Papers by Vesna Popović-Bugarin
Credit Card Fraud Detection Using Supervised Learning Algorithms
Signal Decomposition of Micro-Doppler Signatures
Institution of Engineering and Technology eBooks, May 30, 2014
The micro-Doppler effects detection, estimation, and removal are analyzed in this chapter. Since ... more The micro-Doppler effects detection, estimation, and removal are analyzed in this chapter. Since the most common form of the micro-Doppler signal is a sinusoidal modulated one, a simple technique for its analysis, based on the inverse Radon transform is presented. A method for period estimation is presented as well. The presented method is important for the micro-Doppler analysis itself, as well as for the technique based on the inverse Radon transform. In addition, it is shown that a general micro-Doppler form can be separated from the rigid body, by using TF analysis and the L-statistics. Finally, the micro-Doppler effect in high noise environments is analyzed by using the Viterbi algorithm.
DBSCAN and CLARA Clustering Algorithms and their usage for the Soil Data Clustering
In this paper two clustering algorithms DBSCAN and CLARA were applied over the pedological databa... more In this paper two clustering algorithms DBSCAN and CLARA were applied over the pedological database of Montenegro. Both algorithms clusterize data based on their density distribution. DBSCAN enables discovering clusters of arbitary shapes, without domain knowledge. On the other hand, CLARA forms clusters of approximatly equal size and shape for databases with uniformly spaced data. The used databases is composed of chemical and mechanical-physical parameters of soil samples. There are no clear transitions between different types of soil and large differences in values of their parameters at the boundary points of the clusters. Thus, CLARA is proved to be better for clustering pedologic data, which is confirmed by means of simulations. The results obtained by the CLARA are comparable with the results obtained by the analysis of soil in Montenegro by the expert.
A problem of soil clustering based on the chemical characteristics of soil, and proper visual rep... more A problem of soil clustering based on the chemical characteristics of soil, and proper visual representation of the obtained results, is analysed in the paper. To that aim, K-means and fuzzy K-means algorithms are adapted for soil data clustering. A database of soil characteristics sampled in Montenegro is used for a comparative analysis of implemented algorithms. The procedure of setting proper values for control parameters of fuzzy K-means is illustrated on the used database. In addition, validation of clustering is made through visualisation. Classified soil data are presented on the static Google map and dynamic Open Street Map.
Application of Apriori Algorithm for CRM Improvement - Case Study from Montenegro
S-Method In Radar Imaging
Publication in the conference proceedings of EUSIPCO, Florence, Italy, 2006
Soil and Water Research, 2018
This paper describes the process of digitizing Montenegro’s legacy soil data, and an initial atte... more This paper describes the process of digitizing Montenegro’s legacy soil data, and an initial attempt to use it for digital soil mapping (DSM) purposes. The handwritten legacy numerical records of physical and chemical properties for more than 10 000 soil profiles and semi-profiles covering whole Montenegro have been digitized, and, out of those, more than 3000 have been georeferenced. Problems and challenges of digitization addressed in the paper are: processing of non-uniform handwritten numerical records, parsing a complex textual representation of those records, georeferencing the records using digitized (scanned) legacy soil maps, creating a single computer database containing all digitized records, transforming, cleaning and validating the data. For an initial assessment of the suitability of these data for mapping purposes, inverse distance weighting (IDW), ordinary kriging (OK), multiple linear regression (LR), and regression-kriging (RK) interpolation models were applied to ...
IEEE Access, 2018
The method for detection of complex sinusoids in additive white Gaussian noise and estimation of ... more The method for detection of complex sinusoids in additive white Gaussian noise and estimation of their frequencies is proposed. It contains two stages: 1) sinusoid detection (model order estimation) and coarse frequency estimation, and 2) fine frequency estimation. The proposed method operates in the frequency domain, i.e., it uses the discrete Fourier transform (DFT) as the main tool. Sinusoid detection is performed so that a fixed probability of false alarm is provided (Neymann-Pearson criterion). For both coarse and fine frequency estimations, the three-point periodogram maximization approach is used. Simulations are carried out for variable signal-to-noise ratio, variable frequency displacement between the sinusoids and variable offset from the frequency grid. The proposed method meets the Cramér-Rao lower bound in frequency estimation and practically does not depend on the frequency displacement except for very small displacement values. In terms of model order estimation accuracy, it outperforms the state-of-the-art approaches. The most expensive operation in the method is the calculation of the DFT. Therefore, in terms of calculation complexity, the proposed method is on par with the most efficient algorithms for multiple frequency estimations. INDEX TERMS Cramér-Rao lower bound, discrete Fourier transform, model order estimation, multiple frequency estimation, Neymann-Pearson criterion.
Efficient instantaneous frequency estimation in high noise based on the Wigner distribution
Signal Processing, 2018
Abstract Instantaneous frequency (IF) estimation of signals embedded in high noise is considered.... more Abstract Instantaneous frequency (IF) estimation of signals embedded in high noise is considered. A time-frequency (TF) approach based on the Wigner distribution (WD) is proposed, i.e. the WD is processed in order to enable accurate IF estimation of both monocomponent and multicomponent signals in high noise environment. To that end, a two-step procedure is introduced. In the first step, the auto terms are emphasized with respect to noise. In the second step, the influence of inner interferences, cross terms and noise is suppressed, yielding the final TF representation. The accuracy performance of the proposed method, validated on simulated and real data, is on par with the state-of-the-art methods. However, it is characterized by significantly lower calculation complexity.
The STFT-Based Estimator of Micro-Doppler Parameters
IEEE Transactions on Aerospace and Electronic Systems, 2017
A two-stage technique for estimating micro-Doppler signal parameters has been proposed. In the fi... more A two-stage technique for estimating micro-Doppler signal parameters has been proposed. In the first stage, rough parameter estimations are performed by regression of instantaneous frequency estimate obtained from the short-time Fourier transform. Afterwards, rough estimates are refined in the second stage. The proposed technique has better performance with respect to current state-of-the-art algorithms, and it reaches the Cramer–Rao lower bound for sinusoidal frequency-modulated signal parameters.
Proceedings of the 3rd IEEE International Symposium on Signal Processing and Information Technology (IEEE Cat. No.03EX795)
Capon's method is widely used in the literature in array signal processing and spectral analysis.... more Capon's method is widely used in the literature in array signal processing and spectral analysis. Recently, this method has been extended to the time-frequency analysis. The topic of this paper is to compare the standard spectrogram and its Capon's counterpart, with respect to the several basic parameters: instantaneous frequency estimation, distribution concentration, and resolution of two close components. The analysis is generalized to the distributions from the Cohen class and distributions that take into account phase nonlinearity.
2015 4th Mediterranean Conference on Embedded Computing (MECO), 2015
Instantaneous frequency estimation of signals in a high noise environment is analyzed in the pape... more Instantaneous frequency estimation of signals in a high noise environment is analyzed in the paper. An algorithm based on Ant colony optimization and Wigner distribution is proposed for solving the considered estimation problem. The proposed approach has been applied and tested on mono-component frequency-modulated signals. Numerical examples are given in order to demonstrate the algorithm's performances in the analyzed framework.
IEEE Transactions on Aerospace and Electronic Systems, 2015
A method for accurate and efficient parameter estimation and decomposition of sinusoidally freque... more A method for accurate and efficient parameter estimation and decomposition of sinusoidally frequency modulated signals is presented. These kinds of signals are of special interest in radars and communications. The proposed method is based on the inverse Radon transform property to transform a twodimensional sinusoidal pattern into a single point in a twodimensional plane. Since the signal is well concentrated (sparse) in the inverse Radon transform domain its reconstruction can be performed from a reduced set of observations (back-projections). Theory is illustrated on signals with one and more components, including noise and disturbances, as well as time-frequency patterns that deviate from sinusoidal form.
We propose a method for the phase estimation of a polynomial-phase signal (PPS). Using an approac... more We propose a method for the phase estimation of a polynomial-phase signal (PPS). Using an approach based on the high-order instantaneous moment (HIM) of the PPS, the estimation of phase coefficients boils down to the sinusoid frequency estimation. Since the desired complex sinusoid is embedded in a heavy-tailed noise, standard periodogrambased techniques cannot be used for the frequency estimation. Instead of the standard periodogram, we use the robust M-periodogram, where a non-quadratic loss function is used for fitting of observations corrupted by noise with unknown heavy-tailed distribution. The estimation accuracy is additionally improved using an iterative procedure based on the dichotomous search of periodogram peak. Simulations carried out with several common noise distributions confirm the superiority of the proposed method over the standard one.
Time-Frequency Analysis for Sar and Isar Imaging
NATO Science for Peace and Security Series C: Environmental Security, 2009
In the past 10 years the Time-Frequency (TF) techniques have found significant importance for imp... more In the past 10 years the Time-Frequency (TF) techniques have found significant importance for improvement and refinement of radar images (both SAR and ISAR) and in extraction of objects' features from radar images. The main contributions in this direction are given by V. Chen (Chen and Ling, 2002). Here, we will concentrate our attention on the radar imaging and two
Noise analysis of the high resolution methods in ISAR
ISPA 2005. Proceedings of the 4th International Symposium on Image and Signal Processing and Analysis, 2005., 2005
ABSTRACT Commonly used technique for the inverse synthetic aperture radar (ISAR) signal analysis ... more ABSTRACT Commonly used technique for the inverse synthetic aperture radar (ISAR) signal analysis is a two dimensional Fourier transform. Short time intervals could be used in order to reduce nonstationarity effects. However, resolution of images obtained by using short intervals is highly limited. In this case superresolution techniques, such as Capon, ESPRIT and MUSIC, should be used. All high resolution methods are quite sensitive to the noise. The main topic of this paper is to find a relation between maximum SNR and resolution when 2D Capon method is used in ISAR. 2D Capon method's sensitivity to the noise is analyzed and compared to the 2D Fourier transform application, in case of well separated scattering centers. In order to improve performance of the standard Capon method a modified version with the known number of components is considered.
Sar Images Improvements by Using The S-Method
2006 IEEE International Conference on Acoustics Speed and Signal Processing Proceedings
Synthetic aperture radar processors are generally made with the stationary targets in mind, hence... more Synthetic aperture radar processors are generally made with the stationary targets in mind, hence commonly used technique for the SAR signal analysis is a two-dimensional Fourier transform. Moving targets induce Doppler-shift and Doppler spread in the returned signal, producing blurred or smeared images. Standard techniques for these kinds of the problems are motion compensation and time-frequency analysis application. Both of
Signal Processing, 2011
The problem of non-stationary interference suppression in direct sequence spread-spectrum (DS-SS)... more The problem of non-stationary interference suppression in direct sequence spread-spectrum (DS-SS) systems is considered. The phase of interference is approximated by a polynomial within the considered interval. According to the local polynomial Fourier transform (LPFT) principle, the received signal is dechirped by using the obtained phase approximation and the interference is, in turn, suppressed by excising the corrupted low-pass frequency band. For the estimation of polynomial coefficients, we use the product high-order ambiguity function (PHAF), known for its capability to successfully resolve components of a multicomponent polynomial-phase signal (PPS). The proposed method can suppress interferences with both polynomial and non-polynomial phase. In addition, it can suppress both monocomponent and multicomponent interferences. The simulations show that the proposed method outperforms timefrequency (TF) methods, that successfully deal with multicomponent interferences, in terms of the error probability and computational complexity.
An efficient peak frequency estimator for product high-order ambiguity function
2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2011
In this paper, we consider improvement of parameter estimation accuracy of multicomponent polynom... more In this paper, we consider improvement of parameter estimation accuracy of multicomponent polynomial-phase signals (mc-PPSs) using the product high-order ambiguity function (PHAF). We propose a fine search method that improves the coarse estimate obtained as the position of maximum PHAF bin. The method is iterative and based on the dichotomous search. It significantly reduces the calculation complexity compared to the
Signal Processing, 2010
The local polynomial Fourier transform (LPFT) based algorithm for autofocusing SAR images has rec... more The local polynomial Fourier transform (LPFT) based algorithm for autofocusing SAR images has recently been proposed by the authors. It produces a well focused image of moving targets, without defocusing stationary targets or inducing undesired cross-terms. The drawback of this algorithm is its high computational burden caused by the large number of elements in the set of used chirp-rates. We propose an algorithm with decreased number of elements used for the LPFTbased SAR imaging. The product high-order ambiguity function (PHAF) is applied to estimate parameters of a radar signal. The estimated chirp-rate is used as an initial value for forming the set of chirp-rates. The proposed algorithm has significantly smaller set of chirprate values (tens comparing to several hundreds or thousands used in the previous algorithm version). In this manner, the calculation complexity is significantly reduced. The proposed procedure is fully automated, meaning that it follows the change of motion parameters. In addition, our procedure considers the third-order phase compensation.
Uploads
Papers by Vesna Popović-Bugarin