RU2528391C1 - Method of searching for low-signature mobile objects - Google Patents

Abstract

FIELD: radio engineering, communication.

SUBSTANCE: method employs new operations which maximise the output signal-to-noise ratio and which are based on finding the highest eigen values of correlation matrices used to compensate for coherent interference in the form of a forward signal of an illumination transmitter, and during detection of useful signals obtained after interference compensation and corrected on a given set of hypothetical spatial coordinates, directions and velocities of objects.

EFFECT: more efficient search for low-signature mobile objects.

3 dwg

Description

The invention relates to radio engineering and can be used in monitoring systems of land, sea and air space using direct and scattered objects of radio signals emitted by many uncontrolled and monitored transmitters of electronic systems for various purposes.

Achieving high efficiency in the detection, localization and identification of land, sea and air objects is limited by significant a priori uncertainty in the size, reflecting properties and states (spatial coordinates, direction and speed of movement) of the object, as well as the imperfection of known methods for detecting and tracking moving objects.

The technology of covert detection and tracking of moving objects, using natural radio illumination of targets, created at a variety of frequencies by radio emissions from transmitters for various purposes in the short, meter, decimeter and centimeter wavelengths: broadcast (commercial FM broadcasting, high-definition television), information (communication) and measuring (control, navigation), has not yet received sufficient distribution, despite the fact that it can significantly increase the secrecy and effectiveness of detection, spatial localization and identification of a wide class of moving objects.

There is a method of searching for subtle moving objects [1], which consists in selecting a transmitter emitting a spread spectrum radio signal, synchronously receiving an array of multipath radio signals including the direct radio signal of the transmitter and the radio signals of this transmitter scattered from objects, synchronously converting the ensemble of radio signals received by the antennas into digital signals, direct and compressed scattered signals are formed from digital signals, the selected direct and scattered signals are compared and time All delays, Doppler shifts and directions of arrival of scattered signals, based on time delays, Doppler shifts and directions of arrival, perform detection and spatial localization of airborne objects.

This method does not contain operations to suppress coherent interference in the form of a direct radio signal from the transmitter and, as a result, provides effective detection of only very large closely spaced objects.

More effective is the method of searching for subtle moving objects [2], free from this drawback and selected as a prototype.

According to this method, direct and scattered by mobile objects broadband radio signals emitted by transmitters of various electronic systems are coherently received by the antenna array multi-beam radio signals including direct and scattered radio signals, synchronously transform the ensemble of radio signals received by the antennas into digital signals, digital signals are combined into a matrix digital signal and remember, from the matrix digital signal form the signal of the spatial correlation matrix, to which, together with the guidance vector signal, determined by the azimuthal elevation direction of direct radio signal reception, wavelength, and lattice geometry, is converted into an optimal weight vector signal, the matrix digital signal is converted into a direct digital signal, which is stored in a complex and mutually dependent time shift correlation functions (VKF) between the digital signal of an individual antenna and a direct digital signal, determine the maximum value of the module of each complex VKF and phi they compute the values of the complex VKF corresponding to these maxima, calculate the differential digital signals, form complex two-dimensional cross-correlation functions (DKKF) depending on the time and frequency shifts between each difference digital signal and the direct digital signal, average the modules of the complex DKFV, determine the number of compressed from the maxima of the averaged DKFV signals and record the values of time delay and absolute Doppler shift of each p-th compressed signal,

the components of complex DKFV corresponding to a separate maximum of the averaged DKFV are identified as the pth signal compressed in time and frequency, the values of the components of complex DKFV, the time delay and the absolute Doppler shift of each pth compressed signal are extracted and stored according to the selected values of the rth identified components complex DKFV synthesize a complex two-dimensional angular spectrum, the maximum modulus of which determines the azimuth-elevation direction of arrival of the p-th compressed signal, by pits of delay and absolute Doppler shift and azimuthal elevation direction of arrival detect and determine the spatial coordinates of moving objects.

The prototype method due to the presence of adaptive spatial filtering operations and operations of electronic compensation of coherent interference in the form of a powerful direct radio signal of the backlight transmitter provides detection of a wider class of moving objects. However, the prototype method contains operations for generating time-frequency images of target signals, based on the formation of a classical two-dimensional mutual correlation function, which contains high side lobes masking the signals of distant and weakly scattering objects, on the one hand, and perceived as false targets, on the other.

Thus, the disadvantage of the prototype method is the low search efficiency (high probability of missing targets and false alarms) of subtle moving objects.

The technical result of the invention is to increase the search efficiency of subtle moving objects.

Improving the search efficiency of subtle moving objects is achieved through the use of new operations that maximize the output signal-to-noise ratio and are based on finding the largest eigenvalues of the correlation matrices used in the formation and compensation of the direct signal of the backlight transmitter, which is coherent interference, as well as in the selection and optimal coherent detection of useful signals received after interference compensation and corrected for a given set of hypotheticals physical states (spatial coordinates, direction and speed of movement) of objects.

- номер антенны, объединяют откорректированные цифровые сигналы S_n в матрицу откорректированных сигналов Φ, которую запоминают и преобразуют в пространственную корреляционную матрицу откорректированных сигналов G=Φ^HΦ, находят наибольшее собственное значение λ_max корреляционной матрицы откорректированных сигналов G, а также итерационно формируют зависящий от предыдущего решения вспомогательный матричный сигнал Λ^(k)=Λ^(k-1)(G/λ_max), где k=1, 2, … - номер итерации, Λ⁽⁰⁾=I - единичная матрица, до тех пор, пока энергия разности текущего и запомненного предыдущего частотно-временных изображений не достигнет заданного малого значения, после чего фиксируют текущий номер итерации k_t и формируют прямой цифровой сигнал

, где E - вектор столбец из N единиц, который запоминают, формируют нормированный коэффициент взаимной корреляции

между прямым цифровым сигналом u и откорректированным цифровым сигналом s_n каждой антенны, получают разностный цифровой сигнал каждой антенны

, преобразуют разностный цифровой сигнал каждой антенны

в откорректированный по временной задержке и доплеровскому сдвигу частоты на заданном множестве гипотетических состояний объектов разностный цифровой сигнал

, откорректированные разностные цифровые сигналы

всех антенн объединяют в матрицу разностных сигналов

, которую преобразуют в пространственную корреляционную матрицу разностных сигналов

, находят наибольшее собственное значение корреляционной матрицы разностных сигналов

, после сравнения которого с порогом принимают решение об обнаружении подвижного объекта с текущим гипотетическим состоянием объекта.The technical result is achieved by the fact that in the method of searching for inconspicuous moving objects, which consists in the use of direct and scattered by moving objects broadband radio signals emitted by transmitters of electronic systems for various purposes, coherent antenna array receive multi-beam radio signals, including direct and scattered radio signals, synchronously convert the received antennas, radio signals into digital signals, according to the invention, convert the digital signal of each antenna into an corrector vanny of time delay and Doppler frequency shift for the known state of the transmitter digital signal s _n, where

- the antenna number, combine the corrected digital signals S _n into the matrix of corrected signals Φ, which is stored and transformed into the spatial correlation matrix of the corrected signals G = Φ ^H Φ, find the largest eigenvalue λ _{max of the} correlation matrix of the corrected signals G, and iteratively form depending on of the previous solution, the auxiliary matrix signal Λ ^(k) = Λ ^(k-1) (G / λ _max ), where k = 1, 2, ... is the iteration number, Λ ⁽⁰⁾ = I is the identity matrix, until difference energy is current and undoubtedly of the previous time-frequency images reaches a predetermined low value, and then fix the current number of iterations k _t and form a direct digital signal

where E is a column vector of N units to be memorized, a normalized cross-correlation coefficient is formed

between the direct digital signal u and the corrected digital signal s _{n of} each antenna, a differential digital signal of each antenna is obtained

convert the differential digital signal of each antenna

in the time difference and Doppler frequency shift corrected for a given set of hypothetical states of objects, a difference digital signal

, corrected differential digital signals

all antennas are combined into a matrix of difference signals

, which is converted into a spatial correlation matrix of difference signals

find the largest eigenvalue of the correlation matrix of difference signals

, after comparing it with a threshold, they decide to detect a moving object with the current hypothetical state of the object.

The operation of the method is illustrated by drawings:

figure 1 is a structural diagram of a device for searching subtle moving objects;

figure 2 - simulation results of the search process for invisible moving objects of the proposed method;

figure 3 - simulation results of the search process for inconspicuous moving objects when using the prototype method.

A device that implements the proposed method for searching for inconspicuous moving objects includes a series-connected antenna system 1, an N-channel frequency converter (NRF) 2, an N-channel quadrature sampling device 3, a calculator 4, and a display device 5. In turn, the calculator 4 includes a shift device 4-1, a shaper of differential signals 4-2, a shift device 4-3, a detection device 4-4. The shear device 4-1 and the frequency converter 2 are connected to external systems to obtain information about the radiation parameters of the selected backlight transmitter. In addition, the shift device 4-1 receives from external systems information about the state of the backlight transmitter in the form of: spatial coordinates, direction and speed of movement when choosing a mobile transmitter or only spatial coordinates when choosing a stationary transmitter. Communication with external systems for the purpose of simplification is not shown.

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.

The bandwidth of each channel of the multi-channel RFI 2 provides reception of a broadband signal. In addition, multi-channel RFR 2 and device 3 are made with a common local oscillator, which provides coherent reception of radio signals. For periodic calibration of channels using an external signal source in order to eliminate their amplitude-phase non-identity, the RFI 2 provides the connection of one of the antennas, instead of all the antennas of the array. Calibration by internal signal source is possible. 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 resolution and speed of the ADCs that are part of device 3 are sufficient for direct analog-to-digital conversion of the input signals, then instead of the RFI 2, a frequency-selective band-pass filter and amplifier can be used. In other words, the analog part of the device that implements the proposed method can be built on the principle of direct amplification.

A device that implements the proposed method works as follows.

After receiving information about the radiation parameters of the selected backlight transmitter, the RFI 2 tunes to the frequency of the backlight radio signal.

Multipath radio signals, which include a spread spectrum direct signal emitted by the backlight transmitter and scattered objects from this transmitter, are coherently received by spatially separated receiving antennas of array 1. The total radio signal s _n (t) received by each antenna array 1 is time-dependent t is transferred coherently to RF in 2 lower frequency.

. Комплексные цифровые сигналы $$s_n^{\text{'}}=\left[s_n^{\text{'}}\left(1\right),\dots ,s_n^{\text{'}}\left(z\right),\dots s_n^{\text{'}}\left(Z\right)\right]$$

, где Z - число цифровых отсчетов сигнала, поступают в устройство 4-1, где синхронно регистрируются на заданном временном интервале.The ensemble of received radio signals s _n (t) formed in RFH 2 is synchronously converted in device 3 into an ensemble of complex digital signals $$s_n^{\text{'}\text{'}}\left(z\right)$$

. Integrated Digital Signals $$s_n^{\text{'}\text{'}}=\left[s_n^{\text{'}\text{'}}\left(\mathrm{one}\right),...,s_n^{\text{'}\text{'}}\left(z\right),...s_n^{\text{'}\text{'}}\left(Z\right)\right]$$

where Z is the number of digital samples of the signal, enter the device 4-1, where they are synchronously recorded on a given time interval.

In addition, the following actions are performed in device 4-1: - the digital signal of each antenna is converted to a digital signal s _n, corrected for the time delay and Doppler frequency shift for a known transmitter state.

In this case, information on the state of the transmitter coming from external systems is taken into account in the form of current spatial coordinates, the direction and speed of its movement, if a mobile transmitter is selected, and in the form of only spatial coordinates, if a stationary transmitter is selected.

The conversion is carried out in a known manner [3].

, где z - номер временного отсчета сигнала, µ_n - значение комплексного коэффициента рассеяния цели, u(z) - копия сигнала передатчика, принятого каждой n-ой антенной сигнала корректируется по следующей формуле $$s_n\left(z\right)=s_n^{\text{'}}\left(z+{\tau }_n\right)e^{-i2\pi l_{n}z}={\mu }_{n}u\left(z\right)$$

. Отметим, если передатчик является стационарным, то значение доплеровского сдвига частоты l_n его сигнала равно нулю и принятый каждой n-ой антенной сигнал корректируется только по задержке;So, the expected value of the time delay τ _n is calculated from the spatial coordinates of the transmitter, and the expected value of the Doppler frequency shift l _{n of the} signal received by each nth antenna is calculated from the direction and speed of the transmitter. After that, each countdown $$s_n^{\text{'}\text{'}}\left(z\right)={\mu }_{n}u\left(z-{\tau }_n\right){\mathrm{<figure-callout\; id="2"\; label="e\; i"\; filenames="00000040.png"\; state="\{\{state\}\}">e\; </figure-callout>}}^i$$

where z is the number of the time reference signal, μ _n is the value of the complex scattering coefficient of the target, u (z) is a copy of the transmitter signal received by each n-th antenna signal is adjusted according to the following formula $$s_n\left(z\right)=s_n^{\text{'}\text{'}}\left(z+{\tau }_n\right)e^{-\mathrm{<figure-callout\; id="2"\; label="i"\; filenames="00000040.png"\; state="\{\{state\}\}">i</figure-callout>}2\pi l_{n}z}={\mu }_{n}u\left(z\right)$$

. Note that if the transmitter is stationary, then the value of the Doppler frequency shift _{ln of} its signal is zero and the signal received by each nth antenna is corrected only for the delay;

- the digital signals s _n corrected for the known spatial position of the transmitter are combined into the matrix of corrected signals C = (s ₁ , ..., s _N ) of size Z × N. The matrix of corrected signals C enters the shaper 4-2.

In the shaper 4-2, the following actions are performed:

- the matrix of corrected signals C is stored and converted into N × N spatial — the correlation matrix of corrected signals G = C ^H C;

- find the largest eigenvalue λ _{max of the} correlation matrix of the corrected signals G.

The greatest eigenvalue of the matrix G is found by known methods [4, 5], for example, as the largest root λ _{max of the} polynomial det (G-λI), where det is the determinant, I is the identity matrix.

- an auxiliary matrix signal Λ ^(k) = Λ ^(k-1) (G / λ _max ) depending on the previous solution is iteratively generated, where k = 1, 2, ... is the iteration number, Λ ⁽⁰⁾ = I is the identity matrix, until the energy of the difference between the current and stored previous frequency-time images reaches a predetermined small value;

, где E - вектор столбец из N единиц.- after which the current iteration number k _t is fixed and a direct digital signal is generated

where E is a column vector of N units.

When the auxiliary matrix signal is generated at the first iteration, the signal G / λ _max is generated, and then the first approximation of the time-frequency image signal is found and stored, i.e., Λ ⁽¹⁾ = Λ ⁽⁰⁾ (G / λ _max ) = I (G / λ _max ), since Λ ^(k-1) = Λ ⁽⁰⁾ = I for k = 1, at the second iteration use the approximation obtained at the first iteration, that is, Λ ⁽²⁾ = Λ ⁽¹⁾ (G / λ _max ) for k = 2, etc.

Thus, at the first (k = 1) and subsequent (k = 2, 3, ...) iterations, the auxiliary matrix signal Λ ^(k) is expressed in terms of the auxiliary matrix signal Λ ^(k-1) obtained at the previous iteration.

, где

- евклидова норма вектора, полученных на текущей и предыдущей итерации, с порогом δ. Значение порога выбирается, например, из условия

.At each iteration, the energy of the difference in the time-frequency images is compared

where

is the Euclidean norm of the vector obtained at the current and previous iteration, with a threshold δ. The threshold value is selected, for example, from the condition

.

инициализируется очередная итерация синтеза частотно-временного изображения, на которой номер итерации k увеличивают на единицу, и повторяются операции формирования сигнала Λ^(k)=Λ^(k-1)(G/λ_max), его запоминания и сравнения энергии разности частотно-временных изображений

с порогом δ. При выполнении условия

фиксируется текущий номер итерации k_t, формируется и запоминается прямой цифровой сигнал u.If the condition is not met

the next iteration of the synthesis of the time-frequency image is initialized, on which the iteration number k is increased by one, and the operations of generating the signal Λ ^(k) = Λ ^(k-1) (G / λ _max ), storing it, and comparing the energy of the time-frequency difference difference are repeated images

with threshold δ. When the condition is met

the current iteration number k _t is fixed, a direct digital signal u is generated and stored.

In addition, the following operations are performed in the shaper 4-2:

между прямым цифровым сигналом u и откорректированным цифровым сигналом s_n каждой антенны;- a normalized cross-correlation coefficient is formed

between the direct digital signal u and the corrected digital signal s _{n of} each antenna;

.- a differential digital signal of each antenna is obtained

.

поступает в устройство 4-3.Received in the shaper 4-2 differential digital signal of each antenna

enters the device 4-3.

преобразуется в откорректированный по временной задержке τ_n и доплеровскому сдвигу частоты ω_dn на заданном множестве гипотетических состояний объектов разностный цифровой сигнал

.The device 4-3 differential digital signal of each antenna

a differential digital signal is converted into a difference digital signal corrected by the time delay τ _n and the Doppler frequency shift ω _dn on a given set of hypothetical states of objects

.

в откорректированный разностный цифровой сигнал

также осуществляется известным способом [3]. При этом отдельное гипотетическое состояние объекта описывается несколькими величинами: ожидаемыми пространственными координатами, направлением и скоростью движения подвижного объекта.Converting a difference digital signal of each antenna

into the corrected difference digital signal

also carried out in a known manner [3]. In this case, a separate hypothetical state of the object is described by several values: the expected spatial coordinates, direction and speed of the moving object.

всех антенн объединяются в Z×N матрицу разностных сигналов $$\widetilde{Ц}=\left({\tilde{s}}_1,\dots ,{\tilde{s}}_N\right)$$

. Матрица разностных сигналов $$\widetilde{Ц}$$

преобразуется в N×N пространственную корреляционную матрицу разностных сигналов $$\tilde{G}={\widetilde{Ц}}^H\widetilde{Ц}$$

.After that, the device 4-3 corrected differential digital signals $${\tilde{s}}_n$$

all antennas are combined into a Z × N matrix of difference signals $$\widetilde{\mathrm{Ts}}=\left({\tilde{s}}_{\mathrm{one}},...,{\tilde{s}}_N\right)$$

. Difference Signal Matrix $$\widetilde{\mathrm{Ts}}$$

the spatial correlation matrix of difference signals is converted to N × N $$\tilde{G}={\widetilde{\mathrm{Ts}}}^H\widetilde{\mathrm{Ts}}$$

.

поступает в устройство 4-4.The spatial correlation matrix of difference signals obtained in device 4-3 $$\tilde{G}$$

enters the device 4-4.

известными способами [4, 5].The device 4-4 is the largest eigenvalue of the spatial correlation matrix of difference signals $$\tilde{G}$$

by known methods [4, 5].

сравнивается с порогом. Порог выбирается исходя из минимизации вероятности ложной тревоги. При превышении порога принимается решение об обнаружении подвижного объекта с текущим гипотетическим состоянием объекта, то есть с текущими гипотетическими пространственными координатами, направлением и скоростью движения.Found the largest eigenvalue of the matrix $$\tilde{G}$$

compares with the threshold. The threshold is selected based on minimizing the likelihood of a false alarm. If the threshold is exceeded, a decision is made to detect a moving object with the current hypothetical state of the object, that is, with the current hypothetical spatial coordinates, direction and speed of movement.

To assess the comparative effectiveness of the device that implements the proposed method, a simulation on a PC.

The simulation was carried out using semi-natural data, based on the measured signal of the sound accompaniment of analog television at a frequency of 229.75 MHz, which played the role of a target illumination signal.

Figure 2 and figure 3 presents the time-frequency image scattered by the objects of the radio signals generated by the proposed method and the prototype method, respectively. Both compared methods provide detection of an object signal at a point with coordinates (11000 m, 200 m). However, a lot of false targets are observed in the image obtained by the prototype method.

It follows that the proposed method provides the formation of a better radio image. This reduces the likelihood of false alarms and the likelihood of missing targets and, as a result, increases the efficiency of searching for subtle moving objects.

Thus, through the use of new operations that maximize the output signal-to-noise ratio and are based on finding the largest eigenvalues of the correlation matrices used in the formation and compensation of the direct signal of the backlight transmitter, which is coherent interference, as well as in the selection and optimal coherent detection of useful signals, obtained after interference compensation and corrected on a given set of hypothetical spatial coordinates, directions and velocities zheniya objects, it is possible to solve the problem with achieving said technical result.

Information sources

1. US patent 6703968 B2, cl. G01S 13/87, 2004

2. RU, patent, 2 444 754, cl. G01S 13/02, 2012

3. Reference radar. Ed. M. Skolnik. New York, 1970. Transl. from English (in four volumes) under the general ed. K.N. Trofimova. Volume 1. Basics of radar. Ed. Ya.S. Yitzhoki. M .: “Owls. Radio ”, 1976, 456 p.

4. Wilkison J. X. Algebraic eigenvalue problem. - M.: Nauka, 1970.564 s.

5. Marple S.L. (ml.). Digital spectral analysis and its applications. M .: "World", 1990.584 s.

Claims (1)

A method of searching for inconspicuous moving objects, which consists in the use of direct and scattered by moving objects broadband radio signals emitted by transmitters of various electronic systems, coherently receiving a multi-beam radio signal including a direct and scattered radio signals by the antenna array synchronously converts the received radio signals into antennas into digital signals by converting the digital signal of each antenna into a time delay and Doppler-corrected one Moving the frequency for the known state of the transmitter digital signal s _n, where

- the antenna number, combine the corrected digital signals s _n into the matrix of corrected signals Φ, which is stored and transformed into the spatial correlation matrix of the corrected signals G = Φ ^H Φ, find the largest eigenvalue λ _{max of the} correlation matrix of the corrected signals G, and iteratively form depending on of the previous solution, the auxiliary matrix signal Λ ^(k) = Λ ^(k-1) (G / λ _max ), where k = 1, 2, ... is the iteration number, Λ ⁽⁰⁾ = I is the identity matrix, until difference energy is current and undoubtedly of the previous time-frequency images reaches a predetermined low value, and then fix the current number of iterations k _t and form a direct digital signal

where E is a column vector of N units to be memorized, a normalized cross-correlation coefficient is formed

between the direct digital signal u and the corrected digital signal s _{n of} each antenna, a differential digital signal of each antenna is obtained

convert the differential digital signal of each antenna

in the time difference and Doppler frequency shift corrected for a given set of hypothetical states of objects, a difference digital signal

, corrected differential digital signals

all antennas are combined into a matrix of difference signals

, which is converted into a spatial correlation matrix of difference signals

find the largest eigenvalue of the correlation matrix of difference signals

, after comparing it with a threshold, a decision is made to detect a moving object with the current hypothetical state of the object.

Images