RU2286583C1 - Method for detection and localization of composite signals - Google Patents

Abstract

FIELD: radio engineering, applicable in passive systems of radio monitoring for detection and localization in frequency, time, azimuth and angle of elevation of composite signals in the conditions of priory uncertainty in the form and range of existence.

SUBSTANCE: the result is attained due to optimization of formation of amplitude-phase distribution in the stages of detection and localization of signals, matching of frequency-time ranges tied up by the power spectrum of the detected signal and correlated amplitude-phase distributions, and comparison of the directions of income of signals formed by the algorithms of high resolution of various sub-classes. As a result, the loss of signal power is minimized at a detection and direction finding, as well as abnormal errors of direction finding are precluded and, as a result, the discernibleness of composite monofrequency and single-frequency signals with a low spectral power density on the background of noise and interference is essentially enhanced.

EFFECT: enhanced efficiency of detection and localization in frequency, time, azimuth and angle of elevation the three main classes of composite signals.

5 cl, 10 dwg

Description

The invention relates to radio engineering and can be used in passive radio monitoring systems for detection and localization in frequency, time, azimuth and elevation of complex signals.

With the advent and improvement of communication, location and control systems that use complex signals with a large base, the radiated power of which is specially distributed in the time-frequency domain (wideband single-frequency pseudo-random signals and multi-frequency signals with frequency jumps), problems arise of their effective radio monitoring.

A known method for the detection and localization of complex signals [1], including:

signal reception by two spatially separated receiving channels;

correlation in the time domain of the output signals of the receivers and the formation of a signal describing the mutual correlation function of the signal received by two receivers;

filtering a signal describing the mutual correlation function in the time domain, and highlighting only the central part of the mutual correlation function;

the conversion of the selected central part of the mutual correlation function into a complex mutual spectral density of the signal;

measuring the angle of the phase slope of the mutual spectral density to determine the direction of arrival of the signal;

indication of signal detection results.

The limited spatial selectivity significantly reduces the effectiveness of this method in practice.

More effective is the method of detection and localization of complex signals [2], adopted as a prototype and includes:

1. Coherent reception of time signals x _n (t) by antennas included in the N - element array, where n = 1, (N - antenna number;

2. Synchronous conversion of the received signals x _n (t) into digital signals x _n (z), where z is the number of time reference signal;

сигнала каждой антенны, где q - номер временного отрезка преобразования,

, a l - номер частотного отсчета,

.3. Time-shifting conversion of digital signals x _n (z) to obtain a complex spectral density with a given discreteness in time and frequency

the signal of each antenna, where q is the number of the conversion time span,

, al is the number of the frequency reference,

.

In other words, the input signals at each time interval are divided into frequency channels.

N-мерных векторов спектральных плотностей

с элементами

;As a result of this operation, a time-frequency matrix is formed

N-dimensional spectral density vectors

with elements

;

антенны, выбранной в качестве опорной, и спектральных плотностей остальных антенн

для формирования одночастотного комплексного дискретного амплитудно-фазового распределения (АФР)

, сигнала, зарегистрированного в q-м временном интервале.4. Multiplication at each discrete frequency l of the complex conjugate spectral density

the antenna selected as the reference, and the spectral densities of the remaining antennas

for the formation of a single-frequency complex discrete amplitude-phase distribution (AFR)

, a signal recorded in the qth time interval.

(одночастотных комплексных дискретных АФР

, представляющих собой N-мерные векторы с элементами

. В свою очередь величина

может рассматриваться как одночастотная комплексная дискретная радиоголограмма. В результате формируются АФР для каждого частотно-временного элемента, отличающегося положением (q, l) в анализируемой частотно-временной области.This operation can be considered as the formation of a time-frequency matrix

(single-frequency complex discrete AFR

representing N-dimensional vectors with elements

. In turn, the value

can be considered as a single-frequency complex discrete radio hologram. As a result, an AFR is formed for each time-frequency element that differs by position (q, l) in the analyzed time-frequency region.

5. Determination of the complex spatial cross-correlation coefficients of a single-frequency AFR obtained in each frequency channel with single-frequency AFR obtained in the remaining frequency channels of the reception band in the qth time interval.

элементы которой описываются формулой

где

- нормированные элементы вектор-строки

с элементами

a r - номер частотного канала,

l≠r. При этом использовано обозначение скалярного произведения и нормы N - мерных комплексных векторов в виде

As a result of this operation, a block vector row is formed of complex cross-correlation coefficients

elements of which are described by the formula

Where

- normalized elements of a row vector

with elements

ar is the number of the frequency channel,

l ≠ r. We used the notation of the scalar product and the norm of N - dimensional complex vectors in the form

с порогом и объединение сигналов с частотами, на которых превышен порог, в i-й сигнал, который идентифицируется как обнаруженный сигнал с полосой частот δƒ_i, если полоса δƒ_i непрерывна, или как многочастотный сигнал с полосой частот δƒ_i, если полоса δƒ_i дискретно-непрерывна, принадлежащий одному передатчику с полосой частот δƒ_i, где i=1...Р, а Р - число обнаруженных передатчиков из числа одновременно попадающих в текущую полосу приема;6. Comparison of the modules of the correlation coefficients

with a threshold and combining signals with frequencies at which the threshold is exceeded, into the i-th signal, which is identified as a detected signal with a frequency band δƒ _i if the band δƒ _{i is} continuous, or as a multi-frequency signal with a frequency band δƒ _i if the band δƒ _i discrete-continuous, belonging to one transmitter with a frequency band δƒ _i , where i = 1 ... P, and P is the number of detected transmitters from the number simultaneously falling into the current reception band;

где a_l - двоичные числа (0, 1), отличные от нуля в полосе частот δƒ_i,

- индекс, соответствующий средней частоте сигнала с шириной спектра δƒ_i;7. Averaging single-frequency AFR of the i-th signal in the frequency band δƒ _i and obtaining the average AFR

where a _l are binary numbers (0, 1) that are nonzero in the frequency band δƒ _i ,

- the index corresponding to the average frequency of the signal with the width of the spectrum δƒ _i ;

i-го сигнала для определения реальной части его двумерного комплексного углового спектра

где d_n(m, k) - диаграмма направленности n-й антенны, m=0...М-1 - текущий номер узла сетки по азимуту, М - число узлов по азимуту, k=0...К-1 - текущий номер узла сетки наведения решетки по углу места, К - число узлов по углу места, a

- модельная фазирующая функция, зависящая от конфигурации антенной решетки;8. Use of averaged AFR

ith signal to determine the real part of its two-dimensional complex angular spectrum

where d _n (m, k) is the radiation pattern of the nth antenna, m = 0 ... M-1 is the current number of the grid node in azimuth, M is the number of nodes in azimuth, k = 0 ... K-1 - the current node number of the grid pointing the lattice in elevation, K is the number of nodes in elevation, a

- model phasing function, depending on the configuration of the antenna array;

двумерного комплексного углового спектра.9. Determination of azimuthal and elevation bearings of the i-th signal detected in the reception band, to the maximum of the real part

two-dimensional complex angular spectrum.

The basis of this method is the estimation of the angular proximity of the individual components of the field of the radio emission source using the integrated cross-correlation of single-frequency AFR. The prototype method is effective in the detection and localization of multi-frequency signals with a frequency-hopping frequency, provided that the width of the analysis frequency channel is slightly different from the signal spectrum width at a separate frequency position.

However, under conditions of a priori uncertainty regarding the frequency band of the received signals, the width of each frequency analysis channel is selected several times narrower than the narrowband signal itself. As a result, multi-frequency signals with an abrupt change in the frequency of a narrow-band signal are detected with a loss of efficiency, and wide-band single-frequency signals with a low spectral power density do not differ from noise. In other words, the prototype method does not have adaptability to the spectral width of the detected signals and loses its effectiveness in the detection and localization of broadband multi-frequency and single-frequency pseudo-random signals with a low power spectral density.

In addition, the disadvantages of the prototype method include the limited localization of signal energy only in the frequency domain of analysis. At the same time, the complex signal transmitter distributes the radiated power in the two-dimensional time-frequency domain.

Thus, due to incomplete matching of the time-frequency region occupied by the power spectrum of the detected signal and the time-frequency region of the formation of correlated AFRs, signal power losses are observed and, as a result, a significant decrease in the distinguishability against the background of noise and interference of complex single-frequency and multi-frequency signals with low power spectral density using three main types of broadband modulation [3, p. 10]:

- modulation by shifting, carrier frequency jumps at discrete time instants by an amount specified by a code sequence;

- modulation of the carrier frequency with a digital code sequence with a symbol repetition rate many times greater than the information signal bandwidth;

- linear frequency modulation of the pulses, as a result of which the carrier frequency changes in a wide frequency band for a time equal to the pulse duration.

Improving the efficiency of detection and localization of complex signals using the prototype method can be achieved in several well-known ways: by increasing the base of the antenna array and by increasing the duration of the signal registration interval to increase the signal-to-noise ratio due to correlation accumulation and signal extraction against the background of noise [4]. However, these paths do not radically solve the problem, since they only partially increase the efficiency of detection and localization of complex signals by improving spatial resolution and increasing the signal-to-noise ratio of only that part of the signal that falls within one transformation time period.

The technical result of the invention is to increase the efficiency of detection and localization in frequency, time, azimuth and elevation of a wider class of complex signals with a low power spectral density under conditions of a priori uncertainty regarding their shape and areas of existence.

The technical result is achieved in that in a method for detecting and localizing complex signals, including receiving and synchronizing digitally the signals received by the antennas of the N-element array, time-shifting conversion of the digital signals of each antenna into complex spectral densities with a given discreteness in time and frequency and storing spectral densities, according to the invention, spatial correlation matrices of received signals (CMPS) are formed from spectral densities in distinguishing their position and size of the time-frequency elements of the receiving area, then they transform the corresponding MPSF to form the amplitude-phase distribution (AFR) of the signals in each time-frequency element, decide on the detection and determine the time-frequency localization regions of each detected signal by simultaneously identifying the generated AFR and matching of the time-frequency regions occupied by the generated AFR and a localized signal form spatial correlation matrices s detected signals (KMOS) from the spectral densities belonging to the identified areas of localization, determine the azimuth and elevation direction of arrival of each detected signal transformation algorithm KMOS of high resolution from the subclass based on an analysis of eigenvalues.

Particular cases of the method are possible:

1. The identification of AFR is carried out by cross-correlation of AFR formed in various time-frequency elements.

This increases the energy efficiency of signal detection and localization.

2. Coordination of the time-frequency regions is carried out by selecting the time-frequency regions providing the maximum signal-to-noise ratio of the mutual correlation response.

It also improves the energy efficiency of signal detection and localization.

3. The formation of AFR signals in each time-frequency element is carried out by selecting an eigenvector corresponding to the maximum eigenvalue of each ICPS.

This increases the signal-to-noise ratio of the generated AFR and increases the efficiency of subsequent detection and localization of signals.

4. The formation of AFR signals in each time-frequency element is also carried out by selecting a separate line of ICPS.

This increases the speed of detection and localization of signals.

5. The determination of the azimuthal and elevation directions of arrival of each detected signal is carried out by selecting the eigenmode of CMOS as the AFR and transforming it using a subclass of high-resolution algorithms based on regularization.

This increases the efficiency of determining the directions of arrival of correlated signals.

5. The azimuthal and elevation directions of arrival of each detected signal are determined by selecting a separate CMOS line as the AFR and transforming it using a subclass of high-resolution algorithms based on regularization.

This increases the speed of determining the directions of arrival of signals.

6. The determination of the azimuthal and elevation directions of arrival of each detected signal is also carried out by comparing the azimuthal and elevation directions of arrival of each detected signal obtained by high-resolution algorithms of various subclasses.

This increases the accuracy and reliability of determining the directions of arrival of each detected signal.

Thus, due to the introduction of operations:

- optimizing the formation of AFR both at the stages of detection and localization, and at the stage of measuring the angles of arrival of signals;

- coordination of the time-frequency region occupied by the power spectrum of the detected signal and the time-frequency region of the formation of correlated AFRs;

- comparing the directions of arrival of signals obtained by high-resolution algorithms of various subclasses, minimizing power losses during detection and direction finding of signals, as well as eliminating anomalous direction finding errors and, as a result, significantly increasing the distinguishability of complex single-frequency and multi-frequency signals with a low power spectral density against noise and interference, it is possible to solve the problem with the achievement of the technical result.

The operation of the method is illustrated by drawings:

Figure 1. Block diagram of a device for detecting and localizing complex signals.

Figure 2. The structure of the correlation matrix of received signals (MPSF), formed in the (q, l) -th frequency-time element of the minimum size (Q = 1, L = 1);

Figure 3. The totality of the MHPS formed in the differing position of the time-frequency elements of the minimum size (Q = 1, L = 1).

Figure 4. The totality of MPSFs formed in time-frequency-differing elements of size (Q = 1, L = 2).

Figure 5. Time-frequency regions occupied by the generated AFR and localized signal.

6. Modules of cross-correlation coefficients of AFR formed in different time-frequency domains:

figa - the area of formation of AFR contains 0,055 signal energy (figa);

figb - region of formation of AFR contains 0.5 signal energy (figb);

figv - the area of formation of AFR contains 0.44 signal energy (figv).

These advantages, as well as the features of the present invention will become clear when considering the operation of the device, which implements the proposed method with reference to the accompanying drawing (figure 1).

The device includes a series-connected antenna system 1, an N-channel frequency converter 2, an N-channel analog-to-digital converter (ADC) 3, a fast Fourier transform computer (FFT) 4, an IFRIC shaper 5, an AFR calculator 6, a detection and localization device 7, shaper CMOS 8, a device for measuring angular coordinates 9 and a display unit 10. The output of the transmitter 4 is also connected to the second input of the shaper CMOS 8.

Antenna system 1 contains N antennas with numbers n = 1 ... N, combined in an array. The antenna array can be of any spatial configuration: flat rectangular, flat annular or three-dimensional, in particular, conformal.

Frequency converter 2 is made in the N-channel version with a common local oscillator and with a channel bandwidth of each channel many times greater than the spectrum width of a single transmitter signal. The common local oscillator provides N-channel coherent signal reception, which is the main condition for recording the relative phase difference of the signals received by the set of antennas. In addition, the transducer 2 provides calibration according to the internal signal source. In this case, a noise generator can be used, the output of which can also be connected instead of all antennas for periodic calibration of channels.

If the capacity and speed of the N-channel ADC are sufficient for direct analog-to-digital conversion of input signals, such as, for example, when constructing a radio image in the KB range, then a frequency selective bandpass filter and amplifier can be used instead of converter 2. In other words, the analog part of the device that implements the proposed method can be built on the principle of direct amplification.

The calculator 4, the shaper 5, the calculator 6, the device 7, the shaper 8, the device 9 are constructed according to a multi-channel scheme, which provides maximum performance due to parallel signal processing.

временных отрезках преобразования.The computer 4 contains N parallel FFT modules, each of which contains a buffer random access memory (BOSU), designed to store spectral densities obtained on

time periods of conversion.

параллельных модулей формирования КМПС.Shaper 5 contains

parallel modules for the formation of the IPMC.

The computer 6 contains Z parallel modules forming AFR.

The device 7 contains Z parallel correlation modules and a comparison unit.

Shaper 8 contains P modules for forming CMOS. Device 9 contains P modules for calculating angular coordinates.

A device is operating that implements a method for detecting and localizing complex signals, as follows.

The multi-frequency time signals x _n (t} from the output of the antenna system 1 from the antennas with numbers n = 1 ... N included in the array arrive at the inputs of the converter 2 in the reception band many times greater than the spectral width of a single transmitter signal, and coherently transferred to a lower frequency.

Using ADC 3, the frequency-converted signals x _n (t) are synchronously converted to digital signals x _n (z), where n is the antenna number, az is the signal time reference number.

сигнала каждой антенны, где q - номер временного отрезка преобразования,

a l - номер частотного отсчета,

In each of the N modules of the calculator 4, a time-shifting conversion of digital signals x _n (z) with a given discreteness in time and frequency, complex spectral densities are obtained and stored

the signal of each antenna, where q is the number of the conversion time span,

al is the number of the frequency reference,

возможно применением гребенки цифровых фильтров или, что, как правило, более эффективно с вычислительной точки зрения, алгоритма БПФ, реализующего дискретное Фурье-преобразование q-го временного отрезка сигнала каждой антенны

где

- оператор прямого дискретного Фурье-преобразования по времени [5].Spectral Density

it is possible to use a comb of digital filters or, which, as a rule, is more efficient from a computational point of view, an FFT algorithm that implements a discrete Fourier transform of the qth time segment of the signal of each antenna

Where

- direct discrete Fourier transform operator in time [5].

To ensure the required detail of the moving spectral density acquisition over time, the qth and (q + 1) th signal conversion segments are selected with the necessary overlap.

N-мерных векторов спектральных плотностей

с элементами

As a result of this operation, a time-frequency matrix is formed and stored.

N-dimensional spectral density vectors

with elements

модулей.The spectral densities obtained in calculator 4 are transmitted to shaper 5, which includes

modules.

формируются пространственные корреляционные матрицы принятых сигналов (КМПС) размером N×N в отличающихся положением (q, l) и размерами (Q, L) частотно-временных элементах области приема, где Q,

- размеры элементов по времени и по частоте соответственно.In the shaper modules 5 of spectral densities

spatial correlation matrices of received signals (CMPS) of size N × N are formed in different position (q, l) and sizes (Q, L) time-frequency elements of the receiving area, where Q,

- sizes of elements in time and frequency, respectively.

матриц для фиксированных значений Q и L в отличающихся положением (q, l) частотно-временных элементах области приема. Эти матрицы могут формироваться последовательно или параллельно. В последнем случае отдельный модуль формирователя 5 должен быть построен по параллельной схеме и содержать ξ вычислителей, что с целью упрощения на фиг.1 не показано.In each of the Z modules of the shaper 5 is formed

matrices for fixed values of Q and L in differing position (q, l) time-frequency elements of the receiving area. These matrices can be formed sequentially or in parallel. In the latter case, a separate shaper module 5 should be constructed in a parallel circuit and contain ξ calculators, which is not shown in FIG. 1 for the sake of simplification.

Moreover, in each module of the shaper 5, each of the ξ matrices is formed according to the following formula:

r - номер частотного элемента размером L,

a

- пространственная корреляционная матрица принятых сигналов для q-го элемента по времени и l-го элемента по частоте единичного размера (Q=L=1) с элементами

1≤n'≤N.where j is the number of the temporary element of size Q,

r is the number of the frequency element of size L,

a

- spatial correlation matrix of received signals for the qth element in time and the lth element in frequency of a unit size (Q = L = 1) with elements

1≤n'≤N.

For greater clarity, figure 2 shows as an example the structure of the MSCF formed by a separate shaper module 5 in the (q, l) -th time-frequency element of size (Q = 1, L = 1).

The set of ICMS formed by a separate module of the shaper 5 in different position (q, l) time-frequency elements of minimum size (Q = 1, L = 1) is shown in Fig. 3, and in time-frequency elements of size {Q = 1, L = 2) - in figure 4.

Thus, in the shaper 5 are simultaneously formed

модулей.Obtained in the shaper 5 KMPS go to the computer 6, which also includes

modules.

In each of the Z modules of the calculator 6, the generated MSCFs are converted to form the amplitude-phase distribution (AFR) of the signals in each time-frequency element.

элементами которой являются АФР

сформированные в отличающихся положением (j, r) частотно-временных элементах размером (Q, L).In each module of calculator 6, for a fixed value of Q and L, a block matrix of size

elements of which are PRA

formed in different position (j, r) time-frequency elements of size (Q, L).

Elements of a block matrix can be formed sequentially or in parallel. In the latter case, a separate calculator module 6 is constructed in a parallel circuit and contains ξ calculators, which, for the sake of simplification, is also not shown in Fig. 1.

При этом вычисляется собственный вектор, соответствующий максимальному собственному значению соответствующей КМПС. Полученный собственный вектор выбирается в качестве АФР.In each module of calculator 6 in all time-frequency elements {j, r) for fixed values of Q and L, the amplitude-phase distributions (AFR) are simultaneously formed by the conversion of the corresponding CMPS

In this case, the eigenvector corresponding to the maximum eigenvalue of the corresponding MPSF is calculated. The resulting eigenvector is selected as the AFR.

The choice of the eigenvector of the ICPS as an AFR ensures the use of the maximum possible information contained in the received signals. This, in turn, reduces the threshold signal-to-noise ratio at subsequent stages of signal processing when they are detected and localized.

представляющая собой N-мерный вектор с элементами в виде усредненных по времени и частоте взаимных спектральных плотностей

В связи с этим для повышения скорости обнаружения и локализации сигналов в вычислителе 6 в качестве АФР выбирается отдельная строка КМПС (см. отдельную строку КМПС при (Q=1, L=1) на фиг.2).A separate line of each N × N-sized ICPS can be considered as a multi-frequency radio hologram

which is an N-dimensional vector with elements in the form of mutual spectral densities averaged over time and frequency

In this regard, to increase the speed of detection and localization of signals in the calculator 6, a separate line of the IPMC is selected as the AFG (see a separate line of the IPMC at (Q = 1, L = 1) in FIG. 2).

блочных матриц

размером

элементами которых являются АФР

сформированные в отличающихся положением (j, r) частотно-временных элементах размером (Q, L). Сформированные АФР

поступают в устройство обнаружения и локализации 7.Thus, as a result of this set of operations in the calculator 6 is obtained

block matrices

the size

elements of which are PRA

formed in different position (j, r) time-frequency elements of size (Q, L). Formed PRA

arrive at the detection and localization device 7.

In the device 7, the generated AFR is identified and the time-frequency regions occupied by the formed AFR and the spectrum of the localized signal are matched. As a result of these operations, decisions are made about the detection of signals and the time-frequency regions of localization of each detected signal are determined.

и

сформированных в различных частотно-временных элементах

и

где

r^'≠r, j^'≠j. Для этого сформированные АФР

нормируются

и получаются

блочных матриц

размером

с элементами в виде нормированных АФР

To do this, in each of the Z blocks of device 7, for fixed values of Q and L, the AFR cross-correlation is performed

and

formed in various time-frequency elements

and

Where

r ^' ≠ r, j ^' ≠ j. For this formed AFR

are normalized

and get

block matrices

the size

with elements in the form of normalized AFR

отличающихся размерами (Q, L) частотно-временных элементов формирования АФР

In each of the Z blocks of device 7, for each of the Z block matrices

different sizes (Q, L) of the time-frequency elements of the formation of AFR

-размерная блочная матрица

(коэффициентов корреляции нормированных АФР.- under construction

-dimensional block matrix

(correlation coefficients of normalized AFR.

определяются скалярным произведением

В свою очередь элементы

матрицы

вычисляются по формуле

которая в развернутой форме имеет вид

Block matrix elements

are defined by a scalar product

In turn the elements

matrices

calculated by the formula

which in expanded form has the form

) с порогом и сигналы спектральных плотностей

, соответствующие элементам, в которых превышен порог, предварительно объединяются в ν_Q,L-й сигнал, где ν_Q,L=1...Π_Q,L, Π_Q,L - число обнаруженных сигналов, отличающихся размерами (Q, L) частотно-временных элементов локализации и угловыми координатами, и формируется бинарная (двоичные числа

- соответствует наличию излучения, а 0 - соответствует отсутствию излучения) частотно-временная матрица его локализации, описывающая закон изменения энергии сигнала по частоте и времени;- modules of correlation coefficients are compared (matrix elements

) with threshold and spectral density signals

corresponding to elements in which the threshold is exceeded are previously combined into the ν _Q, Lth signal, where ν _{Q, L} = 1 ... Π _{Q, L} , Π _{Q, L} is the number of detected signals that differ in size (Q, L ) time-frequency elements of localization and angular coordinates, and binary (binary numbers

- corresponds to the presence of radiation, and 0 - corresponds to the absence of radiation) the time-frequency matrix of its localization, which describes the law of change of the signal energy in frequency and time;

), превысивших заданный порог корреляции. Значение порога корреляции зависит от числа элементов антенной решетки и выбирается из условия минимизации вероятности ложных тревог.- for each detected ν _Q, Lth signal, the average value of the modules of the correlation coefficients (matrix elements

) exceeding a given correlation threshold. The value of the correlation threshold depends on the number of elements of the antenna array and is selected from the condition of minimizing the probability of false alarms.

The average values of the modules of the correlation coefficients that exceeded a predetermined threshold of each detected ν _Q, Lth signal are received in the comparison unit of the device 7.

In the comparison unit of device 7, time-frequency regions are selected that provide the maximum signal-to-noise ratio of the mutual correlation response (the average value of the modules of the correlation coefficients), which is equivalent to matching the time-frequency regions of the formation of identifiable AFR elements and the time-frequency region occupied by the power spectrum of the localized signal .

- соответствует наличию излучения, а 0 - соответствует отсутствию излучения) частотно-временная матрица его локализации, описывающая закон изменения энергии сигнала по частоте и времени. Бинарная частотно-временная матрица формируется таким образом, чтобы выполнялось условие

для тех частотно-временных элементов, которые принадлежат частотно-временным областям локализации i-го сигнала, коэффициент корреляции которых превышает заданный порог, в противном случае

For this, the time-frequency localization regions ν _{Q, L} -x of signals detected at different values of averaging intervals (Q, L) are compared, and signals with overlapping regions are combined into the i-th group, where i = 1 ... P, Р - the number of groups. In each i-th group, a signal with a maximum correlation response is selected, which is identified as the i-th detected signal, where i = 1 ... P, P is the number of detected signals. A binary (binary numbers) is formed for each i-th signal

- corresponds to the presence of radiation, and 0 - corresponds to the absence of radiation) the time-frequency matrix of its localization, which describes the law of change of signal energy in frequency and time. The binary time-frequency matrix is formed in such a way that the condition

for those time-frequency elements that belong to the time-frequency regions of localization of the i-th signal, the correlation coefficient of which exceeds a predetermined threshold, otherwise

Correlation and matching operations increase the energy efficiency of detection and time-frequency localization of complex single-frequency and multi-frequency signals with a low power spectral density.

Figure 5 presents the options for dividing the time-frequency reception region containing the localized signal into time-frequency elements. When the time-frequency elements of the minimum size Q = L = 1, the area of formation of the AFR contains only 0,055 signal energy (figa). When the time-frequency elements with the element size Q = 3 and L = 3 (Fig.5b), the AFR formation region contains 0.5 signal energy, which is an example of the best matching of the time-frequency elements of the formation of the AFR and the time-frequency region occupied by the spectrum localized signal power. Frequency-time elements of size Q = 4 and L = 4 are also inconsistent with the region occupied by the power spectrum of the localized signal, since they contain 0.22, 0.055, 0.44 and 0.11 of the signal energy (Fig. 5c).

On figa presents the modules of the cross-correlation coefficients generated as a result of calculating the cross-correlation of the AFR elements of the 4th row of the time-frequency region of the formation of the AFR shown in figa. Physically, this corresponds to the formation of cross-correlation coefficients of AFR only in the frequency domain for a fixed time interval, the interval at number 4 (figa).

On figb presents the modules of the cross-correlation coefficients generated as a result of calculating the cross-correlation of the AFR elements of the 2nd row for the example presented on figb.

On figv presents the modules of the cross-correlation coefficients generated as a result of calculating the cross-correlation of the AFR elements of the 2nd row for the example shown in figv.

As can be seen from the drawings, the coordination of the time-frequency elements of the formation of AFR and the time-frequency region occupied by the power spectrum of the localized signal is achieved for time-frequency elements of size Q = L = 3 (Fig.5b), which corresponds to the maximum correlation response (Fig. 6b).

- соответствует наличию излучения, а 0 - соответствует отсутствию излучения) частотно-временная матрица локализации сигнала, которая поступает в формирователь пространственных корреляционных матриц обнаруженных сигналов (КМОС) 8.Thus, at this stage, the input signal stream is separated by a spatial attribute. In the analyzed time-frequency domain, all signals differing in angular coordinates are detected. For each detected signal, a binary (binary numbers

- corresponds to the presence of radiation, and 0 - corresponds to the absence of radiation) the time-frequency matrix of signal localization, which enters the shaper of spatial correlation matrices of detected signals (CMOS) 8.

принадлежащие выявленным областям локализации каждого i-го обнаруженного сигнала и поступающие из вычислителя 4, из которых вначале формируются корреляционные матрицы принятых сигналов W(q, l) для частотно-временных элементов единичного размера, принадлежащих выявленным частотно-временным областям локализации обнаруженного сигнала. После этого КМОС вычисляется суммированием полученных матриц W(q, l) по частотно-временным элементам (q, l), для которых выполняется условие

In each of the P modules of the driver 8, the CMOS of each i-th detected signal is calculated. Spectral densities are used for this.

belonging to the identified localization regions of each i-th detected signal and coming from the calculator 4, from which the correlation matrices of the received signals W (q, l) are first formed for time-frequency elements of unit size belonging to the detected frequency-time regions of localization of the detected signal. After that, the CMOS is calculated by summing the obtained matrices W (q, l) over the time-frequency elements (q, l) for which the condition

The calculated CMOS of each i-th signal goes to device 9.

In each of the P modules of device 9, the azimuthal and elevation directions of arrival of the i-th detected signal are determined. For this, the CMOS of the ith signal is converted according to one of the known high-resolution algorithms related to a subclass of algorithms based on the analysis of eigenvalues. A subclass of high-resolution radio image generation algorithms based on the analysis of eigenvalues of the spatial correlation matrix includes, for example, MUSIC (multiple signal classification) and EV (eigenvector) algorithms [4].

As a result of the CMOS conversion of the i-th signal, a radio image of the two-dimensional angular spectrum of the signal is formed, the maximum values of which determine the elevation and azimuth directions of its arrival.

To increase the efficiency of determining the directions of arrival of correlated signals in each of the P modules of device 9, the eigenmode of the CMOS is selected as the AFR, according to which using the subclass of high-resolution algorithms based on regularization [6], the angles of arrival of the detected signal are determined.

To increase the speed of determining the directions of arrival of signals in each of the P modules of device 9, a separate CMOS line is selected as the AFS, according to which, using a subclass of high-resolution algorithms based on regularization, the directions of arrival of each detected signal are determined.

In addition, in each of the P modules of device 9, to increase the accuracy and reliability of determining the directions of arrival of each detected signal, the azimuthal and elevation directions of arrival of each detected signal are obtained, obtained by high-resolution algorithms of various subclasses. Signal arrival directions coinciding with a given tolerance are averaged and used as reliable directions.

In order to increase the information content, in the display unit 10, the parameters of the detected transmitters are displayed using a geographical map of the area, including the time-frequency region occupied by the detected signal, as well as the azimuth and elevation angle of the transmitter.

Thus, the performance of the above actions on the signals provides a significant increase in the efficiency of detection and localization in frequency, time, azimuth and elevation of a wider class of complex signals under conditions when the shape and areas of existence of the controlled signals are a priori unknown. Improving efficiency is achieved by:

- optimizing the formation of AFR both at the stages of detection and localization, and at the stage of measuring the angles of arrival of signals;

- coordination of the time-frequency region occupied by the power spectrum of the detected signal and the time-frequency region of the formation of correlated AFRs;

- comparing the directions of arrival of signals obtained by high-resolution algorithms of various subclasses, minimizing power losses during detection and direction finding of signals, as well as eliminating anomalous direction finding errors and, as a result, significantly increasing the distinguishability of complex single-frequency and multi-frequency signals with a low spectral power density against the background of noise and interference.

где

- ширина спектра обнаруживаемого сигнала, δ_ƒ- ширина дискреты анализа по частоте. При F_с=50 кГц и

получаем В=15,80 раз (11,98 дБ).So, for example, the energy gain in comparison with the prototype, achieved only through the coordination of the frequency regions during the detection and localization of the signal, reaches a value

Where

is the width of the spectrum of the detected signal, δ _ƒ is the width of the frequency analysis samples. At F _c = 50 kHz and

we get B = 15.80 times (11.98 dB).

Information sources

1. US, patent, 5955993, CL. G 01 S 3/02, 1999

2. RU, patent, 2190236, cl. G 01 S 5/04, 2002

3. Dickson R.K. Broadband systems. - M.: Communication, 1979.

4. Johnson D.H. Application of spectral estimation methods to problems of determining the angular coordinates of radiation sources // TIIER. - 1982. - T. 70. No. 9. - S.126.

5. Marpl.-ml. S.L. Digital spectral analysis and its applications. - M .: Mir, 1990 .-- 584 p.

6. Shevchenko V.N. Estimation of the angular position of sources of coherent signals based on regularization methods // Radio Engineering. - 2003. - No. 9. - C.3-10.

Claims (5)

1. A method for detecting and localizing complex signals, including receiving and synchronizing digitally the signals received by the antennas of the N-element array, time-shifting conversion of the digital signals of each antenna into complex spectral densities with a given discreteness in time and frequency and storing spectral densities, characterized in that from the spectral densities form spatial correlation matrices of received signals (MPSF) in different position and size of the frequency-time nth elements of the receiving region, then the corresponding MPSF is converted to form the amplitude-phase distribution (AFR) of the signals in each time-frequency element, a decision is made about the detection and the time-frequency regions of localization of each detected signal are determined by simultaneously identifying the generated AFR and matching the time-frequency The areas occupied by the generated AFR and the localized signal form spatial correlation matrices of the detected signals from the spectral planes. At the same time, based on the analysis of the eigenvalues of the spatial correlation matrix of each detected signal, a radio image of its two-dimensional angular spectrum is formed, from the maxima of which determine the elevation and azimuth directions of its arrival. 2. The method according to claim 1, characterized in that the identification of AFR is carried out by cross-correlation of AFR formed in various time-frequency elements. 3. The method according to claim 1, characterized in that the coordination of the time-frequency regions is carried out by selecting the time-frequency regions, providing the maximum signal-to-noise ratio of the mutual correlation response. 4. The method according to claim 1, characterized in that the formation of AFR signals in each time-frequency element is carried out by selecting an eigenvector corresponding to the maximum eigenvalue of each ICPS. 5. The method according to claim 1, characterized in that the formation of AFR signals in each time-frequency element is also carried out by selecting a separate line of the IPMC.

Images