Search

Cuckoo search algorithm (CSA) is relatively a new optimization technique with less control parameters and strong exploration ability. Due to the random search associated with CSA, it requires large number of functional evaluations for obtaining optimal solution. An improved algorithm, named as improved global-best-guided CSA, is presented here based on the best solution of previous iteration for the optimal design of multiplierless two-dimensional recursive digital filters. The most important feature of the proposed algorithm is that it is completely self-adaptive with no tuning parameters, whereas in CSA the replacement factor needs to be adjusted. The proposed algorithm exhibits 52% improvement in fitness function evaluation (for p = 2) and the execution time is reduced by 56% in comparison with the existing algorithms. Further, the proposed algorithm has been tested for several benchmark problems and found to exhibit significant performance improvement. [ABSTRACT FROM AUTHOR]

In this paper, a new sign-normalized least-mean-square adaptive filtering algorithm based on IIR spline adaptive filter (IIR-SAF-SNLMS) is proposed. By using the absolute value of the a posteriori error as the cost function and solving the optimization problem, the proposed algorithm achieves robustness against impulsive noise. Furthermore, to further improve the performance of the IIR-SAF-SNLMS, its variable step-size variant is proposed. Simulation results in the identification of the IIR-SAF nonlinear model show that the proposed algorithms provide better tracking and steady-state performance as compared to the existing spline algorithms. [ABSTRACT FROM AUTHOR]

This paper presents development and implementation of a novel digital signal processing algorithm for on-line estimation of peak value of the fundamental component of a non-sinusoidal power system signal. The algorithm relies on stable band-pass first and second degree digital integrator based rigorous mathematical deduction and recombines itself with zero-crossing avoiding technique. Compared with the well-established technique such as the enhanced phase-locked-loop (EPLL) based system, the proposed algorithm provides higher degree of immunity and insensitivity to harmonics and noise, and faster response. Based on simulation studies, performances of the proposed algorithm at different operating conditions have been presented and its accuracy and response time have been compared with the EPLL systems. Moreover, a simple laboratory setup implemented by MATLAB and dedicated hardware is built to verify the performance of the proposed algorithm in real time applications. The framework of laboratory prototype has been given out and the necessary modification in proposed algorithm for on-line implementation has been discussed. [ABSTRACT FROM AUTHOR]

The arithmetic mean and the order statistical median are two fundamental operations in signal and image processing. They have their own merits and limitations in noise attenuation and image structure preservation. This paper proposes an iterative algorithm that truncates the extreme values of samples in the filter window to a dynamic threshold. The resulting nonlinear filter shows some merits of both the fundamental operations. Some dynamic truncation thresholds are proposed that guarantee the filter output, starting from the mean, to approach the median of the input samples. As a by-product, this paper unveils some statistics of a finite data set as the upper bounds of the deviation of the median from the mean. Some stopping criteria are suggested to facilitate edge preservation and noise attenuation for both the long- and short-tailed distributions. Although the proposed iterative truncated mean (ITM) algorithm is not aimed at the median, it offers a way to estimate the median by simple arithmetic computing. Some properties of the ITM filters are analyzed and experimentally verified on synthetic data and real images. [ABSTRACT FROM AUTHOR]

In this article, we show that when targets are closely spaced, traditional tracking algorithms can be adjusted to perform better under a performance measure that disregards identity. More specifically, we propose an adjusted version of the joint probabilistic data association (JPDA) filter, which we call set JPDA (SJPDA). Through examples and theory we motivate the new approach, and show its possibilities. To decrease the computational requirements, we further show that the SJPDA filter can be formulated as a continuous optimization problem which is fairly easy to handle. Optimal approximations are also discussed, and an algorithm, Kullback–Leibler SJPDA (KLSJPDA), which provides optimal Gaussian approximations in the Kullback–Leibler sense is derived. Finally, we evaluate the SJPDA filter on two scenarios with closely spaced targets, and compare the performance in terms of the mean optimal subpattern assignment (MOSPA) measure with the JPDA filter, and also with the Gaussian-mixture cardinalized probability hypothesis density (GM-CPHD) filter. The results show that the SJPDA filter performs substantially better than the JPDA filter, and almost as well as the more complex GM-CPHD filter. [ABSTRACT FROM PUBLISHER]

Abstract: This paper presents a new recursive algorithm to directly obtain the sum of the autocorrelation functions of the M sequences belonging to a Complementary Set of Sequences (M-CSS). The proposed algorithm allows a digital filter with only one stage to be obtained, what significantly reduces the total number of memory requirements, multiplications and additions to be performed in comparison with the use of M correlators, even when these correlators could have an efficient implementation. The use of this algorithm can be of great relevancy since there is an increasing interest in the use of CSS (or derived sequences such as Loosely Synchronized—LS-sequences or Zero Correlation Zone—ZCZ-sequences) in the field of communications and sensor systems. [Copyright &y& Elsevier]

The discrete pulse transform (DPT) is a decomposition of an observed signal into a sum of pulses, i.e., signals that are constant on a connected set and zero elsewhere. Originally developed for 1-D signal processing, the DPT has recently been generalized to more dimensions. Applications in image processing are currently being investigated. The time required to compute the DPT as originally defined via the successive application of LULU operators (members of a class of minimax filters studied by Rohwer) has been a severe drawback to its applicability. This paper introduces a fast method for obtaining such a decomposition, called the Roadmaker's algorithm because it involves filling pits and razing bumps. It acts selectively only on those features actually present in the signal, flattening them in order of increasing size by subtracing an appropriate positive or negative pulse, which is then appended to the decomposition. The implementation described here covers 1-D signal as well as two and 3-D image processing in a single framework. This is achieved by considering the signal or image as a function defined on a graph, with the geometry specified by the edges of the graph. Whenever a feature is flattened, nodes in the graph are merged, until eventually only one node remains. At that stage, a new set of edges for the same nodes as the graph, forming a tree structure, defines the obtained decomposition. The Roadmaker's algorithm is shown to be equivalent to the DPT in the sense of obtaining the same decomposition. However, its simpler operators are not in general equivalent to the LULU operators in situations where those operators are not applied successively. A by-product of the Roadmaker's algorithm is that it yields a proof of the so-called Highlight Conjecture, stated as an open problem in 2006. We pay particular attention to algorithmic details and complexity, including a demonstration that in the 1-D case, and also in the case of a complete graph, the Roadmaker's algorithm has optimal complexity: it runs in time \cal O(m), where m is the number of arcs in the graph. [ABSTRACT FROM AUTHOR]

Abstract: This article proposes an unbiased plain gradient algorithm for a second-order adaptive IIR notch filter with constrained poles and zeros. The proposed algorithm employs removing a dominant parameter that produces inherent bias. By using this technique, the performances are improved with slight expense in computational complexity. In this paper, theoretical analysis for deriving the estimations of bias and mean square error (MSE) at steady state are presented in closed form. Moreover, the stability bound of the algorithm is also derived. To confirm the analytical results, the computer simulations are provided to corroborate the effectiveness of the proposed algorithm. Furthermore, the performances of the algorithm are also compared with the plain gradient (PG) and modified plain gradient (MPG) algorithms. [Copyright &y& Elsevier]

Abstract: The suppression of late reverberation by spectral subtraction tends to degrade disproportionally low-level signal regions and signal transients. This work proposes two novel relaxation criteria that can constrain such problems in a signal-dependent fashion. These criteria were found to improve the performance of state-of-the-art late reverberation suppression algorithms when used in conjunction with a perceptually motivated non linear filtering stage. Objective results indicate that the proposed method is robust over a wide range of practical reverberation scenarios. [Copyright &y& Elsevier]

Digital Finite Impulse Response (FIR) filters are frequently used in Digital Signal Processing (DSP) by virtue of stability and easy implementation. The problem of designing the multiplier block of a digital FIR filter has received a significant amount of attention during the last decade, as the filters require a large number of multiplications, leading to excessive area, delay, and power consumption even if implemented in a full custom integrated circuit. Previous works have focused on the design of filters with minimum area by replacing the multiplication operations with constant coefficients by addition, subtraction, and shifting operations. Since shifts can be implemented with only wires in hardware, the design problem can be defined as the minimization of the number of addition/subtraction operations to implement the coefficient multiplications. This problem is generally known as the MCM problem. In the last decade, many efficient algorithms have been proposed for the MCM problem. These algorithms can be categorized in two classes: Common Subexpression Elimination (CSE) and graph-based algorithms. While the CSE algorithms basically find the common non-zero digit patterns on the representations of the constants, the graph-based algorithms are not restricted to a particular number representation and synthesize the constants iteratively by building a graph. However, in the algorithms proposed for the MCM problem, an addition/subtraction operation is assumed to be a two-input operation that is generally implemented using a Ripple Carry Adder (RCA) block yielding great latency in the implementation of MCM. In high-speed applications, particularly in DSP systems, Carry-Save Adder (CSA) blocks are preferred to RCA blocks taking into account the increase in area. Despite the large number of algorithms designed for the minimization of addition/subtraction operations based on RCAs, there are only a few algorithms proposed for the optimization of the number of CSA blocks. Although these algorithms give good results, they are based on heuristics, i.e., provide no indication on how far from the minimum their solutions are. To the best of our knowledge, there is no exact algorithm proposed for the optimization of the number of CSA blocks in MCM. In this paper, an exact CSE algorithm designed for the minimization of the number of CSA blocks is introduced. In the exact CSE algorithm, initially, all possible implementations of filter coefficients and partial terms using CSAs are found when filter coefficients are defined under a number representation, and a combinational network that represents the implementations of constants is constructed. Then, the minimization of the number of CSA blocks problem is defined as a 0-1 Integer Linear Programming (ILP) problem with a cost function to be minimized and constraints to be satisfied.… [ABSTRACT FROM AUTHOR]

Abstract: The digital processing of chromatographic thin-layer plate images has increasing popularity among last years. When using a camera instead of flatbed scanner, the charged coupled device (CCD) noise is a well-known problem—especially when scanning dark plates with weakly fluorescing spots. Various techniques are proposed to denoise (smooth) univariate signals in chemometric processing, but the choice could be difficult. In the current paper the classical filters (Savitzky–Golay, adaptive degree polynomial filter, Fourier denoising, Butterworth and Chebyshev infinite impulse response filters) were compared with the wavelet shrinkage (31 mother wavelets, 3 thresholding techniques and 8 decomposition levels). The signal obtained from 256 averaged videoscans was treated as the reference signal (with noise naturally suppressed, which was found to be almost white one). The best choice for denoising was the Haar mother wavelet with soft denoising and any decomposition level larger than 1. Satisfying similarity to reference signal was also observed in the case of Butterworth filter, Savitzky–Golay smoothing, ADPF filter, Fourier denoising and soft-thresholded wavelet shrinkage with any mother wavelet and middle to high decomposition level. The Chebyshev filters, Whittaker smoother and wavelet shrinkage with hard thresholding were found to be less efficient. The results obtained can be used as general recommendations for univariate denoising of such signals. [Copyright &y& Elsevier]

From a metrological point of view, a measurement process rarely consists of a direct measurement but, rather, of digital signal processing (DSP) performed by one or more instruments. The measurement algorithm makes the numerical results available as functions of acquired samples from input signals. Moreover, when repeated direct measurements are performed, one may speak about interactions in subsequent results (and it may be dependent on the type of instrument being used). With mathematical formalism, the complex relations involved can be described, although again, an indirect measurement result would be obtained. Regardless, no matter what kind of process is being examined, the distribution of the uncertainty associated with the measurement needs to be known. To express a measurement result with its associated uncertainty, the recommendations of the ISO Guide need to be met. Many published papers have proposed the use of fuzzy intervals to describe both the systematic and statistical effects of repeated measurements on the distribution of their results. In this paper, we use a random-fuzzy model, the single measure is represented as a fuzzy set, and the propagation of the possibility distribution through the DSP stage (which simply consists of an average operation) is performed using the extension principle of Zadeh based on a particular triangular norm: the so-called Dombi's. [ABSTRACT FROM AUTHOR]

We propose a method for determining the time delay of frequency-shift keyed signals on the basis of the digital filtering algorithm insensitive to carrier-frequency shifts. The filter coefficients correspond to the information-optimal solution of the problem of variance minimization at the linear-filter output. Efficiency of the proposed method is shown under conditions of additive noise and inexact knowledge of the center frequency of the processed-signal spectrum. The developed algorithm permits simple realization for real-time operation on the basis of a digital signal processor. [ABSTRACT FROM AUTHOR]

An algorithm that is suitable for the computation of symmetrical components is presented in this paper. The structure of the algorithm consists of four decupled linear modules: the first for adaptive filtering of input signals, the second for frequency estimation, the third for symmetrical components estimation, and the fourth for the amplitudes of the component estimation. Filters are used to minimize the noise effect and to eliminate the presence of harmonics. A simple method for online designing of digital filters of sinusoidal signals is used in this paper. This method uses closed forms for calculation of filter coefficients and requires modest computations. The symmetrical components estimation through adaptive transformation matrix of phase shifters used in this paper makes it possible to get instantaneous symmetrical components independently on the frequency variation with modest-computation requirements. The technique used in this paper provides accurate frequency estimates with the maximum error of 0.002 Hz and symmetrical components estimation with the maximum error of 0.03% for SNR = 60 dB in about 25 ms. The algorithm is not sensitive to power-system frequency changes and to the harmonic distortion of the input signals. To demonstrate the performance of the developed algorithm, computer-simulated data records have been processed. It has been found that the proposed algorithm is suitable for real-time applications. [ABSTRACT FROM AUTHOR]

New algorithm for measurement of mean count rate of pulses from radiation detectors is described. The algorithm is based on the control of a bank of digital IIR filters governed by the state of mean count rate—stationary or transient. The algorithm solves successfully the compromise between the contradictory requirements for low statistical fluctuations in steady state and fast response to sudden changes of mean count rate. The design of the bank of HR filters is carried out by using the Chebyshev approximation of the second kind in order to achieve the high selectivity and high stopband attenuation applying as low as possible order of the constituent filters. All designed filters are of the second order. The static and dynamic characteristics of the new algorithm and the corresponding computational efficiency are tested and compared with those of the traditional ratemeter algorithms. The state of mean count rate, stationary or transient, is determined by a circuit for estimation of the gradient of mean count rate. The numbers of threshold levels within the adaptation logic circuit and their mutual positions have been optimized to achieve the best compromise. [ABSTRACT FROM AUTHOR]

In this paper, efficient architectures for realizing the recursive discrete cosine transform (DCT) and the recursive inverse DCT (IDCT) are proposed. By respectively folding the inputs of the DCT and the outputs of the IDCT, efficient formulations of the DCT and IDCT are derived to construct a transform kernels. The data throughout per transformation is twice that of the existing methods by spending only half of the computational cycles used by the single folding algorithms. To further improve efficiency, the double folding recursive architectures of the DCT and IDCT are developed. The computational cycles of the DCT are half of the single folding method, and the data throughout of the IDCT is twice that of the single folding method. The regular and modular properties of the proposed recursive archietectures are suitable for very scale integration (VLSI) implementation. With high throughout advantage, the proposed structures could be implemented with less power consumption, which could be applied to low rate video in mobile and portable information appliances. [ABSTRACT FROM AUTHOR]

It is known that linear phase FIR digital filters fall into four types considering the symmetry of impulse response. For example, filters of type 2 can be used at the time of partitioning and reconstruction of signals in multirate signal processing, and other types have similar special applications. Moreover, the linear phase FIR digital filters with flat passband and equiripple stopband characteristics are very important filters in practice because they have better time domain characteristics and less ringing than Chebyshev filters with equiripple stopband and passband characteristics. However, conventional design methods are restricted to type 1. In this paper, we therefore propose a method of design of linear phase FIR digital filters with flat passband and equiripple stopband characteristics for all four types. The proposed function is composed of two functions, one with a flat characteristic in the passband and the second with equiripple characteristics in the stopband. The latter function is obtained by using the well-known Remez algorithm only for the stopband. © 2002 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 86(4): 89–95, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.1148 [ABSTRACT FROM AUTHOR]

This paper proposes the parallel construction of the second-order infinite impulse response (IIR) adaptive digital filter (ADF) as an adaptive digital filter for eliminating multiple narrowband signals mixed in the broadband signal. Then the pole-control method for the filter is proposed. In the proposed pole-control, the pole radius and the pole argument are controlled separately by different algorithms. In this study, the pole argument is corrected by estimating the deviation of the central frequency of the signal to be eliminated from the convergent value of the multiplier coefficient that determines the zero. The pole radius is controlled by the method that determines the direction of correction, using the parallel-connected three ADFs with the same pole argument and a slightly different pole radius. It is verified by a simulation that those pole-control methods work effectively. It is demonstrated that the pole radius and the pole argument can be controlled simultaneously. [ABSTRACT FROM AUTHOR]

The unifying treatment of a general gradient adaptive algorithm studied here is based on the principles of minimum mean square method and instrumental variable method by formulating the general oblique projection algorithm, including finite impulse response (FIR) and serial-parallel infinite impulse response (IIR)-type configurations. The oblique projection algorithm of low-order past-value data space is derived in which the dimension of past-due data space can be easily adjusted, including the least mean square (LMS) method, learning-identification method and gradient instrumental variable method (GIVE). Next, the new analysis model which describes the convergence process of general gradient algorithm is shown and the convergence conditions are indicated for the first and second moment with arbitrary inputs. It is found that this condition is more severed as compared with the white input. Moreover, the process of convergence of these conditions is studied by simulation. [ABSTRACT FROM AUTHOR]