting the finite alphabet are used. For such techniques, and for a general underdetermined MIMO system, we study conditions under which the channel and the data for each user are blindly and semi-blindly identifiable. We consider the stochastic maximum likelihood (SML) criterion in which the unknown input symbols are modeled as discrete random variables. We apply the expectation-maximization (EM) algorithm in the frequency domain to get blind and semi-blind channel estimates for each user in the MISO case. We also present a recursive EM solution that updates the channel and noise estimates at each time instant. Simulations show that users can be separated, even at low SNR. Furthermore, semi-blind estimation allows for a more robust estimation solution since a possible singularity problem is avoided. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Variable-rate data sampling for low-power microsystems using modified Adams methods

    Page(s): 3182 - 3190
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (554 KB) |  | HTML iconHTML  

    A method for variable-rate data sampling is proposed for the purpose of low-power data acquisition in a small footprint microsystem. The procedure enables energy saving by utilizing dynamic power management techniques and is based on the Adams-Bashforth and Adams-Moulton multistep predictor-corrector methods for ordinary differential equations. Newton-Gregory backward difference interpolation formulae and past value substitution are used to facilitate sample rate changes. It is necessary to store only 2m+1 equispaced past values of t and the corresponding values of y, where y=g(t), and m is the number of steps in the Adams methods. For the purposes of demonstrating the technique, fourth-order methods are used, but it is possible to use higher orders to improve accuracy if required. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Formulation and comparison of two detectors of independent timing jitter in a complex harmonic

    Page(s): 3043 - 3052
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (768 KB) |  | HTML iconHTML  

    Two detectors of symmetrically distributed independent timing jitter in a data record composed of a complex harmonic in additive white Gaussian noise are proposed. The proposed detectors are computationally efficient, and although they are formulated using asymptotic results, they may be effectively used with small sample lengths under a wide range of conditions. The conditions required for consistency of the detectors are derived and examined for important special cases. The performances of the detectors are analyzed using simulations. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The DSFPN: a new neural network and circuit simulation for optical character recognition

    Page(s): 3198 - 3209
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (614 KB) |  | HTML iconHTML  

    A new type of neural network for recognition tasks is presented. The network, which is called the "dynamic supervised forward-propagation network" (DSFPN), is based on the forward only version of the counterpropagation network (CPN). The novel DSFPN is trained using a supervised algorithm and can grow dynamically during training, allowing allographs in the training data to be learned in an unsupervised manner. Training times are comparable with the CPN while giving better classification accuracies than the popular multilayer perceptron (MLP). Data preprocessed using Fourier descriptors show that, on average, the DSFPN is trained in 1353 times fewer presentations than the MLP networks and gives best recognition accuracy of 98.6%. Moreover, data preprocessed using wavelet multiresolution analysis gives a very high recognition accuracy; the best accuracy is 99.792%. Results show the effectiveness of the DSFPN and justify a hardware implementation to enable fast data classification. A circuit implementation for the DSFPN competitive middle layer is presented, and simulation results show that it can perform reliable pattern recognition at a rate of over 100 kHz. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Joint state and parameter estimation for a target-directed nonlinear dynamic system model

    Page(s): 3061 - 3070
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (655 KB) |  | HTML iconHTML  

    We present a new approach to joint state and parameter estimation for a target-directed, nonlinear dynamic system model with switching states. The model, recently proposed for representing speech dynamics, is called the hidden dynamic model (HDM). The model parameters, subject to statistical estimation, consist of the target vector and the system matrix (also called "time-constants"), as well as parameters characterizing the nonlinear mapping from the hidden state to the observation. We implement these parameters as the weights of a three-layer feedforward multilayer perceptron (MLP) network. The new estimation approach is based on the extended Kalman filter (EKF), and its performance is compared with the traditional expectation-maximization (EM) based approach. Extensive simulation results are presented using both approaches and under typical HDM speech modeling conditions. The EKF-based algorithm demonstrates superior convergence performance compared with the EM algorithm, but the former suffers from excessive computational loads when adopted for training the MLP weights. In all cases, the simulated model output converges to the given observation sequence. However, only in the case where the MLP weights or the target vector are assumed known do the time-constant parameters converge to their true values. We also show that the MLP weights never converge to their true values, thus demonstrating the many-to-one mapping property of the feedforward MLP. We conclude that, for the system to be identifiable, restrictions on the parameter space are needed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Natural frequency-based neural network approach to radar target recognition

    Page(s): 3191 - 3197
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (505 KB) |  | HTML iconHTML  

    Time domain response-based neural networks and frequency domain response-based neural networks have been proposed for radar target recognition. We propose a natural frequency-based neural network for radar target recognition. Our scheme takes advantage of an aspect angle independence of a natural frequency. It is shown from experimental results that a natural frequency based-neural network using the first natural frequency pair is superior to a time domain response-based neural network in the case of a single aspect angle and that a natural frequency based-neural network using the first natural frequency pair or the first two natural frequency pairs is superior to a time domain response-based neural network in the case of a multiple aspect angle. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Sampling theorems for uniform and periodic nonuniform MIMO sampling of multiband signals

    Page(s): 3152 - 3163
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (859 KB) |  | HTML iconHTML  

    We examine a multiple-input multiple-output (MIMO) sampling scheme for a linear time-invariant continuous-time MIMO channel. The input signals are modeled as multiband signals with different spectral supports, and the channel outputs are sampled on either uniform or periodic nonuniform sampling sets, with possibly different but commensurate intervals on the different outputs. This scheme encompasses Papoulis' generalized sampling and several nonuniform sampling schemes as special cases. We derive necessary and sufficient conditions on the channel and the sampling rate that allow stable perfect reconstruction of the inputs or, equivalently, perfect inversion of the channel. From an implementation viewpoint, we note that it is desirable that the reconstruction filters have continuous frequency responses. We derive necessary and sufficient conditions that guarantee this continuity property. The frequency responses of the reconstruction filters are specified as solutions to a system of linear equations. Finally, we demonstrate that perfect reconstruction may be possible, even when the channel outputs are sampled at an average rate that does not allow the reconstruction of any output from its samples alone. In certain instances, this average rate can achieve the recently presented fundamental bounds on MIMO sampling density. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Hermite neural network correlation and application

    Page(s): 3210 - 3219
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (552 KB) |  | HTML iconHTML  

    A method for obtaining the correlation of two Hermite neural networks is developed. The method is based on the fact that a Hermite function is unchanged by the Fourier transform, which allows an expression for the correlation to be obtained directly from the network weights without the need for the Fourier transform. Comparative results with other neural network correlation methods are presented on simulated radar signals. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Stochastic time-frequency analysis using the analytic signal: why the complementary distribution matters

    Page(s): 3071 - 3079
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (863 KB) |  | HTML iconHTML  

    We challenge the perception that we live in a "proper world", where complex random signals can always be assumed to be proper (also called circularly symmetric). Rather, we stress the fact that the analytic signal constructed from a nonstationary real signal is, in general, improper, which means that its complementary correlation function is nonzero. We explore the consequences of this finding in the context of stochastic time-frequency analysis in Cohen's class. There, the analytic signal plays a prominent role because it reduces interference terms. However, the usual time-frequency representation (TFR) based on the analytic signal gives only an incomplete signal description. It must be augmented by a complementary TFR whose properties we develop in detail. We show why it is still advantageous to use the pair of standard and complementary TFRs of the analytic signal rather than the TFR of the corresponding real signal. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A class of regular biorthogonal linear-phase filterbanks: theory, structure, and application in image coding

    Page(s): 3220 - 3235
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2192 KB) |  | HTML iconHTML  

    This paper discusses a method of regularity imposition onto biorthogonal linear-phase M-band filterbanks using the lattice structure. A lifting structure is proposed for lattice matrix parameterization where regularity constraints can be imposed. The paper focuses on cases with analysis and synthesis filterbanks having up to two degrees of regularity. Necessary and sufficient conditions for regular filterbanks in terms of the filter impulse response, frequency response, scaling function, and wavelets are revisited and are derived in terms of the lattice matrices. This also leads to a constraint on the minimum filter length. Presented design examples are optimized for the purpose of image coding, i.e., the main objectives are coding gain and frequency selectivity. Simulation results from an image coding application also show that these transforms yield improvement in the perceptual quality in the reconstruction images. The approach has also been extended to the case of integer/rational lifting coefficients, which are desirable in many practical applications. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The KaGE RLS algorithm

    Page(s): 3094 - 3104
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (662 KB) |  | HTML iconHTML  

    A new fast recursive least squares (RLS) algorithm, the Kalman gain estimator (KaGE), is introduced. By making use of RLS interpolation as well as prediction, the algorithm generates the transversal filter weights without suffering the poor numerical attributes of the fast transversal filter (FTF) algorithm. The Kalman gain vector is generated at each time step in terms of interpolation residuals. The interpolation residuals are calculated in an order recursive manner. For an Nth-order problem, the procedure requires O(Nlog2N) operations per iteration. This is achieved via a divide-and-conquer approach. Computer simulations suggest that the new algorithm is numerically robust, running successfully for many millions of iterations. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the phase condition and its solution for Hilbert transform pairs of wavelet bases

    Page(s): 3293 - 3294
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (204 KB) |  | HTML iconHTML  

    In this paper, the phase condition on the scaling filters of two wavelet bases that renders the corresponding wavelets as Hilbert transform pairs is studied. An alternative and equivalent phase condition is derived. With the equivalent condition and using Fourier series expansions, we show that the solution for which the corresponding scaling filters are offset from one another by a half sample is the only solution satisfying the phase condition. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Stochastic resonance in locally optimal detectors

    Page(s): 3177 - 3181
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (333 KB) |  | HTML iconHTML  

    The aim of the paper is to show that the nonlinear effect known as stochastic resonance, which corresponds to the improvement of the processing of information by noise, occurs naturally in some detection problems. We illustrate this by studying the problem of detecting a small amplitude sinusoid in non-Gaussian noise. We show that in some cases, the nonlinearity that appears in locally optimal detectors can be viewed as a stochastic resonator. If the parameters of the locally optimal detector (LOD) are not well tuned, the performance can be improved by the addition of noise. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Analysis of post-combiner equalizers in cosine-modulated filterbank-based transmultiplexer systems

    Page(s): 3249 - 3262
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1167 KB) |  | HTML iconHTML  

    In this paper, we present an analysis of the post-combiner equalizers used to compensate for channel distortion in cosine-modulated-based transmultiplexer systems. Such equalizers have been widely used in multicarrier modulation (MCM) systems that use cosine modulated filterbanks (CMFB) for signal modulation and demodulation. By making the reasonable assumption that the number of subchannels in the system is large enough such that each subchannel can be approximated by a constant complex gain, we derive close-form equations for optimum post-combiner transfer functions. These transfer functions are found to be finite impulse response (FIR) and closely related to the prototype filter in CMFB analysis and synthesis blocks. Moreover, by making use of the results of our analysis, we propose a new post-combiner structure in which the number of adaptive parameters are about one third of those in the earlier reports. This reduction in the number of adaptive parameters results in a three-fold increase in convergence speed of the least mean square (LMS) algorithm when it is used for adaptation of the post combiners. We also present a convergence analysis of the post-combiner when the LMS algorithm is used for adaptation. We study the correlation matrix whose eigenvalues determine convergence behavior of the LMS algorithm. We show that all eigenvalues of this matrix are equal. This implies that in the case of the post-combiner, the LMS algorithm experiences no slow mode of convergence, as is the case in many other applications. Computer simulations that corroborate our theoretical findings are also presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Markovian source separation

    Page(s): 3009 - 3019
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (701 KB) |  | HTML iconHTML  

    A maximum likelihood (ML) approach is used to separate the instantaneous mixtures of temporally correlated, independent sources with neither preliminary transformation nor a priori assumption about the probability distribution of the sources. A Markov model is used to represent the joint probability density of successive samples of each source. The joint probability density functions are estimated from the observations using a kernel method. For the special case of autoregressive models, the theoretical performance of the algorithm is computed and compared with the performance of second-order algorithms and i.i.d.-based separation algorithms. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An algorithm for unit-norm equation error system identification based on the method of multipliers

    Page(s): 3080 - 3085
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (442 KB) |  | HTML iconHTML  

    It is known that the unit-norm constraint for equation-error based system identification is superior to the monic constraint since it produces consistent estimates for white measurement noise and also presents better approximation properties in reduced-order cases. Here, a new algorithm for unit-norm equation-error adaptive filtering is proposed. This scheme is inspired by the constrained optimization technique known as the method of multipliers. An analysis of stationary points and mean convergence properties is developed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Blind, adaptive channel shortening by sum-squared auto-correlation minimization (SAM)

    Page(s): 3086 - 3093
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (600 KB) |  | HTML iconHTML  

    We propose a new blind, adaptive channel shortening algorithm for updating the coefficients of a time-domain equalizer in a system employing multicarrier modulation. The technique attempts to minimize the sum-squared auto-correlation terms of the effective channel impulse response outside a window of desired length. The proposed algorithm, known as "sum-squared auto-correlation minimization" (SAM), requires the source sequence to be zero-mean, white, and wide-sense stationary, and it is implemented as a stochastic gradient descent algorithm. Simulation results are provided, demonstrating the success of the SAM algorithm in an asymmetric digital subscriber loop (ADSL) system. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Bearing estimation for a distributed source of nonconstant modulus ounds and analysis

    Page(s): 3027 - 3035
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (672 KB)  

    The paper considers the problem of estimating the bearing of a single, far-field source, surrounded by local scatterers, using passive sensor array measurements. An associated source-bearing estimation problem is formulated, and the Crame´r-Rao lower bound is evaluated. In addition, a comprehensive analysis is performed of the maximum likelihood estimates that (due to mismodeling) assume a constant modulus source, and the degradation in performance is quantified as a function of the source's empirical variance. It is shown that, for a limited price in terms of mean square error, the constant modulus maximum likelihood estimator can replace the optimal nonconstant modulus estimator. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.

Aims & Scope

IEEE Transactions on Signal Processing covers novel theory, algorithms, performance analyses and applications of techniques for the processing, understanding, learning, retrieval, mining, and extraction of information from signals

Full Aims & Scope

Meet Our Editors

Editor-in-Chief
Zhi-Quan (Tom) Luo
University of Minnesota