ALBERT

Notes: Abstract The paper presents a simple method for solving the optimal (LQG) control problem on an infinite time horizon for linear plants described by proper rational transfer function matrices. Since the class of proper plants discussed in the paper is more general than the one commonly used, additional properties of solutions are presented. Two numerical examples illustrate the theoretical considerations. The examples have been performed on an IBM PC using the proposed algorithm. This algorithm is a part of a public ASMCS package developed by the author for analysis, synthesis, and simulations of multi-input, multi-output (MIMO) control systems.

Notes: Abstract We study the behavior of a Hilbert network (i.e., a finite or countably infinite network whose variables are in a Hilbert space and in which the associated total energy is finite) whose elements are affected by perturbations. More specifically, we will give estimates for a change of the current distribution caused by (a) perturbations of the elements of the nominal network when the voltage sources are fixed, and (b) a change of voltage sources in a network whose elements are perturbed. The conditions given in our theorems imply insensitivity and robust stability of the nominal network. The applications of the results are illustrated by an example of an infinite network.

Notes: Abstract A new lattice filter structure to model two-dimensional (2-D) autoregressive (AR) fields is proposed. The proposed structure utilizes and extracts the information contained in the backward prediction error fields and their delayed versions. The main idea is to use two sets of reflection coefficients corresponding to two quadrant filters and to increase the number of reflection coefficients with the order of the lattice filter. Increasing the number of reflection coefficients at each stage produces a sufficient number of independent parameters to model AR fields up to order three, which is an improvement over the existing 2-D lattice filter structures. The improvement is confirmed by computer simulations. In addition, a relationship between the reflection coefficients and the AR coefficients is derived. It is also shown that the entropy contained in the backward prediction error field vector of the proposed structure is closer to the input entropy when compared to those contained in existing 2-D lattice filters.

Notes: Abstract This paper addresses the problem of global asymptotic stability in recursive first hyperquadrant causalm-D digital filters. A set of simple-to-check 1-D conditions necessary for stability of them-D system is given. Generalizations tov-dimensional (v〈m) subsystems are also provided. These derived conditions are useful in determining asymptotic convergence ofm-D digital filters implemented in fixed or floating point arithmetic. The set of 1-D conditions can be considered the analogous result to practical bounded input bounded output (BIBO) stability for linearm-D systems.

Notes: Abstract A new efficient algorithm for calculating the weighting coefficients of maximally linear, FIR digital differentiators is presented. Simple closed-form explicit and recursive formulas are derived in a very straightforward manner. Moreover, a simple recursive equation is established, relating coefficients of two digital differentiators of adjacent ranks.

Notes: Abstract This paper presents a sequence of pseudorandom arrays with triangular symmetry. The sequence of arrays is also pseudorandom in nature, so it can be called a pseudorandom sequence of arrays, or PRSA. A circuit for the PRSA generator is given and some interesting properties of this type of PRSA are discussed. Also, some of the concepts presented in this paper are clarified with an example.

Notes: Abstract This work addresses the design of LoG filters in the frequency domain within a structure formed by the cascade of quasi-Gaussian and discrete Laplacian filters. The main feature of such a structure is that it requires half the number of convolutions of the classical structure in which the LoG transfer function is expressed as the sum of two separable transfer functions of 1-D Gaussian and LoG type. Such a perspective allows one to rephrase the design of IIR and FIR filters for edge detection as a frequency domain approximation problem solvable by standard digital filter design tools. The zero-phase IIR solutions have a good performance at low orders and approximation errors practically independent of the aperture parameter. The characteristics of the nearly linear-phase IIR filters solving the problem suggest the consideration of linear-phase FIR filters with zeros constrained on the unit circle. The use of such filters leads to remarkable computational savings with respect to the filters designed by impulse response sampling. The agreement between the edge values obtained by the filters designed according to the scheme proposed in this work and those obtained by standard techniques is very good.

Notes: Abstract Magnetostatic surface wave (MSSW) devices made with pure and gallium-substituted yttrium-iron garnet (Ga:La-YIG) films are described. These devices include nondispersive and dispersive delay lines, band-pass filters, oscillators, and resonators. By controlling the magnitude of the bias magnetic field and the temperature of operation, it is possible to tune these devices over a wide frequency range extending from 0.3 to 4 GHz and from 3 to 18 GHz using Ga:La-YIG and pure YIG films, respectively. These devices could be used in pulse compression radar, microscan receivers, complicated Fourier transform processors, and fundamental oscillator circuits. In this paper, we briefly show results for pure YIG devices tunable in C and X bands and discuss, in detail, the performance of the Ga:La-YIG devices for UHF applications.

Notes: Conclusion Significant improvements are required in the performance of MSW dispersive delay lines and filter banks before they are ready for systems application. Typically delay lines with bandwidths of 1 GHz or greater, differential delays in the range 200 ns to 1μs, and minimum phase errors (〈±1 °) are required for large (∿40 dB) dynamic range compressive receivers. However, techniques are evolving (see rest of this issue) in this relatively new area of technology which will allow systems performance requirements on phase errors to be met. Possible approaches to low phase error dispersive delay lines include reflective arrays, stepped ground planes, and multiple YIG films. The stepped ground plane technique is the most advanced and uses an optimization approach to the delay-line design, which results in a minimum phase error [20]. Ultimately the minimum achievable phase error will be limited by reflections from transducers and multiple mode effects in the delay lines. The MSW compressive receiver requires parallel advances in high-speed digital processing techniques to achieve its full potential. The MSW filter bank provides a simple channelization technique applicable up to approximately 20 GHz. Narrowband channels with 10 dB insertion loss, 3 dB bandwidths of 10 to 40 MHz, and 50 dB bandwidths of 30 to 120 MHz are possible with the already demonstrated techniques. Broader bandwidth channels in the range 50 to 200 MHz with flat passband response require improved transducer design techniques. The channelized receiver does not require extremely high-speed operations but, since a large number of channels are involved, size and cost become very significant.

Notes: Abstract In recent studies, it has been verifiedheuristically andexperimentally (via simulations) that instability in power systems due to a fault occurs when one machine or a group of machines, called thecritical group, loses synchronism with the remaining machines. Using energy functions associated with a critical group (rather than system-wide functions), transient stability results which are less conservative than other existing results, have been obtained. The existence and identity of a critical group is ascertained in these studies byoff-line simulations. In this paper, we present results, for power systems with uniform damping, which establishanalytically theexistence and theidentity of the critical group of machines due to a given fault. We also present a result to determine estimates of the domain of attraction of asymptotically stable equilibrium points in power systems. The results presented herein can potentially be usedon-line to determine which machines belong to a critical group, and to use this information for corrective action (e.g.,shedding of the critical generators orfast valving for these generators). The applicability of the present results is demonstrated by means of a specific example (a 162-bus, 17-generator model of the power network of the State of Iowa).

Notes: Abstract This paper is a brief historical review of linear singular systems, followed by a survey of results on their solution and properties. The frequency domain and time domain approaches are discussed together to sketch an overall picture of the current status of the theory.

Notes: Abstract In this paper, we present a new class of two-dimensional FIR-median hybrid (FMH) filters, which we call separable FMH filters, for image noise smoothing. In a separable FMH filter, a one-dimensional FMH filter is applied to the rows and the columns of an image successively. The deterministic properties of separable and cross window FMH filters are discussed. Under certain assumptions, it is proved that a root of a separable FMH filter is a root of the corresponding cross window FMH filter. The noise attenuating properties of a separable FMH filter are studied and compared with those of the separable median filter, the two-dimensional median filter with a square window, and the cross window FMH filter.

Notes: Abstract The Chinese remainder theorem is a fundamental technique widely employed in digital signal processing for designing fast algorithms for computing convolutions. Classically, it has two versions. One is over a ring of integers and the second is over a ring of polynomials with coefficients defined over a field. In our previous papers, we developed an extension to this well-known theorem for the case of a ring of polynomials with coefficients defined over a finite ring of integers. The objective was to generalize number-theoretictransforms, which turn out to be a special case of this extension. This paper focuses on the extension of the Chinese remainder theorem for processing complex-valued integer sequences. Once again, the present work generalizes the complex-number-theoretic transforms. The impetus for this work is provided by the occurrence of complex integer sequences in digital signal processing and the desire to process them using exact arithmetic.

Notes: Abstract A method to improve the stability of the forced response of a recursive digital filter with nonlinearities due to finite wordlength is presented. An analysis of applying a stability criterion on the nonlinear filter is performed by transforming the problem into a mathematical problem of finding a zero set. As a result of the mathematical analysis, one can find a maximal sector size that bounds an equivalent nonlinearity, so that stability is ensured. This sector size is influenced by the choice of the desired nonlinearity characteristic appropriate to the problem, and the amount of the input scaling needed. Another improvement of the filter's stability is obtained by adding a feedback gain, the value of which is determined by using the zero set-finding method. Simulation results showing the improved stability properties of a filter designed by this method are presented.

Notes: Abstract In this paper a fully systolic, bit level, three-port unconstrained adaptor, which constitutes the main nontrivial building block of ladder wave digital filters, is generated. The one-dimensional binary convolution is used as the underlying algorithm for the implementation of a multiplication. The Isb-first input data organization approach is adopted and thecanonical mapping methodology is used to fully systolize the unconstrained parallel three-port adaptor at the bit level with piplining period a=1. The technique is based on a transformation of the adaptor's signal-flow graph, so that unidirectional data flow takes place. A ring-systolic scheme is proposed for implementing communications among adaptors.

Notes: Abstract We consider the problem of robust stochastic adaptive control of not necessarily minimum phase systems in the presence of unmodelled dynamics. Stochastic gradient algorithms with parameter projection and modified gain sequence are used for the estimation of the unknown controller parameters. Global stability of the adaptive system is achieved without requiring the strictly positive real condition and the persistency exciting condition to be satisfied.

Notes: Abstract A theoretical framework for the investigation of the qualitative behavior of differential-algebraic equations (DAEs) near an equilibrium point is established. The key notion of our approach is the notion of regularity. A DAE is called regular locally around an equilibrium point if there is a unique vector field such that the solutions of the DAE and the vector field are in one-to-one correspondence in a neighborhood of this equilibrium point. Sufficient conditions for the regularity of an equilibrium point are stated. This in turn allows us to translate several local results, as formulated for vector fields, to DAEs that are regular locally around a given equilibrium point (e.g. Local Stable and Unstable Manifold Theorem, Hopf theorem). It is important that these theorems are stated in terms of the given problem and not in terms of the corresponding vector field.

Notes: Abstract The classical notion of the λ-generalized nullspace, defined on a matrixA εR n×n,where λ is an eigenvalue, is extended to the case of ordered pairs of matrices(F, G), F, G ε R m×nwhere the associated pencilsF − G is right regular. It is shown that for every α εC ∪ {∞} generalized eigenvalue of (F, G), an ascending nested sequence of spaces {P α i ,i=1, 2,...} and a descending nested sequence of spaces {ie495-02 i=1, 2,...} are defined from the α-Toeplitz matrices of (F, G); the first sequence has a maximal elementM α * , the α-generalized nullspace of (F, G), which is the element of the sequence corresponding to the index τα, the α-index of annihilation of (F, G), whereas the second sequence has the first elementP α * as its maximal element, the α-prime space of (F, G). The geometric properties of the {M α i ,i=1, 2,...,τα and {P α i ,i=1, 2,...sets, as well as their interrelations are investigated and are shown to be intimately related to the existence of nested basis matrices of the nullspaces of the α-Toeplitz matrices of (F, G). These nested basis matrices characterize completely the geometry ofM α * and provide a systematic procedure for the selection of maximal length linearly independent vector chains characterizing theα-Segre characteristic of (F, G).

Notes: Abstract The robustness problem of stability for large-scale uncertain systems with a class of multiple time delays is addressed in this paper. By applying the complex Lyapunov stability theorem, the matrix measure techniques, and norm inequalities, a new approach for solving a general case of the above problem is proposed. Several robust stability conditions, delay-dependent or delay-independent, are derived to guarantee the asymptotic stability and exponential stability of the uncertain large-scale time-delay systems. Moreover, these obtained results can also be applied to the stabilization design.

Notes: Abstract Successful speech recognition is highly dependent on appropriate speech segmentation. The poor efficiency of the sequential detection of abrupt changes in the signals with relatively short stationary intervals, as is the case with speech signals, can be improved by the off-line maximum likelihood segmentation algorithm. In this paper the new segmentation algorithm is presented. For the a priori known number of segments, the algorithm determines such signal partitions for which the sum of segment distortion is minimal. The generalized maximum likelihood distortion measure has been introduced, and has proven to be particularly efficient on short signal segments. In the case of an unknown number of segments, its estimate is obtained comparing the reduction of the distortion. The asymptotic properties of the distortion sequence have been analyzed, which led to the definition of the presented segmentation algorithm. The introduced measure can be applied both to the AR and ARMA models. The segmentation algorithm is verified on test signals as well as on the natural speech signal, for which the pitch synchronous framing scheme is applied. The experimental results also include a comparison of the AR and ARMA model-based segmentations. The first results show that ARMA model-based segmentation gives somewhat better results than the AR model algorithm.

Notes: Abstract A new bit-serial architecture for implementation of high order FIR filters is introduced, as well as example FPGA and CMOS realizations. This structure exploits the simplicity of coefficients that consist of two power-of-two terms to yield efficient implementations. Quantization effects are discussed and a simple block scaling method for reducing rounding and truncation noise in high order filters is also presented.

Notes: Abstract This paper describes an implementation of 2-D FIR and IIR linear digital filters via VLSI array processors. The underlying realization structures are based on the matrix decomposition approach. The 2-D concurrent processing is used in order to implement the row and column delays within the cycle time. A high degree of concurrency is achieved by exploiting the pipelining of the array processors with the inherent parallelism of the matrix decomposition structure. The resulting structures are modular, and regular, use only local communication and internal local feedback loops, and achieve high throughput and sampling rates.

Notes: Abstract This article studies dynamical systems under perturbations. We prove an Equilibrium Equivalence Theorem that guarantees that the dynamics of the system remains unchanged under perturbations with certain fairly general assumptions. We also prove that a power system is robust with respect to parameter changes in generic situations. Concepts of equilibrium equivalence and equilibrium equivalence structural stability are developed and are applied to studies of bifurcations of vector fields on noncompact manifolds. A constructive approach to equilibrium equivalence structural stability verification is emphasized. General results on structural stability of vector fields on differential manifolds are established and important applications of this theory to stability analysis are considered.

Notes: Abstract The problem of designing a stabilizing compensator for a control system to achieve prescribed initial value constraintsυ (i)(0+)=yi is considered. Indeed, modulo certain technical conditions, such a compensator exists if and only if yi=0;i= 0,1,...,r〈p〉 +r〈t〉 −2; wherer〈p〉 is the relative degree of the plant andr〈t〉 is the relative degree of the system input. This theorem is derived and a complete parameterization of the set of compensators that achieve the prescribed design constraints is formulated.

Notes: Abstract Wideband source location in array signal processing has received much attention in the literature lately. Methods such as the Coherent Signal Subspace (CSS) method proposed by Wang and Kaveh [12], and the signal subspace method used by Cadzow [3], are typical of the approaches used to tackle the multiple wideband source location problem. Most of these methods are variations of the narrowband high-resolution methods. Grenier [5], on the other hand, has applied the idea of time-dependent Auto-Regressive (AR) modeling [7] for a nonstationary process to the frequency domain AR modeling of the sensor outputs in a linear array and has been able to produce good results for a wideband signal. The AR coefficients in the model are expanded in a set of frequency-dependent basis functions. The choice of the basis functions was deemed immaterial and the method works even when only one snapshot of the array output is available. In this paper, we re-examine this method and present an extension of the frequency-dependent AR modeling approach to a planar array. It is shown that the use of a set of sinc functions for representing the frequency-dependent AR coefficients accurately tracks their evolution in the frequency domain, and gives superior performance compared to that when power or Legendre functions are used. We also propose two methods for smoothing the spatial spectra, from which the source locations are determined. Comparison with the CSS method are also presented.

Notes: Abstract A novel and complete method is presented for constructing the generalized Bezout identity in polynomial form of the transfer matrix from its state-space representation of a reachable and observable system. As a result, one can easily apply well-developed synthesis and analysis theories to systems described in the frequency domain.

Notes: Abstract The aim of this paper is to obtain the uncertain value set in the complex plane for systems with real and complex parameters that are known to lie inside a ball in a weightedl p-norm. It generalizes previously available results and may be used to test the robust stability of polynomials whose coefficients lie in a weighted lp-ball.

Notes: Abstract This paper describes a new approach—a Fixed-Level Least Squares (FLLS) method for linear-phase FIR filter design. It is intended for rejection of the Gibbs phenomenon through the introduction of a set of equally spaced fixed levels in the transition band and subsequent redefinition of the approximated and weighted functions. Detailed mathematical solutions of the problem as well as many examples are given. The results in graphical form are shown as an output of the FLLS software model.

Notes: Abstract In this paper we examine order reduction of parabolic systems using modal truncation. The parabolic distributed system is first approximated using the Galerkin method. The system matrices have a special structure that allows us to find the approximate spectrum of the parabolic system. To do this we compute approximate inverses of tridiagonal, diagonally dominant symmetric matrices. The approximation leads to algorithms of orderO(n), as opposed to traditional algorithms of orderO(n), wheren is the order of the system. Finally, an example is presented to illustrate the proposed algorithm.

Notes: Abstract We give results concerning the problem of approximating the input-output maps of nonlinear discrete-time approximately finite-memory systems. Here the focus is on the linear dynamical parts of the approximating structures, and we give examples showing that these linear parts can be derived from asingle prespecified function that meets certain conditions. This is done in the context of an approximation theorem in which attention is focused on what we call “basic sets.” We also consider the related but very different problem of approximating nonlinear functionals using lattice operations or the usual linear ring operations. For this problem we give criteria, not just sufficient conditions, for approximation on compact subsets of reflexive Banach spaces (any Hilbert space is a reflexive Banach space).

Notes: Abstract New calibration methods are presented for bearing estimation with sensor location uncertainties, which are based on eigenvalue decomposition of the sample covariance matrix and three or more nondisjoint sources in known bearings. The methods can be applied to arbitrary arrays, including linear arrays, require fewer computations, and are suitable for low signal-to-noise ratio (SNR) cases. Computer simulations are given to illustrate the performance of the proposed methods.

Notes: Abstract The Wigner bispectrum of multicomponent signals is studied, and its modified and reduced forms are introduced. A generalization of the presented forms to the Wigner higher-order spectra (WHOS), in the case of multicomponent signals, is provided. From our previous work it is known that cross terms removal (reduction) is possible for odd-order spectra with equal numbers of conjugated and nonconjugated terms. Here, we extend the analysis to even-order spectra. The theory is illustrated by examples.

Notes: Abstract We describe methods to establish identifiability and information-regularity of parameters in normal distributions. Parameters are considered identifiable when they are determined uniquely by the probability distribution and they are information-regular when their Fisher information matrix is full rank. In normal distributions, information-regularity implies local identifiability, but the converse is not always true. Using the theory of holomorphic mappings, we show when the converse is true, allowing information-regularity to be established without having to explicitly compute the information matrix. Some examples are given.

Notes: Abstract The extendibility of estimated correlation bisequences from an available sampled data array is described in terms of the generating functions of associated block Toeplitz with Toeplitz block (BTTB) matrices. The periodogram-based correlation bisequences are shown to be extendible. It is shown that the method of resultants and subresultants is convenient for generating the nonlinear constraints in the optimization problem which is solved iteratively for power spectrum estimation. A nontrivial example illustrates the concepts developed.

Notes: Abstract A multiresolutional approach is presented for effectively recognizing three-dimensional (3D) objects. The approach is both pose and scale invariant. A multiresolutional model base is constructed, and multiscale edges of the object are detected using the wavelet transform. The minimum alignment between model base and object is realized by the linear mapping scheme.

Notes: Abstract A doubly recursive algorithm for time domain convolution with a piecewise linear weighting function is presented that combines the speed of a recursive (IIR) digital filter with the flexibility and ease of design of a nonrecursive (FIR) digital filter. The approach approximates the desired FIR weighting function by a sum-of-triangles weighting function. ForL triangles (or triangle pairs for a linear phase filter) the algorithm is of orderLN. The approximation improves with the number of triangles. A significant advantage of the algorithm compared to FFT filtering or direct convolution is that there is no necessity of a tradeoff between frequency response accuracy and computation time per output point as the data spacing decreases in the filtered signal. The computational complexity is dependent on the number of triangles chosen, not the width of the weighting function, so the algorithm is especially effective for filters with an inherently wide FIR weighting function.

Notes: Abstract A solution of the problem of correlation analysis of stochastic amplitude-angle modulated signals is presented. The solution is obtained on the basis of the theory of linear stochastic processes for the general case of nonstationary correlated Gaussian noises in both modulation channels. These noises are regarded as the results of a linear filtration of Gaussian white noise processes. As an example of the analysis a case of band-pass modulation noises is considered.

Notes: Abstract Two-dimensional (2-D) periodically shift-variant (PSV) digital filters are considered. These filters have applications in processing video signals with cyclostationary noise, scrambling digital images, and 2-D multirate signal processing. The filters are formulated in the form of a Givone-Roesser (GR) state-space model with periodic coefficients. This PSV model is then presented in block form as a shift-invariant system that also has the same GR state-space form. This block form has reduced computations and ease of analysis. An algorithm is developed that transforms any given 2-D PSV GR system to its equivalent shift-invariant model. Invertibility of this model is an important consideration, especially in image scrambling applications. A condition is established for the invertibility of the shift-invariant model of the 2-D PSV system. Also, the inverse system can be easily computed from the original. The established results are illustrated with an example.

Notes: Abstract In a recent paper a method is described for constructing certain approximations to a general element in the closure of the convex hull of a subset of an inner product space. This is of interest in connection with neural networks. Here we give an algorithm that generates simpler approximants with somewhat less computational cost.

Notes: Abstract In this paper, we gave an exponential estimate for the solutions of singular systems of differential equations with a delay and establish a comparison inequality for the systems. Based on this we estabish the criteria for asymptotic stability for the composite singular systems.

Notes: Abstract This paper presents the VLSI architectures for three-level correlator design based on 1-μm CMOS technology. The architecture performs very high speed, real-time, three-level cross-correlation of signals. Two architectures, one for serial incoming samples of signals (serial data) and the other for stored signal samples (parallel data), are described in the paper.

Notes: Abstract This paper gives the existence of and bounds for the branch voltages and currents in a given infinite network, which on each branch are related by a maximal monotone resistance function. We give a simpler proof than that of Minty, with smaller bounds, in the finite network case.

Notes: Abstract An efficient procedure is presented for computing the frequency-response templates of uncertain rational transfer functions depending on two independent interval polynomials. It consists of a modified Cohen-Sutherland algorithm for the intersection test of two rectangles and a pivoting algorithm. Because the proposed procedure traces out the boundary of a frequency-response template directly, it is, in general, more efficient than the previously proposed algorithms.

Notes: Abstract The potential for data compression in using fractal interpolation functions (FIFs) is realized by the construction of a set of multirate filters. The filter tap weights are determined by optimizing the energy contents of a preselected set of frequency bands. This filter bank implementation of the FIF is successfully used to compress data simulated in a tracking environment.

Notes: Abstract A first-order autoregressive filter is altered by changing the constant gain to two or more gains that cyclically alternate in time. The advantages of this system are shown, and the relation to linear autoregressive moving average difference equations of higher order is derived.

Notes: Abstract This paper refers to the fast implementation of the positional forward acceleration of the end effector of revolute robotic arms with spherical wrists, using the distributed arithmetic technique. The acceleration of the end effector is calculated by a cascade configuration of two pipelined arrays that calculate the Jacobian matrix and its time derivative, as well as the centrifugal-Coriolis and linear accelerations. These partial accelerations are then added in the adder tree. The building blocks of the arrays are the distributed arithmetic-based circuits that implement the matrix-vector multiplications involved in the calculations. The digit-serial configuration of the proposed implementation of the positional forward acceleration of the end effector is described. The serial and the parallel configurations may result as special cases of the digit-serial configuration. The proposed distributed arithmetic (DA) implementation of the positional forward acceleration may be applied, after appropriate modifications, to the general case of robots having either revolute or prismatic joints, with any type of wrist.

Notes: Abstract This paper addresses the calculation of the extrema of the sin x/x function. First the Newton-Raphson method is used, which allows us to obtain the extrema locations very fast through the use of a recursion formula. Then a second approach is proposed, which gives the extrema locations and the extrema amplitudes in the form of series expansions. Simple, accurate algebraic expressions are derived.

Notes: Abstract We give some existence results for a resistive network in which the components are neither voltage nor current controlled; that is, they are merely unicursal. In fact we allow coupling. Degree arguments give existence and bounds. We study several ways of avoiding the requirement of eventual passivity. No-gain and passive multiterminal elements are included. The results are extended to infinite networks.

Notes: Abstract Random sampling is one of the methods that can overcome the Nyquist limit when evaluating a frequency spectrum of a signal. However, the computational complexity becomesN 2 as the FFT cannot be used. A new approach, called hybrid additive random sampling, is proposed. This new scheme is devised by concatenating random sampling sequences in such a way that symmetry is created in the transform kernel for reducing the computational effort while the anti-alias property is maintained. A savings of the least 75% in computation is achieved. The sampling scheme is also found to be suitable for parallel implementation. In this paper, the algorithms for generating the sampling sequence and evaluating the spectrum are described in detail. The performances of the scheme in terms of noise, accuracy, etc., are compared with genuine random sampling and another approach proposed previously. The advantages and limitations are included.

Notes: Abstract Centralized methods for source location using sensor arrays have computational and communication burdens that increase significantly with the number of sensors in the array. Therefore, these methods may not be usable in the applications involving very large arrays. In such applications, the data processing may need to be decentralized. This paper introduces two methods for decentralized array processing, based on the recently proposed MODE algorithm. For prescribed nonoverlapping subarrays, both methods are shown to be statistically optimal in the sense that asymptotically they provide the most accurate decentralized estimates of source location parameters. The problem of subarray selection to further optimize the estimation accuracy is only briefly addressed. The two methods are intended for different types of applications: the first should be preferred when there exist significant possibilities for local processing or for parallel computation in the central processor; otherwise the second method should be preferred. The accuracy of the two decentralized methods is compared to the centralized Cramér-Rao bound, both analytically and numerically, in order to provide indications about the loss of accuracy associated with decentralized processing.

Notes: Abstract We present in this paper a recursive-in-order least-squares (LS) algorithm to compute efficiently the parameters of a 2-D Gaussian Markov random field (GMRF) model. The algorithm is based on the fact that the least-squares estimation of the parameters of a 2-D noncausal GMRF model is consistent and the coefficient matrix in the normal equation has near-to-block-Toeplitz structure. Hence, it has a Levinson-like form for the updating of model parameters by introducing auxiliary variables. Moreover, this paper proposes the concept ofrecursive path for 2-D recursive-in-order algorithms, and points out that there exists a tradeoff between fast computation of the parameters and accurate choice of model support; a compromise recursive path is then suggested where the orders change alternately in two directions. The computational complexity of the developed algorithm is analyzed, and the results show that the algorithm is more efficient when either the image size or the model support is larger. It is found that the total number of multiplications (mps) involved in the new algorithm is only about 14% of that in the conventional LS method when the image size is 512 × 512 and the neighbor set of the model is a 17 × 17 window. Computer simulation results using the recursive-in-order algorithm developed in this paper and the conventional LS method are given to verify the correctness of the new algorithm.

Notes: Abstract In this paper, we present a hybrid space-time-filtered Viterbi receiver using multiple antennas for co-channel interference (CCI) reduction and intersymbol interference (ISI) equalization in a slow Rayleigh fading channel. In this approach, a space-time filter is first applied at the antenna outputs to maximize the signal-to-interference-plus-noise ratio (SINR), and the scalar output is then sent to a Viterbi equalizer. We propose a closed-form solution to jointly determine the weight vector for the space-time filter and the channel vector for the Viterbi equalizer. We also examine the need for a whitening filter prior to the Viterbi equalizer and show that it only marginally improves the performance. Simulation results are provided to validate our approach and to compare the performance of our receiver with that of different existing receivers.

Notes: Abstract Prosthetic heart valves have been responsible for extending the life spans and improving the quality of life of many people with serious heart conditions. Even though the heart valves are extremely reliable, eventually they are susceptible to the long-term fatigue and structural failure effects expected for mechanical devices operating over long periods of time. In [2] a classification procedure was developed using spectral features obtained from acoustic signals to determine the condition of the prosthetic heart valve. Although this classification procedure has produced very encouraging results, this method still lacks a fundamental physical description of the sounds produced by the valve during normal operation. In order to obtain a better understanding of the valve acoustic response, we have performed a set of anechoic tests. In this paper, we describe the anechoic experiment and also present limited transient response results. This transient information will eventually be used to identify and improve the features used to classify the valve condition.

Notes: Abstract In this two-part study we present a new design methodology for neural classifiers. The design procedure utilizes a multiclass vector quantization (MVQ) algorithm for information extraction from the training set. The extracted information suffices to specify the hidden layer in a canonical neural network architecture. The extracted information also leads to the specification of neuron inhibition rules and subsequently the design of the hidden layer-to-output map. In Part I of the study we focus attention on the MVQ algorithm and how it is used to extract information from a training set. The extracted information is referred to as thecodebook. The codebook is used to directly specify the hidden layer. This specification can take the form of a perceptron layer, a radial basis layer, or a heterogeneous layer involving a mixture of neuron types. These and otherh-layer specifications are determined directly from the same extracted information. The MVQ codebook also suffices to scale the activation function of each neuron. In Part II we consider the nonsimplistic hidden layer-to-output map design. We note that the MVQ algorithm, as it extracts information, decomposes the design set into disjoint neighborhoods. For each neighborhood we identify subsets of the hidden layer neurons, which are significant sensors for the neighborhood. For each such subset we construct an output map. Inhibition rules are established to ensure that the proper output map is activated. In benchmark simulations the overall design exhibits excellent performance, to the extent that we are hard pressed to identify bounds on performance, if any.

Notes: Abstract A transform domain image tagging orwatermarking method that survives image cropping (and, hence, is “holographic”) was proposed at Bell Labs in September 1994. This report analyzed in detail the various properties of this method and introduces an optimal procedure for watermark recovery.

Notes: Abstract Cross terms are an inherent consequence of the second order nature of Cohen's class TFDs (Time-Frequency Distributions) [5], [6]. They are manifest in a TFD of multicomponent signals as spurious artifacts arising from interactions between the various signal components, and they can often appear at times and/or frequencies inconsistent with the underlying physical nature of the signal, causing misinterpretation [2], [3], [4]. There are many time frequency distributions that avoid the cross term effect; the best are the Choi-Williams ED (Exponential Distribution) [1] and Levin's IPS (Instantaneous Power Spectrum) [9]. In this paper we combine the cross term reducing philosophy of the ED and IPS to obtain a new TFD that most effectively reduces the cross term effect. Surprisingly, the new TFD also satisfies most desired TFD properties.

Notes: Abstract An algorithm for the exact computation of the frequency responses of linear interval systems is obtained. For the computation a sectoring stage is added to an elimination algorithm to eliminate interior curves. It is shown that the intersection of a ray and the frequency response set is an empty set or a line segment whose endpoints are extrema of the functions defined. Required mathematical analysis tools for the development of the sectoring algorithm and an illustrative example are provided.

Notes: Abstract The geometric theory of the domain of an ordered pair (F, G) of matrices or the geometry associated with matrix pencils provides a unifying framework for the study of algebraic, dynamic and feedback properties of linear singular systems. The key concepts and tools of the geometry are the notions of the (F, G)-, (G, F)-invariance and a set of subspace sequences. In this paper, an alternative characterization of these sequences is given based on the properties of the partitioned null spaces of appropriate sequences of Toeplitz matrices defined by the (F, G) pair. The results provide a simple procedure for the computation of the limit spaces of these sequences and clearly cover corresponding problems of the singular, implicit systems theory.

Notes: Abstract This paper presents a new concept of narrowband and broadband active noise compression (ANCom) systems which improves the conventional active noise equalization (ANE) technique to meet some practical requirements. By using a variable equalization coefficient, the narrowband ANCom system can automatically switch among the four different operating modes of narrowband ANE, according to the primary noise power. As a result, it can “compress” the narrowband noise into a desired power range when the primary noise power varies drastically. Compared to a conventional broadband ANE system, the novel broadband ANCom system not only has the ability to shape the residual noise power spectrum, but can also compress the residual noise power into a predetermined dynamic range as the narrowband AN-Com does. Theoretical analyses are conducted for both the narrowband ANCom and broadband ANCom systems. Simulation results show that the proposed algorithms work well in both static and dynamic situations.

Notes: Abstract The cornerstone of the theory of discrete-space single-input single-output linear systems is the idea that every such system has an input-output mapH that can be represented by a convolution or the familiar generalization of a convolution. This thinking involves an oversight, which, for the case of bounded inputs mapped continuously into bounded outputs, was recently corrected by adding an additional term to the representation. Here we give a more general result that addresses an important larger family of inputs.

Notes: Abstract In this paper the problem of robust stabilizing a linear, time-invariant singular system is studied. The characterization is given in terms ofH ∞-bounded perturbations to the numerator and denominator factors of its normalized left coprime factorization. An optimal stability margin is provided in terms of the definition of the Hankel norm of a singular system. The Hankel norm is computed using two generalized Lyapunov equations.

Notes: Abstract Of concern is the propagation of distortionless surface waves in a medium that may be nonuniform relative to depth. Distortionless wave propagation in inhomogeneous media was discussed by V. Burke, R. J. Duffin and D. Hazony, inQuart. Appl. Math., 183–194 (1976). Accordingly, the media could be modeled by a distributed electrical ladder network, nonuniform along the axis. We give a two-dimensional development based on Hooke's law and Newton's law which leads to the well-known case of Rayleigh waves in homogeneous media. It will be seen that the available pool of propagation modes greatly increases when high-pass propagation is included. The emphasis is on media where the elastic coefficients track one another as a function of depth. Special cases are studied in detail showing that as a disturbance travels along the surface, it may assume a broadband phase change, which translates into a shape distortion in the time domain, which is periodic with distance. Applications may be found in acousto-optics, in situ monitoring of elongated bodies, high-frequency SAW filters, microstrips, and any situations where surface waves are used in an environment of high precision or relatively large distances.

Notes: Abstract This study presents a linear output-based controller for stabilizing a rigid-link flexible-joint electrically driven (RLFJED) robot manipulator. The proposed controller ensures local exponential stability under some uncertainty conditions. It is assumed that the velocity signals from the link side are not measurable. The controller is analyzed by using tools for pole placement by an output-feedback in the framework of the linear system theory. Some useful structural properties of the systems under consideration have been studied. Applications of the results to the set-point regulation control problem are considered.

Notes: Abstract In this paper we consider an adaptive controller with vanishing gain and excitation of the reference signal. We use the burst recovery concept to show that all signals in the adaptive loop remain uniformly bounded. We also show that the mean-square performance converges so that the adaptive system is optimal in the sense that the parameter estimation error and the one-step ahead prediction error are uncorrelated in the mean despite the presence of the unmodeled dynamics.

Notes: Abstract Finite homogeneous Markov chains ξ, which admit invariant probability distributions, can be defined by the cycloids { $$\bar C_k $$ } (closed polygonal lines whose consecutive edges have various orientations that do not necessarily determine a common direction for $$\bar C_k $$ ) occurring in their graphs. These Markov chains are called cycloid chains, and the corresponding finite-dimensional distributions are linear expressions on the cycloids { $$\bar C_k $$ } with the real coefficients αk. Then the collection {{ $$\bar C_k $$ }, {αk}}, called the cycloid decomposition of ξ, gives a minimal description of the finite-dimensional distributions that, except for a choice of the maximal tree, uniquely determines the chain ξ. Furthermore, the cycloid decompositions have an interpretation in terms of the transition probability functions expressing the same essence as the known Chapman-Kolmogorov equations.

Notes: Abstract We study solutions of the “linear system in a saturated mode” $$\begin{array}{*{20}c} {(M)} & {x' \in Tx + c - \partial I_{D^n } x.} \\ \end{array} $$ We show that a trajectory is in a constant face of the cubeD n on some interval (0,d]. We answer a question about comparing the two systems: (M) and $$\begin{array}{*{20}c} {(H)} & {\begin{array}{*{20}c} {Cu' = T\upsilon + c - R^{ - 1} u,} & {\upsilon = G(\lambda } \\ \end{array} u)} \\ \end{array} $$ . As λ→∞, limits ofv corresponding to asymptotically stable equilibrium points of (H) are asymptotically stable equilibrium points of (M), and the converse is also true. We study the assumptions to see which are required and which may be weakened.

Notes: Abstract In this paper we introduce a new computational method for solving the diffusion equation. In particular, we construct a “generalized” state-space system and compute the impulse response of an equivalent truncated state-space system. In this effort, we use a 3D finite element method (FEM) to obtain the state-space system. We then use the Arnoldi iteration to approximate the state impulse response by projecting on the dominant controllable subspace. The idea exploited here is the approximation of the impulse response of the linear system. We study the homogeneous and heterogeneous cases and discuss the approximation error. Finally, we compare our computational results to our experimental setup.

Notes: Abstract A simple state-space approach for the four-block singular nonlinearH ∞ control problem is proposed in this paper. This approach combines a (J, J′)-lossless and a class of conjugate (J, J′)-expansive systems to yield a family of nonlinearH ∞ output feedback controllers. The singular nonlinearH ∞ control problem is thus transformed into a simple lossless network problem that is easy to deal with in a network-theory context.

Notes: Abstract The stability of time-varying autoregressive (TVAR) models is an important issue in many applications such as time-varying spectral estimation, EEG simulation and analysis, and time-varying linear prediction coding (TVLPC). For stationary AR models there are methods that guarantee stability, but the for nonadaptive time-varying approaches there are no such methods. On the other hand, in some situations, such as in EEG analysis, the models that temporarily exhibit roots with almost unit moduli are difficult to use. Thus we may need a tighter stability condition such as stability with margin 1−ϱ. In this paper we propose a method for the estimation of TVAR models that guarantees stability with margin 1−ϱ, that is, the moduli of the roots of the time-varying characteristic polynomial are less than or equal to some arbitrary positive number ϱ for every time instant. The model class is the Subba Rao-Liporace class, in which the time-varying coefficients are constrained to a subspace of the coefficient time evolutions. The method is based on sequential linearization of the associated nonlinear constraints and the subsequent use of a Gauss-Newton-type algorithm. The method is also applied to a simulated autoregressive process.

Notes: Abstract In this paper, a novel VLSI algorithm for the computation of a two-dimensional discrete cosine transform is proposed. The 2D-DCT equation can be expressed by the sum of high order cosine functions, and the algorithm can be realized by combining a highly efficient first order recursive structure with some simplified matrix multiplications, which results in highly regular hardware architecture and simple routing. The algorithm has temporal and spatial locality of connection and can be segmentized for pipeline operations, so the computation time is greatly reduced. Owing to the simplicity in hardware structure, it is especially good for VLSI implementation.

Notes: Abstract We examine a generalized matched-filter problem in which the interference is a nonstationary process generated by passing white noise through a general linear time-varying filter. First a matched filter is constructed by transforming the problem into an equivalent formulation involving stationary interference and a time-varying propagation channel. Whereas the response of a time-invariant matched filter is sampled at its peak, the response of this time-varying matched filter is normalized before sampling to account for variations in the signal power. Next a matched filter is constructed using a spectral characterization of the nonstationary interference. This construction is then used to formulate a simplified solution for the case where the rate of variation in the nonstationary interference is sufficiently small. The different solutions are illustrated by a numerical example.

Notes: Abstract The purpose of this paper is to investigate the approximation capability of elliptic basis function (EBF) neural networks. The main results are: (1) A necessary and sufficient condition for a functionS′ (R 1) to be qualified as an activation function in the hidden layer of an EBF neural network is given. (2) Every nonzero functionG ε L 2(R n ) is qualified to be an activation function in the elliptic neural network to approximate any function in L2(Rn). (3) As applications, we give new proofs of the theorems concerning the approximation capability of affine basis function (ABF) neural networks and generalized radial basis function (GRBF) neural networks (including radial basis function neural networks) with arbitrary activation functions. In particular, we solve the problem of the approximation capability of sigma-pi neural networks.

Notes: Abstract The frequency response of linear interval systems under PID controllers is analyzed. A two-stage elimination algorithm is presented to describe the boundary of the response in the complex plane. Illustrative examples are included.

Notes: Abstract This article deals with a new intelligent analog circuit design system. Unlike previous design methods, this approach introduces a formal circuit representation for both the numerical and heuristic knowledge of the design system. The predicate logic circuit representation is proposed as a new formal analog circuit description language. The language's syntax and semantics provide a precise symbolic description of analog circuit functionality at higher levels of hierarchy and network component connectivities, together with CMOS transistor sizes as the lowest level of hierarchy. It is shown how sentence conversion rules of language grammer can be used to derive transistor level circuits from input performance specifications. Language sentences have the form of clauses. In addition to clause representation, a frame representation to reflect VLSI design hierarchy is also introduced. An original bidirectional inference mechanism with elements of hypothesis has been introduced to infer designs from the knowledge in clause and frame representations. The unique feature of this circuit knowledge representation is its ability to automate the analog CMOS circuit design process. The design methodology is described in detail. The proposed iterative closed loop design system adopts an expert system approach to provide topological synthesis, uses a SPICE circuit simulator to evaluate the circuit performance, and uses a new diagnostic expert system to provide advice on how to improve the design. The implementation of the methodology and associated experimental results for simple CMOS operational amplifier designs are also presented.

Notes: Abstract This paper presents a comparison between the traditional image processing method and the area vector concept as well as the new technique of artificial neural networks. Freeman chain coding is considered in the study, and the principle of segmentation may be based and implemented for further investigations resulting from the proposed work. The pattern recognition concept is analyzed and defined through the sigmoid function and the determination of the threshold of a gray image for an object. The block schemes for the given protocols are summarized in a single scheme for illustration and comparison purposes. The synthetic pictures are generated and investigated regarding the dependence of computer vision on the contents of the artificial neural network. The normalization technique is included to eliminate noise and zooming problems. The minimum computational time for image processing with the generated pictures is also determined. The rate of deflection in the computational time is recommended for sensing the minimum computational time according to the variation of the number of hidden units in the hidden layer. A three-layer neural network has been used. The study of gray binary imaging for color pictures is illustrated to save computational time and effort.

Notes: Abstract Of concern is an environment made up of signals and systems tightly confined both in time and frequency. Such an environment is often encountered in transmission line circuits, radar, sonar, and optical circuits, and when the principal signals are well-defined sharp pulses. It will be seen that once this environment is achieved, the signals and systems possess some attractive properties. A conventional system may preserve the symmetry of a propagating signal or change its symmetry from even to odd or from odd to even. Another system may be used to predict the arrival of an incoming pulse with a high degree of accuracy. Electrical networks may also be associated with these properties. Approximation problems, existence theorems, and realization schemes will be addressed and developed.

Notes: Abstract In a previous paper [1], a new algorithm for ARMA spectral estimation of stationary time series has been presented. The algorithm is based on nonlinear least squares fit of the sample partial autocorrelations to the partial autocorrelations generated by the assumed ARMA model. This paper explores the statistical properties of the above algorithm, including some numerical examples of the asymptotic variance of the estimated parameters, as compared to the Cramer-Rao bound. The results confirm the good performance of the algorithm and suggest an improvement in its implementation.

Notes: Abstract This paper develops a new digitally redesigned pulse-amplitude modulated (PAM) controller for a continuous-time input time-delay/nondelay system with nonsynchronous sampling. The concept of the law of the mean from the input integral calculus is utilized for the development of the equivalent digital predictor controllers from the available analog predictor controllers. As a result, the digitally controlled states closely match the original continuous-time states. To implement the developed discrete-time state-feedback PAM controller, this paper also develops an ideal discrete-time state using nonsynchronously sampled input-output data of the continuous-time input delay/nondelay system without establishing a dynamic observer.

Notes: Abstract The model-matching problem for systems described by external models is considered in frameworks of both external and input-output equivalence. Necessary conditions for the solvability of the problem are produced, and it is shown that in certain cases these conditions are also sufficient. In the case where necessary and sufficient conditions exist, the solutions of the problem are obtained in a constructive way and a parametrization of solutions is given.

Notes: Abstract This paper addresses the invariance property of Gaussian signals, originally derived by Bussgang, which characterizes the input/output moment relation of a hybrid nonlinear moment (HNM) estimator based on a zero-memory nonlinearity (ZMN) g(y). Some re-derivations of this property are reviewed, and an original, direct, and simple proof is presented (Appendix 1). The paper then derives a new interpretation of this property (Theorem 1) that shows a moment-sense equivalence between g(y) and a linear mappingh 1(y) whose coefficients a0 and a1 are completely characterized in terms ofg(y) and are shown to be optimal in a mean square error (MSE) sense. A direct and very interesting byproduct of this interpretation is a simple linear relationship between the input and output of the HNM estimator involved. The property is then generalized (Theorem 2) to signals other than Gaussian, resulting in an infinite cumulant series expansion of the HNM estimator output, whose coefficients are all characterized in terms ofg(y). Applications of Theorem 1 to some ZMNs commonly used in signal processing and control theory are presented that clearly illustrate the power and elegance of the invariance property. Finally, some conclusions are given.

Notes: Abstract In this paper the connection between ‘mass’M, ‘resistance’ϱ and ‘commute time’θ for random walks on graphs is further explored, and the relationθ=2M ·ϱ is proved. An extension of the result is made to multigraphs, which are an extension of the graph concept where a black box is treated like an edge.

Notes: Abstract Let {S(A):A ∈A}, whereA is a subset of an infinite-dimensional normed linear spaceL, be a class of general nonlinear input-output systems that are governed by operator equations relating the input, state, and output, all of which are in extended spaces. IfQ is a given operator from a specified set ¯D i, of inputs into the space of outputs ¯H 0, the problem we consider is to find, for a given ɛ〉0, a “parameter”A ε∈A such that the transmission operatorR(A ε) ofS(A ε) furnishes a nearly best (or ɛ-best) approximation toQ from allR(A),A ∈A. Here the “distance” betweenQ andR(A) is defined as the supremum of distances betweenQz andR(A)z taken over allz ∈ ¯D i. In Theorems 2 through 5 we show that ifS(A) is “normal” (Definition 2),A satisfies some mild requirement andL contains a fundamental sequence, then establishingA ε∈A reduces to minimizing a certain continuous functional on a compact subset ofR n, and thus can be carried out by conventional methods. The applications of results are illustrated by the example of a model-matching problem for a nonlinear system, and of optimal tracking.

Notes: Abstract It is shown that the elementsG of a large class of input-output maps can be uniformly approximated arbitrarily well using a certain structure if and only ifG is continuous. For the case considered the system inputs and outputs are defined on a discrete set {0, 1,...,a 1}×...{0, 1,...,a m }, in which a1,...,a m are positive integers. Our approximating structure involves certain functions that can be chosen in different ways. For the special case in which these functions are taken to be certain polynomial functions, the input-output map of our structure is a generalized discrete Volterra series. Our results provide an analytical basis for the use of such series.

Notes: Abstract In this paper, the stability robustness of deterministic state feedback discretetime linear quadratic (LQ) optimal regulators for the performance index with cross-product terms is analyzed. Guaranteed stability margins for such a type of LQ optimal regulator are suggested for the first time. These stability margins are obtained on the basis of a modified return difference equality and are expressed directly in terms of the elementary cost and system matrices. Sufficient conditions to guarantee the required stability margins are presented. Finally, the connection between the suggested stability margins and the selection of weighting state, input, and cross-product matrices is investigated, and useful guidelines for choosing proper weighting matrices are presented.

Notes: Abstract A novel computationally efficient realization of sharp linear-phase finite impulse response (FIR) bandstop filters is proposed. The synthesis scheme for the bandstop filters is derived from variations of the frequency-response-masking technique. Five realization structures are presented in this paper for the synthesis of five different classes of bandstop filters. Approximate expressions for the optimal value of the impulse response up-sampling ratio (M) and the corresponding number of multipliers are derived.

Notes: Abstract Estimating the covariance sequence of a wide-sense stationary process is of fundamental importance in digital signal processing (DSP). A new method, which makes use of Fourier inversion of the Capon spectral estimates and is referred to as theCapon method, is presented in this paper. It is shown that the Capon power spectral density (PSD) estimator yields an equivalent autoregressive (AR) or autoregressive moving-average (ARMA) process; hence, theexact covariance sequence corresponsing to the Capon spectrum can be computed in a rather convenient way. Also, without much accuracy loss, the computation can be significantly reduced via an approximate Capon method that utilizes the fast Fourier transform (FFT). Using a variety of ARMA signals, we show that Capon covariance estimates are generally better than standard sample covariance estimates and can be used to improve performances in DSP applications that are critically dependent on the accuracy of the covariance sequence estimates.

Notes: Abstract Results in the study of signal processing based on the use of parameter structural modeling (PSM) are presented. First, we introduce a special form of time-series modeling based on signal-dependent building blocks. Such modeling is used in the design of a nestedform transversal structure, known as a composite filter, based on a shift-invariant finite impulse resonse (FIR) as well as infinite impulse response (IIR) building blocks. The newly proposed composite PSM model (CPSM) possesses a unique feature, namely, its ability to suppress one signal of a given structure, while at the same time being ideally transparent to another one. The intrinsic property of this proposed CPSM is its enhanced insensitivity with respect to noise as well as its ability to fast track, in contrast to the commonly used linear line-enhancer based on conventional autoregressive moving average (ARMA), thus leading to a more practically sound processing of short-duration signals. It is shown that the proposed time-series modeling based on CPSM can be effectively applied towards the separation of superimposed signals of heavily overlapping spectra. Next, the parameter-invariant nonlinear structural signal representation based on shift-invariant CPSM is presented. The use of this model in the design of annihilation operators (AO) is described, and composite parameter-free structural modeling (CPFSM) is developed. Based on this model, two canonical forms of the parameter-invariant null filters (PINF) are presented, and their use in the suppression of a given class of signals, independently of the values of theira priori unknown parameters, is illustrated. The paper also presents some simulation examples illustrating the application of the proposed CPSM and CPFSM in solving problems of detection and parameter estimation in the presence of highly non-Gaussian, mainly signal-like interferences.

Notes: Abstract This work is motivated by the need forFaithful digital simulation of cellular neural networks (CNNs) that maintains most of their qualitative properties of stability and convergence. An interconnection of nonlinear digital filters mimicking behaviors of the analog CNNs is proposed, and the main properties are studied in detail. The discrete model obtained is proven to have the same convergence properties as the original analog network. The key to this development is the use of anAppropriate discretization scheme. Our discrete approximation to the nonlinear state-space representation of cellular neural networks is such that the Lyapunov function used to show convergence in analog cellular neural networks is still a Lyapunov function (when appropriately discretized) for our nonlinear digital filter network. This is in contrast to other digital simulations of CNNs, which have not been proven to preserve the convergence properties. The network of nonlinear digital filters so introduced thus adds another item to the catalog of digital filters obtained viaappropriate discretization of analog circuits, e.g., wave digital filters, orthogonal filters, and certain other of their more recently studied nonlinear counterparts.

Notes: Abstract In this paper, the problem of blind channel identification and equalization is reviewed, and some recent results are presented.

Notes: Abstract This paper presents in summarizing form a description of halfband filters and the related symmetrical Hilbert transformers. It starts with the two complemetary relations by which halfband filters are defined and the consequences for their impulse responses. The idealized versions of the frequency responses of halfband lowpasses and Hilbert transformers are introduced, and the related tolerance schemes that realized systems must satisfy are described. Using their frequency responses, the transformation of one filter type into the other is presented in general form. The design of finite impulse response (FIR)-halfband filters and their relation to corresponding Hilbert transformers are recalled, using maximally flat and Chebyshev approximations as examples. It is shown that the relation between both types of systems can be used for the infinite impulse response (IIR) case as well. The design of IIR-halfband filters is presented for systems with approximately linear phase and for those with minimum phase again for maximally flat and Chebyshev approximations. The design methods are partly new. The general procedure for the transformation into Hilbert transformers yields noncausal solutions, one of which is already known from the literature. By modifying this operation, phase-splitting systems are obtained, one of them related to corresponding continuous ones, discussed in papers published around 1950. Another system with approximately linear phase corresponds to a paper presented in 1987. Finally, the coupled form of these phase splitting allpasses is found to be a Hilbert transformer with precise phase difference, but with deviations of the magnitudes of the frequency responses.

Notes: Abstract The matched field processing (MFP) localization performance of very low frequency (VLF) arrays operated in the deep ocean basins appears to be limited more by uncertainty in the sound velocity profile (SVP) than by low signal-to-noise ratio (SNR). We present a new robust variation of MFP designed to be less sensitive to velocity error in weakly inhomogeneous environments. We analyze the computational requirements of this and other MFP algorithms. When either the search volume is large of the acoustic array is large, computational efficiency is an issue. We present an efficient MFP implementation for the conventional MFP algorithm and our robust algorithm. We show that parallel implementation of these algorithms may allow real-time performance.

Notes: Abstract By applying results from homotopy theory, new conditions are obtained for the existence and uniqueness of an equilibrium for a class of continuous-time feedback neural networks which contains the Hopfield model as a special case. Next, new criteria are established for the global asymptotic stability of the unique equilibrium of this class of neural networks by utilizing Lur'e-type Lyapunov functions and the stability theory for systems of differential inequalities. Several practical stability testing conditions are given. As a special case, criteria are derived for the global asymptotic stability of Hopfield neural networks. This is followed by a robustness analysis of the class of neural networks considered. The results obtained are then applied to an optimization problem.

Notes: Abstract Functional artificial neural networks (FANNs) are artificial neural networks (ANNs) in which the synaptic weights are “functions” rather than numbers. Thus the signals in such networks are analog, and the action of a synapse on a signal passing through it takes place in the form of a scalar product inL 2 between the functional weight and the signal. In this paper, four classes of FANNs are introduced. They result from the solution of a nonparametric optimization problem in a generalized Fock space (GFS) of abstract Volterra series under interpolating or smoothing input-output training data constraints. Two of these classes of FANNs correspond to the interpolating case and are represented by what we call the (two-layer)optimal interpolating (OI) FANN and theoptimal multilayer neural interpolating (OMNI) FANN. The remaining two classes correspond to the smoothing case. We name their representations as the (two-layer)optimal smoothing (OS) FANN and theoptimal smoothing multilayer artificial neural (OSMAN) FANN. In addition to providing the background and the derivation of these FANNs, this paper presents a novel approach to their implementation. This approach does away with the computationally cumbersome use of functional weights. Instead, the effect of these weights is provided by linear time-invariant differential equation models of which those weights are impulse responses. These are implemented by a linear filter bank. This approach thus leads to simple and meaningful causal realizations of FANNs which we call Dynamical FANNs or simply D-FANNs.

Notes: Abstract In this paper, we study the problem of maximizing an objective function over the discrete set {−1, 1} n using a neural network. It is now known that a binary (two-state) Hopfield network can take, in the worst case, an exponential number of time steps to find even a local maximum of the objective function. In this paper, we carry this argument further by studying theradius of attraction of the global maxima of the objective function. If a binary neural network is used, in general there is no guarantee that a global maximum has a nonzero radius of attraction. In other words, even if the optimization process is started off with the neural network in an initial state that isadjacent to the global maximum, the resulting trajectory of the network may not converge to the nearby maximum, but may instead go off to another maximum. At the same time, another set of recent results shows that, if ananalog neural network is used to optimize the same objective function, thenevery local maximum of the objective function has a nontrivial domain of attraction, and conversely, the only equilibria that are attractive are the local maxima of the objective function. This raises the question as to whether analog neural networks offer some advantages over binary neural networks for optimizing the same objective function. As a motivation for this line of inquiry, we study the problem of decoding an algebraic block code using a neural network. It is shown that the binary neural network implementation has the undesirable property thatall the global maxima of the objective function have azero radius of attraction. In contrast, if an analog neural network is used to maximize exactly the same objective function, the region of attraction of each maximum contains not only the associated “orthant” of the state space, but also some points not in this orthant. In other words, the analog implementation exhibits the desired tolerance to transmission errors, whereas the binary neural network does not have this property. With this motivation, two open questions are posed that provide a program of research for studying the possible superiority of analog neural networks over binary neural networks.

Notes: Abstract We present a new statistical technique for average power estimation in sequential circuits. Because of the feedback loops, power dissipations of sequential circuits in consecutive clock cycles are temporally correlated. The existence of data correlation makes it unsuitable to apply conventional techniques to average power inference, because the sample variance is no longer a maximum likelihood estimator. The convergence criterion derived from the biased variance estimation will be overly optimistic, causing power simulation to stop prematurely at a lower-than-specified estimation accuracy. To overcome this problem, we propose a systematic approach for modeling the power dissipation behavior of sequential circuits as an autoregressive random process. An accurate process variance can be obtained by the model parameters, which enables the derivation of a robust confidence interval of the average power. The interval is checked for convergence against a user-specified accuracy criterion. An iterative procedure is developed to invoke these steps repeatedly until the convergence specification is met. For a set of benchmark sequential circuits, this technique yields high accuracy and efficiency.

Notes: Abstract A potential application of MSW technology lies in the area of time delay for future low-sidelobe wide-bandwidth phased array antennas. High-precision MSW electronically tunable analog time-delay units in transmit/receive modules in phased arrays have the potential of greatly enhancing antenna system capabilities, by increasing instantaneous operating bandwidth and decreasing sidelobe levels, over phased array systems using only phase shifters or switched lines for beam steering and control. This paper provides a status report of MSW time delays for such arrays.

Notes: Abstract This paper presents a straightforward algebraic method for designing feedback loops in the frequency domain. The emphasis here is on control system design, but the technique is applicable to active filter design as well. The object is to produce a practical analog filter with a minimum of design effort. The algebraic solution to the design problem is presented, and several examples are explored.

Notes: Abstract This correspondence presents a RELAX (RELAXation-based) algorithm for angle, polarization, and waveform estimation of completely polarized narrowband electromagnetic plane waves arriving at a uniform linear co-centered orthogonal loop and dipole (COLD) array. We use numerical examples to demonstrate the performance of the RELAX algorithm for parameter estimation with the COLD array and compare it with the performance of the MODE (Method of Direction Estimation) algorithm. We show that this RELAX algorithm can outperform MODE and yet be computationally faster than MODE.