Showing papers by "Joos Vandewalle published in 1988"
TL;DR: A multiprocessor architecture is presented that is suited for the customized and automated VLSI realization of complex low-to-medium-speed digital signal processing applications and allows for an efficient hardware realization utilizing the inherent parallelism of a particular application.
Abstract: A multiprocessor architecture is presented that is suited for the customized and automated VLSI realization of complex low-to-medium-speed digital signal processing applications. The proposed architecture is constructed from a set of flexible and parameterizable data paths, a selection of powerful control units (for decision-making tasks), and a number of protocols for fast interprocessor communication. The flexible nature of this system allows for an efficient hardware realization utilizing the inherent parallelism of a particular application. The effectiveness of the approach is substantiated with the synthesis of several test vehicles, such as a pitch-extraction algorithm for speech, in terms of the defined architecture. >
80 citations
TL;DR: The authors describe the properties of nongeneric TLS problems and prove that the proposed generalization remains optimal with respect to the TLS criteria for any number of observation vectors in B if additional constraints are imposed.
Abstract: Total least squares (TLS) is one method of solving overdetermined sets of linear equations $AX \approx B$ that is appropriate when there are errors in both the observation matrix B and the data matrix A. Golub and Van Loan (G. H. Golub and C. F. Van Loan, SIAM J. Numer. Anal., 17 (1980), pp. 883–893) introduced this method into the field of numerical analysis and developed an algorithm based on the singular value decomposition. However in some TLS problems, called nongeneric, their algorithm fails to compute a finite TLS solution. This paper generalizes their TLS computations in order to solve these nongeneric TLS problems. The authors describe the properties of those problems and prove that the proposed generalization remains optimal with respect to the TLS criteria for any number of observation vectors in B if additional constraints are imposed. Finally, the TLS computation is summarized in one algorithm which includes the proposed generalization.
75 citations
TL;DR: In this paper, the identification of models for multivariable linear time invariant systems from noisy input-output observations is studied from a geometrical point of view, where the singular value decomposition allows to estimate the order of observable part of the system and its state space model matrices.
58 citations
11 Apr 1988
TL;DR: It is shown how the singular value decomposition allows the order of the observable part of the system and its state-space model matrices to be estimated and an example with an industrial plant identification is presented.
Abstract: Some geometrically inspired concepts are studied for the identification of models for multivariable linear time-invariant systems from noisy input-output observations. Starting from a fundamental highly structured input-output matrix equation, it is shown how the singular value decomposition allows the order of the observable part of the system and its state-space model matrices to be estimated. Moreover, conditions for persistency of excitation of the inputs and the behavior of the algorithm when the data are perturbed by noise can easily be studied from a geometrical point of view. The singular values allow these concepts to be quantified. An example with an industrial plant identification is presented. >
46 citations
TL;DR: A new and accurate method is described for the accurate characterization of worst-case limit-cycle behaviour and other finite word-length effects in digital filters based on a simulated annealing procedure with a general scope to keep the required CPU times reasonable.
Abstract: The accurate characterization of worst-case limit-cycle behaviour and other finite word-length effects in digital filters constitutes a challenging and important optimization application. Indeed, for a chip implementation this can lead to a significant reduction in the signal word-length and hence to considerable savings in terms of the final chip area. This paper describes a new and accurate method for this investigation which is based on a simulated annealing procedure with a general scope. Novel methods for dynamically and efficiently updating the essential parameters of the annealing schedule keep the required CPU times reasonable.
The quality of the results obtained with our optimization routine and the general applicability of our approach are further substantiated with some promising results for the quantization of coefficients in arbitrary digital filters.
40 citations
TL;DR: In this article, a method for inverting square singular systems is presented, which explores the possibility of singular-system representation, and provides a simple and closed form for the constructed inverse.
Abstract: A method for inverting square singular systems is presented. The construction explores the possibility of singular-system representation, and provides a simple and closed form for the constructed inverse. One important property is that the minimality of the inverse (in the sense of Verghese) is equivalent to that of the original system. >
38 citations
IMEC1
TL;DR: The general applicability of the approach is further substantiated with some promising results for the solution of clustered TSP problems with many cities, and hierarchial partitioning is established as an important as a important property for distingushing applications which are very well-suited for simulated annealing from those which are not.
37 citations
TL;DR: In this paper, the implementation of a high-performance data encryption standard (DES) data encryption chip is presented, which is a single chip of 25 mm/sup 2/ in 3- mu m double-metal CMOS.
Abstract: The implementation of a high-performance data encryption standard (DES) data encryption chip is presented. At the system design level, cryptographical optimizations and equivalence transformations lead to a very efficient floorplan with minimal routing, which otherwise would present a serious problem for data-scrambling algorithms. These optimizations, which do not compromise the DES algorithm nor the security, are combined with a highly structured design and layout strategy. Novel CAD tools are used at different steps in the design process. The result is a single chip of 25 mm/sup 2/ in 3- mu m double-metal CMOS. Functionality tests show that a clock of 16.7 MHz can be applied, which means that a 32-Mb/s data rate can be achieved for all eight byte modes. This is the fastest DES chip reported yet, allowing equally fast execution of all four DES modes of operation, due to an original pipeline architecture. >
35 citations
01 Apr 1988
TL;DR: This paper describes a successful hardware implementation of the RSA algorithm as an 120-bit bit-slice processor, which may be interconnected without additional circuitry to obtain arbitrary word lengths.
Abstract: This paper describes a successful hardware implementation of the RSA algorithm. It is implemented as an 120-bit bit-slice processor, which may be interconnected without additional circuitry to obtain arbitrary word lengths. With 512-bit operands, exponentiation takes less than 30 milliseconds.
24 citations
TL;DR: The Partial Total Least Squares (PTLS) subroutine as mentioned in this paper solves the PTLS problem AX ≈ B by using a Partial Singular Value Decomposition (PSVD), thereby improving the computational efficiency of the classical TLS algorithm.
23 citations
Journal Article•
TL;DR: In this paper, a successful hardware implementation of the RSA algorithm is described, implemented as an 120-bit bit-slice processor, which may be interconnected without additional circuitry to obtain arbitrary word lengths.
Abstract: This paper describes a successful hardware implementation of the RSA algorithm. It is implemented as an 120-bit bit-slice processor, which may be interconnected without additional circuitry to obtain arbitrary word lengths. With 512-bit operands, exponentiation takes less than 30 milliseconds.
11 Apr 1988
TL;DR: A structure for the subband acoustic echo canceler is proposed that avoids filter banks in the sending line and the consequent delay and an additional adaptive filter is included.
Abstract: A structure for the subband acoustic echo canceler is proposed that avoids filter banks in the sending line and the consequent delay. In this arrangement of the filter banks, the conventional least-mean-square algorithm is modified and an additional adaptive filter is included. Simulation results are given, and a real application is reported. >
01 Jan 1988
01 Jan 1988
07 Jun 1988
TL;DR: It is proved that the authors' algorithm performs better than the conventional LMS algorithm in a nonstationary environment.
Abstract: The performance of a least-mean square (LMS) adaptive filter with respect to the feedback factor mu in both stationary and nonstationary environments are outlined and investigated by simulations. A mu -vector LMS algorithm that is an extension of the conventional LMS algorithm is proposed. In the proposed algorithm, different values of feedback factor are used for different taps. Computer simulation results are given and it is proved that the authors' algorithm performs better than the conventional LMS algorithm in a nonstationary environment. Using the proposed algorithm, a two-stage adaptive system is suggested, the first stage of which is an adaptive line enhancer. The application of the method to biomedical signal processing is discussed. >
TL;DR: In this paper, it was shown that any complex rational matrix G(s) can be factorized over the complex field as a product of n complex rational matrices each of which has McMillan degree 1, where n is the maximum degree of the matrix.
Abstract: It is shown that any complex rational matrix G(s) can be factorized over the complex field as a product of n complex rational matrices each of which has McMillan degree 1, where n is the McMillan degree of G(s). When confined to the real field, G(s) with real coefficient can be factorized as a product of a sequence of real national matrices with McMillan degree 1 or 2, where the McMillan degrees of the factors add up to n. >
TL;DR: In this article, a geometrically inspired algorithm is derived for identification of state space models for multivariable linear time-invariant systems using noisy input-output measurements.
Abstract: In this paper, a geometrically inspired algorithm is derived for identification of state space models for multivariable linear time-invariant systems using noisy input-output measurements. The algorithm contains two conceptual steps which allow a robust implementation using SVD techniques:
1.
Using canonical correlation, first a state vector sequence is realized.
2.
The system matrices are then obtained at once from the least squares solution of a set of linear equations.
TL;DR: In this paper, it is shown that when a rational matrix is factorized in matrix fraction description (MFD), redundancy at infinity may be introduced, which can be characterized precisely as the common divisor of the numerator and the denominator matrices at infinity.
Abstract: The development reported in Part 1 of this series is continued by presenting a unifying theory for rational matrices at infinity. The cornerstone for the development is again the homogeneous form for polynomial and rational matrices. It is shown that when a rational matrix is factorized in matrix fraction description (MFD), redundancy at infinity may be introduced. This redundancy can be characterized precisely as the common divisor of the numerator and the denominator matrices at infinity. Moreover, when the numerator and the denominator matrices are completely coprime, the determinantal degree of the denominator is the McMillan degree of the rational matrix.
TL;DR: In this paper, two algorithms for subspace inverse iteration are presented, and their convergence properties and applicability in solving TLS problems are analyzed based on the convergence rate and the operation counts in each iteration step.
07 Jun 1988
TL;DR: A realization algorithm is presented which determines a singular system realization from an arbitrary right matrix fraction description (MFD) which possesses some important properties which are similar to those of Wolovich's realization (1972).
Abstract: A realization algorithm is presented which determines a singular system realization from an arbitrary right matrix fraction description (MFD). Neither the properness nor the column-reducedness of the right MFD is required. The realization also possesses some important properties which, although in a generalized sense, are similar to those of Wolovich's realization (1972). >
TL;DR: In this paper, a survey is presented of some recently obtained results in the problem of identifying linear relations from noisy data, starting from a geometrically inspired definition of noise and linear relations, the mathematical problem is formulated.
01 Nov 1988
TL;DR: In this paper, an autoregressive moving-average spectral method is presented for the analysis of human gastric myoelectrical activity by means of cutaneous electrodes attached on the abdominal skin, called electrogastrography.
Abstract: The recording of human gastric myoelectrical activity by means of cutaneous electrodes attached on the abdominal skin, called electrogastrography (EGG), provides a noninvasive method of studying the electrogastric behavior. The normal frequency of the gastric signal is about 0.05 Hz. However, sudden changes of its frequency have been observed and are generally considered to be related to the gastric motility disorder. This suggests that spectral analysis could serve as a valuable tool for practical purposes. An autoregressive moving-average spectral method is presented. It is based on the adaptive IIR filter and therefore provides both time and frequency information of the signal. The method has been applied to the running spectral analysis of the cutaneous EGG signals. The spectral analysis of the postprandial EGGs is presented. >
TL;DR: A simple algebraic criterion is presented to check if a given 2-D MIMO filter described by a nonsingular transfer matrix is in fact a cascade of two 1-D filters with a minimal number of inputs and outputs.
Abstract: A simple algebraic criterion is presented to check if a given 2-D MIMO (multi-input/multioutput) filter described by a nonsingular transfer matrix is in fact a cascade of two 1-D filters with a minimal number of inputs and outputs. The criterion offers the advantage that whenever a filter is found to be separable, it is almost trivial to construct the two 1-D filters. >
01 Jan 1988
07 Jun 1988
TL;DR: In this paper, a zero delay subband acoustic echo canceller (AEC) is proposed, where the filter banks are realized by a simplified weighted overlap-add method and oversampling is used to avoid aliasing problems.
Abstract: The design of a zero delay subband acoustic echo canceller (AEC) is presented. The filter banks are realized by a simplified weighted overlap-add method and oversampling is used to avoid aliasing problems. Complex adaptive filters are used within the filter banks. The design considerations and hardware implementation are discussed in some detail. Parallel processing, pipeline techniques, short machine cycle time and dedicated buses are applied in the hardware implementation to realize a 2048-tap AEC in a single custom chip with existing technology. Simulation results are given and its advantages are shown to be sufficient echo return loss, echo removal in the whole frequency range, and computational saving. >
01 Nov 1988
TL;DR: It is shown that the propagation of the gastric activities in the stomach can be detected from abdominally recorded EGGs (electrogastrograms) and phase shifts have been observed from all of the volunteers and patients with the position of the stomach localized by X-ray.
Abstract: The authors show that the propagation of the gastric activities in the stomach can be detected from abdominally recorded EGGs (electrogastrograms). The conventional measurement method is not adequate for detecting the propagation. The stomach must be carefully localized before the measurement. The electrodes should be attached on the abdomen along the longitudinal axis of the stomach. Phase shifts have been observed from all of the volunteers and patients with the position of the stomach localized by X-ray. It is easier to detect phase shifts in fasting than in postprandial situations. The reason could be that there exist more contractions in the stomach in the postprandial situation. These contraction-related gastric activities do not propagate as regularly as the slow waves, which makes the cutaneous EGGs more complicated. >
01 Jan 1988
TL;DR: In this paper, l'application de la methode des moindres carres complets qui a ete imaginee pour resoudre les systemes surdetermines d'equations lineaires de la forme Ax~b dans lesquels le vecteur d'observation b aussi bien que la matrice A sont inexact.
Abstract: Discussion sur l'application de la methode des moindres carres complets qui a ete imaginee pour resoudre les systemes surdetermines d'equations lineaires de la forme Ax~b dans lesquels le vecteur d'observation b aussi bien que la matrice A sont inexacts
TL;DR: This work focuses on the implementation on a PC network of exercise sessions for students in the ESAT laboratory of the K.U. Leuven on the use of a computer for solution methods, conditions, restrictions and properties of certain techniques.