Russian
Scientific & Technical
Journal

POTENTIAL LIMIT AND QUALITY INDICATORS OF SUPER-RAYLEIGH SIGNAL RESOLUTION
Klimov S.A., zuk_house@rambler.ru

In this article the directions of search for potential limit of super-Rayleigh resolution are systematized. The simple and convenient resolution quality indicator system for performance evaluation of full signal resolution is established. The question of existence of super-Rayleigh resolution potential limit in solving radar detection-resolution problem is answered. It is shown, that resolving curves of single signals are determined rigidly and they cannot be exceeded if priori statistical information is given. Then full resolution probability curves determine the limit of resolution for fixed false alarm probability and coefficient of correlation of signals. The degree of approximation to potential limit of super-Rayleigh resolution for a large number of digital spectral analyses algorithms is evaluated by using simulation modeling.

Keywords: resolution, super-Rayleigh resolution, resolution limit, quality indicators of signal resolution, digital spectral analyses.

DIGITAL PROCESSING OF RADAR PULSE-TRAIN WAVEFORMS USING LOW-ORDER WEIGHT FUNCTIONS
Zaytsev G.V., gennady-zaytsev@yandex.ru

The paper analyzes processing algorithm of pulse-train waveforms in digital radar receiver. The algorithm consists of range gating and doppler-filter-bank calculation in each range gate. Filter bank is realized by spectral analysis with preliminary weighting to reduce filter sidelobes. Several versions of spectrum calculations are discussed.
Each waveform with specific duration requires specific weight function. So for radars with wide set of pulse trains the number of weighting coefficients to be stored becomes unacceptable. Main part of the paper tackles this problem. To simplify implementation the paper suggests using synthesized by the author weight functions with simple cosine-sum representation, given roll-off rate, and minimum sidelobe level for the given width of the main lobe. Simple representation allows calculating weights in real time during processing instead of their storage. Any of these functions proved to be near-optimal for wide set of waveform parameters. For further simplification, step weight function are examined with step length being equal to pulse repetition period. It is shown that for suggested functions with large roll-off rate transition to step functions leads to insignificant performance degradation.

Keywords: quasicontinuous signal, optimum filtering, weight function, digital signal processing.

OPTIMIZATION OF COHERENT-WEIGHT PROCESSING SYSTEMS OF MULTIFREQUENCY SIGNALS
Popov D.I., adop@mail.ru

The methods of optimization of coherent-weight processing systems of multifrequency radar signals on probabilistic criterion are considered. The block diagrams of adaptive processing systems at continuous and discrete modes are given. The analysis of processing systems of multifrequency signals is carried out.

Keywords: adaptation, probabilistic criterion, moving targets, multifrequency signals, optimization, clutter, processing systems, blind speeds.

IMPLEMENTATION OF STOCHASTIC SIGNAL PROCESSING ALGORITHMS IN REAL ARITHMETIC
Ratynsky M.V., Petrov S.V., m3v5r7@inbox.ru, petrovsv@lest.ru

Four problems of stochastic signal processing in receiving system with antenna array are discussed: detection, enumeration and arrival direction estimation of signal sources and adaptive space filtering. The solution of the problems is implemented generally in complex arithmetic. It is proved that solution of all the problems can be implemented in real arithmetic when antenna array has central symmetric structure. The results of complex and real solutions are really identical for all cases, but in the second case (real arithmetic) the amount of computations is four times less as compared with the first one (complex arithmetic).

Keywords: signal processing algorithms, stochastic signals, signal detection, sources enumeration, arrival direction estimation, adaptive space filtering, real arithmetic.

A NEW METHOD OF FINDING PROBABILITY CHARACTERISTICS IN THE OUTPUT OF NONLINEAR SYSTEMS
Bartenev V.G., Bartenev M.V., syntaltechno@mail.ru

The paper presents a new method of finding probability characteristics in the output of nonlinear systems using elliptical properties of some distributions. Elliptically symmetric distributions are second-order distributions with probability densities whose contours of equal height are ellipses. This class includes the Gaussian, Laplace distributions and others which can be generated from certain first-order distributions. Such desirable features for the description of the second-order statistics may be used in the transformation analysis of a random signal by the nonlinear devices. These elliptically symmetric distributions simplify the evaluation of the output probability characteristics. That is the base of the suggested method in this paper which is illustrated by real losses of detection with analytically developed low false alarm rate of nonlinear multiplier with small integrated samples when the output distribution differs substantially from Gaussian.

Keywords: Elliptically symmetric are second-order distributions, the output probability characteristics of the nonlinear devices, detection probability, false alarm probability, multiplier - integrator.

METHODS AND RESULTS SYNTHESIS OF APERIODIC BINARY SEQUENCES AND POLYPHASE BARKER SEQUENCES
Potekhin E., Shuvalov A., Leukhin N., potegor@yandex.ru, shuvalov.as@gmail.com

The article provides information about the methods of generation, building and searching for optimal binary sequences till length 76 and multiphase Barker sequence till length 52. Search algorithms and analytical conclusions are described in detail. Examples of sequences and their basic characteristics such as merit factor and linear complexity are given.

Keywords: binary sequences, Barker code, the autocorrelation function, the side lobes, multiphase sequences.

REDUCING COMPLEX NOISES IN SPEECH USING FILTERING IN MODULATION DOMAIN
Azarov I.S., Vashkevich M.I., Lihachev D.S., Petrovsky A.A., azarov@bsuir.by, vashkevich@bsuir.by, palex@bsuir.by

The paper presents a method for noise reduction which is capable of handling noises of different nature such as white, colored quasiperiodicor reverberation. This all-purpose method is based on filtration in modulation domain. In order to obtain modulation signals source speech is decomposed into overlapping analytical subband signals using a filter bank. Then each analytical signal is parametrized in terms of instantaneous amplitude and frequency. Subchannel samples are grouped by instantaneous frequency values and their overall intensity envelopes are evaluated. Envelopes are filtered with a band-pass filter which retains speech modulations and attenuates noise modulations. Finally subband gains are calculated and applied to subchannel signals and the cleared speech signal is reconstructed. The paper contains experimental results that show that the proposed method has a better performance than its known counterpart based on RASTA (RelAtive SpecTrA).

Keywords: speech signal, filtering, modulation area, noise reduction, synthesis, evaluation of quality.

SOME ISSUES OF SPECTRAL EFFICIENCY INCREASING OF COMMUNICATION SYSTEMS: NONORTHOGONAL TRANSMISSION
Bakulin M.G, Kreindelin V.B., Shumov A.P., m.g.bakulin@gmail.com, vikrend@gmail.com, magnit2250@mail.ru

Extension of Mazo's concept of faster-than-Nyquist (FTN) signaling to pulse trains that modulate a bank of subcarriers is considered. This method corresponds to two dimensional FTN signaling. The signal processing is similar to orthogonal frequency division multiplex (OFDM) transmission but the subchannels are not orthogonal. Despite nonorthogonal pulses and subcarriers, the method achieves the isolated-pulse error performance and it does so in as little as half the bandwidth of ordinary OFDM system.

Keywords: wireless systems, nonorthogonal transmission, spectral efficiency, OFDM, subcarriers bank, Mazo concept.

THE METHOD FOR PITCH FREQUENCY ESTIMATION IN PRESENCE OF NOISE
Gai V.E., iamuser@inbox.ru

A pitch frequency estimation algorithm in presence of noise is proposed. The algorithm can be used in the development of speaker identification systems. Results of research proving the effectiveness of the proposed algorithm are presented.

Keywords: speaker identification, speech signal, pitch frequency, active perception theory, interference, frequency estimate.

ADAPTIVE ALGORITHM OF RECURSIVE DIGITAL FILTER PARAMETER DETERMINING FOR ACOUSTIC FEEDBACK COMPENSATION
Kondratyev K.V., Uglev V.A., Nepomnyaschy O.V., Sergeevich V.N., kondratyev-37@mail.ru, uglev-v@yandex.ru

This article considers the question of the audio signal adaptive filtering with a digital filter that suppresses acoustic feedback. The algorithm for adjusting the parameters of the recursive digital filter is put forward, this algorithm allows to take into account the changing dynamics of the acoustic environment in the room. Filter characterization is realized each time while in the acoustic path information gap lasting of the order of 1-2 seconds is changing wanted signal. The results of testing the algorithm for signal processing with a loudspeaker system showed that high efficiency of acoustic feedback suppression (up to 93%). It is also considered the prospect of the proposed filter application for the advanced systems of man-machine interaction.

Keywords: audio, sound pressure, an adaptive filter, a recursive algorithm, noise compensation, correlation function.

If you have any question please write: info@dspa.ru