Method for determining coordinates of moving radiation source using passive sonar

Abstract

FIELD: source's radiation.SUBSTANCE: hydroacoustics in the near field of a source's radiation, specifically passive sonar. In the method of passively determining the coordinates of a moving radiation source, a signal is received in combination with interference by a wide-aperture receiving system (RS) of M antennas (M≥3) in the Fresnel zone, preliminary processing is carried out and indication within a given sector of directions a and an interval of ranges D of the observation field is provided, then signal marks (SM) are observed, a selection of SMs for automatic tracking S with coordinates <αiS,DiS> is chosen, where i is the number of the next FFT cycle. The proposed method includes the assignment of a virtual reference point (VRP) Rrefi, placed in coordinates <αiS,riS>, which coincide in direction, but differ in range from the actual location of the radiation source S, the transformation of the input sample of signals from the outputs of the antennas of the RS by multiplying it by the table of the input M-dimensional sample HXiSR, obtained as a result of multiplying two phase vectors and the “matrix inversion” transformation. The phase vectors are obtained from two sets of signal propagation times to the antenna centers Am (m=1,…, M): [tSmrefi] - from the expected point Srefi and [tmrefiR] - from the virtual point Rrefi. At the next FFT cycle, the signal arrays of the input sample Xi from the antenna adders are multiplied by HXiSR, obtaining the transformed signal samples [XmiSR]. To determine the direction αsand ranges Ds deviations δαi and δDi are determined by measuring the parameters of the discriminatory characteristic (DC): <c1α, c1R> and <c2α, c2D>, then the differences in the signal arrival times from the VRP to the centers of each mn-th pair of antennas τRmni(arefi,rrefi)=tRmi-tRni, as well as weighting coefficients wRαmniand wRDmni are calculated, taking into account the difference in the geometry of the relative position of the RS antennas and the radiation source. Deviations of direction δαi and rangeδDi are calculated as δαi=c1ai / c2ai and δDi=c1Di / c2Di. Estimates ai*, Di* are obtained by adjusting αкуаi, Drefi by the values of δαi, δDi respectively, and the current coordinates of the radiation source as(ti), Ds(ti) at the time ti, and the forecast of the reference values αrefi+1, Drefi+1 for the next (i+1)-th step are determined by the known methods of dynamic smoothing.EFFECT: ensuring an increase in the accuracy of determining the range by focusing a hydroacoustic signal, received from the actual location of the radiation source S, by a passive sonar at a virtual point R on a nonlinear range scale DR by precisely phasing it at a virtual point R taking into account the difference in the phase vectors of the signal received by the RS antennas from these points.1 cl, 3 dwg

Description

[0001] The invention relates to the field of hydroacoustics in the near field of a source's radiation, specifically to passive sonar (PHL), and can also be used in atmospheric acoustics and passive radar.

[0002] For passive location in the near field of the source (Fresnel zone), a PGL with extended (wide aperture) receiving systems (RS) from M hydroacoustic receivers (M≥3) or from spaced directional local antennas (hereinafter simply “RS antennas”) and means for preliminary and inter-antenna spatio-temporal processing of information coming from the outputs of the adders of the M RS antennas in the form of a sequence of samples in the frequency domain of fast Fourier transform (FFT) samples, hereinafter referred to as the “input sample” are used [X S m, (s+n)i (ƒ k )], i is the number of the next FFT array, S is the location of the hydroacoustic emission source (HES), (s+n) is the signal plus interference, ƒ k- the frequency of the k-th FFT element. To determine the coordinates of the IGI, two-coordinate survey indicators with a coordinate grid <direction / range> in a given observation field in a given direction sector are used: α∈[α min ÷α max ] and the range interval: D∈[D min ÷D) max ] [Gamper L.E. Optimal spatio-temporal processing in passive sonar systems with spaced antennas / / Scientific and technical collection “Hydroacoustics”. Issue 10 St. Petersburg, 2009. Pp. 49-60.]. Inter-antenna spatio-temporal processing ensures observation of the “useful” signal of each radiation source on the PGL indicator in the form of a “signal mark” (SM) by focusing the PS at the point of the supposed location of the IGI [RU Patent No. 2507531, IPC G01S 3 / 80, No. 2012147666 / 07, declared 08.11.2012, published 20.02.2014] and reading the sought estimates of the IGI coordinates <α* s ,D* s>by the position of the maximum of the SO on the coordinate scales of the KS ("maximum" reading, the cross-section of the maximum of the SO is an even function). Along with the means of survey and preliminary localization of the location of the IG, in passive sonar the task of accurately determining the coordinates and automatic tracking (AT) of a moving IG and / or a PGL carrier is relevant. In noise direction finding (NDF), discrimination technology is used for this - measuring the sign and magnitude of the deviation of the location of the IG from the reference point (RP) ("zero" reading, the output characteristic of the discriminator (OC) is an odd function) [Loskutova G.V., Polkanov K.I. Spatial-frequency and frequency-wave methods for describing and processing hydroacoustic fields. St. Petersburg: "Science". 2007], which, at the beginning of the AS, uses the position of the maximum of the radar on the WP overview indicator, and then the results of current measurements and the forecast of the radar for the next step of the AS using known methods of dynamic accumulation [Kuzmin S.Z. Digital processing of radar information. Moscow: Sov. radio.

[1967] In the near zone, discrimination of the direction deviation δ is used. α and range δ D in parallel, using the characteristics of the discriminators in the direction of DX α and by the range of the DH D , with different parameters of operations that take into account the features of the geometric model for each of these tasks [Gamper L.E. Criteria for assessing the range of passive sonar / / NTS Hydroacoustics. 2010. Issue 11(1). P. 50], [Patent of the Russian Federation No. 2623831, IPC G01S 3 / 80, application No. 206135357, 08 / 30 / 2016, published 06 / 29 / 2017]. A problematic feature of the PGL in the near zone is the asymmetric increase in the width of the response of the PS in its range section depending on the value of the measured range D, or more precisely on the ratio B ПС / D, where B ПС - the basic size of the PS. At the same time, the error in determining the range deviation δ increases D from OT due to an asymmetric decrease in the steepness of the discrimination characteristic (DC) D) [Gamper L.E., Criteria for estimating the range of passive sonar / / NTS Gidroakustika. 2010. Issue 11(1). P. 50]. Such a deformation of the PS response in the range of distances of the IGS from the PS in the near zone is due to the geometry of the system <<IGS / circular front of the IGS signal / PS antenna>> [Le Cadre JP. Performance analysis of wavefront curvature methods for range estimation of a moving source / / IEEE transactions on aerospace and electronic systems. - 1995. - Vol. 31. - No. 3. - Pp. 1082-1103], which differs significantly from the geometry of the far zone of the IGS field. For brevity, we call this dependence the range estimation error σ D [D*] "geometric." This is the reason for its multiple increase at large distances within the near zone of the IGI, significantly reducing the range of feasibility of the PGL, already defined by the boundary of the near zone, which is the subject of the proposal under consideration. Accordingly, we will call the dependence σ "energetic." D[D*], caused by known changes in the levels of the IGI signals and interference, both in the near and far zones, under specific conditions of sound propagation in the environment [Konson A.D. Acoustic signal in the ocean / / Information hydroacoustics.: St. Petersburg, Publishing House of St. Petersburg ETU "LETI". 2023. pp. 7 - 18], which are not related to the content of this application. In the well-known expression for σ D[D*] For maximum likelihood estimates, these two dependencies are the product of two factors. Therefore, reducing and stabilizing the "geometric" factor is essential for ensuring the effective coverage of the PGL throughout the entire observation interval, right up to the far boundary of the near zone. The "energy" factor has no specificity for the near or far zone and is not relevant to the problem under consideration here; therefore, it can have any value depending on the specific adopted values ​​of the model parameters. Therefore, without losing the essence of the operations discussed and the clarity of the illustrations, the following materials, examples, and illustrations adopt the artificial condition of constant signal levels at the PS antennas for all values ​​of the IGI coordinates.

[0003] To overcome the described deformation of the CO and DH DSometimes, well-known methods of changing the graphic image according to the "distorting mirror" principle are used by changing the linear scale of one or another axis of the graph according to a given nonlinear rule [P.A. Bakut et al. Questions of the statistical theory of radar. Vol. II, p. 861]. In some cases, this allows for symmetry and constancy of the shape of the graphic object within the scale, but cannot improve the resolution of objects and the accuracy of range determination, since the information field of the picture does not change, but is only graphically deformed. A more promising method for the problem posed here is focusing the PS at the location of some virtual source R, with the same signal X s (t) (in sum with interference), obtained from a real source S, but located at a point with coordinates <α R =α s , D R =r(Ds)>, associated with the desired coordinates s , Ds> given nonlinear dependence: r s =γ ф Ds​-1 , where γ ф - scale factor of the dimension m 2 .

[0004] A similar method of focusing the PS is used both in observation indicators [RU Patent No. 2788476, IPC G01S 3 / 80, declared 2021138452, 12.12.2021, reg. in the Civil Code of the Russian Federation on 19.01.2023], and in discriminators for precise determination and AS coordinates [RU Patent No. 2797779, IPC G01S 3 / 80, No. 2022123377, declared 31.08.2022, published 08.06.2023].

[0005] The closest to the proposed technical solution for the stated task and a set of common features is the “Method for determining the range to a moving radiation source using a passive sonar” [RU Patent No. 2797779, IPC G01S 3 / 80, No. 2022123377, 08 / 31 / 2022, published 06 / 08 / 2023], including the reception of signals with a circular wave front by a wide-aperture receiving system (RS) of M antennas (M≥3), located in space in a known manner, with a frequency response h k, preliminary processing, indication in the observation field within the specified boundaries of directions α and ranges D, ensuring detection of radiation sources, localization <α* s , D* s > the location of the IGI selected by the observer using the reference frame and the indicator's coordinate scales. At each successive i-th signal sample, a virtual reference focusing point of the receiving system R is assigned. oni with coordinates <α oni , r oni >.

[0006] To accurately determine coordinates and auto-tracking, this method uses the measurement of the sign and magnitude of the deviation of the coordinate estimate δ αi , and δ Di , contained in the current i-th signal sample, from the reference point <α oni , D oni >, obtained from previous measurements. To measure δ αi , and δ Di , calculate the pairwise differences in signal delays from the reference point to the centers of the m-th and n-th antennas of the PS i mni =t mi -t mi as well as weighting coefficients wai =∂τ mni / ∂α and w Di ==∂τ mni / ∂D, (m, n)=l,…, M, and measure the pairwise complex cross power density spectra (PCPS) of the input sample signals [x s mi ] of all mn-th pairs of PS antennas. Having measured the deviations δ αi , δ Di determine the estimates of α i *, D i * adjusting α oni , D oni by the value δ αi , δ Di accordingly, and the current coordinates of the radiation source α s (t i ), D s (t i ) at time t i and forecast of reference values ​​α oni+1 , D oni+1 for the next (i+1)-th step, the results of dynamic smoothing of the previous measurements are determined [Kuzmin S.Z. “Digital processing of radar information. Moscow: Sov. radio. 1967].

[0007] The advantage of the prototype method is the potential for passive determination of coordinates and automatic tracking by the receiving system in the near field of a moving IGi with an increase in the accuracy of range determination as a result of the fixed sharpening and symmetry of the main maximum of the PS response in its range cross-section, invariant to the value of the measured coordinate D. In this case, the peak of the PS and SO response and the zero value of the range discriminator's DH are located at the desired point D=Ds on the standard linear, uniform scale D.

[0008] The disadvantage of the prototype method is the need for preliminary adaptation of the signals of the real input vector sample X s =[X s mi ] from the radiation point S due to the mismatch of phase vectors for focusing (compensating) it at the virtual point R.

[0009] The objective of the proposed invention is to increase the accuracy of determining the range of a passive sonar using the focusing of the PS at a virtual point R on a plane with a nonlinear range scale over the entire range of near-field ranges, by transforming the input signal sample in order to adapt it to the geometry of the point R.

[0010] The technical result of the invention is to increase the accuracy of determining the range by focusing a hydroacoustic signal, received from the actual location of the radiation source S, by a passive sonar at a virtual point R on a nonlinear range scale D R , associated with the main, linear scale D by a nonlinear dependence D R (D S ), by precisely phasing it at the virtual point R, taking into account the difference in the phase vectors of the signal arriving at the PS antennas from these points.

[0011] To ensure the specified technical result in a method for passively determining the coordinates of a moving radiation source, comprising: receiving a signal in sum with interference by a wide-aperture receiving system of M antennas (M≥3) in the near zone (Fresnel zone), preliminary processing, including cyclic FFT and orientation of the PS antennas in the direction of the hydroacoustic radiation source, indication within a given sector of directions a and an interval of ranges D of the observation field with a coordinate grid (CG), observation of signal marks, selection of the CO for automatic tracking of a point S with coordinates <α S i , D S i >, where i is the number of the next FFT cycle, and at each next i-th signal sample, a virtual reference focusing point of the receiving system R is assigned oni , with coordinates <α oni , r oni > and determine the current estimates of the direction to the IGI α* i , and range D i* new features have been introduced with automatic tracking, namely:

[0012] -virtual reference point R oni placed in coordinates <α s i , r s i >, which coincide in direction, but differ in distance from the actual location of the radiation source S: α s i , r s i =γ ф D i s-1 , where γ ф - a scale factor equal to the square of the lower limit of the observation interval D assigned by the observer min .

[0013] - to determine the current value of the range by the passive sonar at each i-th FFT cycle, a conversion table H is organized SR Xi input M-dimensional sample, dimension M×M, using the well-known “rule of transformation of vector coordinates” [Maltsev A.I. Fundamentals of Linear Algebra - M.: Nauka. - 1975. - P. 81]. To form the table H SR Xiaccording to trigonometry formulas, using the coordinates of the centers of the PS antennas x m and the speed of sound in medium C, two sets of signal propagation times to the centers of the antennas of the PS A are calculated m (m=1,…,M): - from the expected (reference) point S oпi - [t s moпi ] with coordinates <α s i , D s i > and - from the virtual point R опi - [t R moпi ] with coordinates <α s i , r s i >. For each mn-th pair of antennas, the PS organizes two complex phase vectors at each frequency ƒ k : for the expected point S oпi : [exp(j2πƒ k t s moпi )]⇒[e s moпi ], and for the virtual point R oпi : [exp(j2πƒ k t s moпi )]⇒[e R moпi ], where ƒ k- discrete frequency in the reception range corresponding to the k-th element of the Fourier transform of the received signal, ⇒ - designation sign (for brevity), m=1,…, M. For brevity, the indices ƒ are used below. k we omit. At each frequency, these vectors are multiplied according to the rules of multiplication of complex vectors and subjected to the standard "matrix inversion" transformation: [[e s moпi ] [e s moпi ]* T ] -1 ⇒N SR Xi , obtaining the transformation table H SR Xi . At the next FFT cycle, the signal arrays of the input sample X i from antenna combiners [ ], containing the "useful" signal s m and interference n m , multiplied by H SR Xi , receiving the transformed signal samples [H SR mi ], as if transferred from the IGI from point S to point R (index (s m +n m ) we omit for brevity). To determine the direction α sand range D s first use the virtual reference point R oпi with coordinates <α s i , r s i > and the phase vector [e R moпi ], determining the deviations δ α andδ D and according to the signals of the input M-dimensional sample by measuring the parameters of the discriminatory characteristic, its “discriminatory” <c 1α , c 1D > and "normalizing" <c 2α , c 2D > parts. To do this, the differences in the signal arrival times from the virtual reference point to the centers of each mn-th pair of antennas τ are calculated. R mni (α oпi , r oпi )=t R mi -t R ni , as well as weighting coefficients w R αmni =∂τ R mni / ∂α and w R Dmni =∂τ R mni / ∂D taking into account the difference in the geometry of the relative position of the PS antennas and the radiation source at point R oпi .

[0014] - Then, for each mn-th pair of antennas, “discriminatory” values ​​of DH are obtained 1mni i-th FFT cycle, summing over frequency ƒ k imaginary parts of the PVSPM of each mn-th pair of antennas in the reception band <ƒ н , ƒ в > with each term multiplied by the frequency response coefficient h k , at frequency ƒ k and by the coefficient e kmni compensation for the difference in signal propagation times from the VOT to the antenna centers of each mn-th pair of antennas τ mni :

[0015] e kmni (τ R mn i(α oпi , r oпi ))=ехр(j2πƒ k (t moпi -t noпi )).

[0016] - Then the “normalizing” values ​​of DH are determined 2mni for each mn-th pair of antennas, summing over frequency ƒ k real parts of the WSPM with each term multiplied by h k , by the compensation coefficient e kmni and per square of frequency ƒ k 2 .

[0017] - Then the estimates of the deviations of the direction δαi and the range δ are determined Di : δ α i=c 1αi / c 2αi and δ Di =c 1Di / c 2Di , for which the values ​​with are summed up by index mn 1mni with weighting coefficients w R αmni and w R Dmni accordingly, obtaining the values ​​c 1αi and c 1Di which are divided into sums by index mn of quantities c 2mni with the same weighting coefficient squared (w R αmni ) 2 , receiving from 2αi and (w R Dmni ) 2 , receiving c 2Di . Ratings a i *, D i * obtained by adjusting α oпi , D oпi by the values ​​δ αi , δ Di - respectively, and the current coordinates of the radiation source a s (t i ), D s (t i ) at time t i and forecast of reference values ​​α oпi+1 D oпi+1for the next (i+1)-th step, it is determined by known methods of dynamic smoothing [Kuzmin S.Z. “Digital processing of radar information”, Moscow: Sov. radio. 1967].

[0018] The advantage of the proposed method for determining coordinates using a passive sonar is the implementation of an increase in the accuracy of determining the range of the PGL using the focusing of the PS on a plane with a nonlinear range scale as a result of adapting the input sample of the PS from a real point S to a virtual focusing point R by converting it using the H table SR xi .

[0019] The claimed technical result is achieved by introducing a phase transformation of the received signal, which is ensured by calculating the propagation times of the radiation source signal to the antenna centers from the expected point S oni , t s moni and from the virtual point R oni , t R moni , forming the transformation table H SRxi , applying it to the input sample and compensating for the propagation times of the converted signal. The introduction of new features allows the passive sonar to provide high resolution and accuracy in determining the range to the radiation source at all distances in the near zone, from the lower boundary of the observation interval D min to the upper boundary of the near field zone of the IGI.

[0020] The essence of the proposed invention is explained in Fig. 1-3.

[0021] Fig. 1: Diagram of geometric relationships <<receiving system / radiation source S i / virtual source R i >> using the example of a three-antenna PS, M=3. In the diagram of Fig. 1: A1, A2, A3 are the PS antennas; x, m are the axis of placement of the PS antenna centers; O is the origin (PS center); Si<α si , D si > - the actual location of the hydroacoustic signal emission source at the time of the i-th sample of the IGI signal: α si - direction, D si - distance to source S i; R i - location of the virtual focal point PS<α si R i (Si)>,L S 1, L S 2, L S 3- ray trajectories from point Si to the centers of antennas PS:, L R 1, L R 2, L R 3- ray trajectories from the virtual focal point R i (S i ) to the centers of the PS antennas.

[0022] In the diagram of Fig. 2: an example of a diagram of the algorithm of operations of a passive sonar with the transformation of the input sample to a virtual point R, implementing the proposed invention: 1 - reception of signal M by antennas PS A1÷A М , with preliminary processing (blocks 1.1÷1.M). 2 - transformation of the real input signal sample [ ] coming from point S, into the transformed sample [X SR mi ], as if coming from a virtual point R, by multiplying the input vector by the transformation table H SR xi .

[0023] 3 - measurement of pairwise mutual power density spectra (PMPS) of converted signals of mn pairs of PS antennas. 4 - measurement of the discriminatory component c 1i deviations α and D from the virtual reference point R oni . 5 - measurement of deviation in direction δ αi . 6 - formation of the H conversion table SR xi , by product of phase vectors and “matrix inversion”: [e s moni ]⋅[e R moni ]* Т =>H SR xi 7 - calculation of phase vectors [e s moni ]⋅[e R moni ]. 8 - measurement of normalizing components c2 αi and c 2Di deviations from the reference point R oпi , by α and D respectively. 9 - measurement of the deviation in range δ Di . 10 - calculation of signal trajectories from points S and R. 11 - control and indication of passive sonar. 12 - determination of coordinates, accumulation, dynamic smoothing of estimates.

[0024] In blocks 2, 6 and 7, at each successive i-th FFT cycle, the input sample transformation algorithm is performed.

[0025] In Fig. 3, a numerical example is shown of the relationship between the dependencies of the difference in signal delays τ mn in the mn-th pair of antennas of the PS PGL from the distance to the IGI D (Lm at Xm=0 and Ln at Xn=25 m, X is the distance from the center of the central antenna to the center of the antenna under consideration) for a conventional (prototype L R ), and for the transformed (proposed L SR ) of the input sample.

[0026] The claimed method, using the algorithm (Fig. 2) embedded in the passive sonar, is implemented as follows. Using control panel 11, the angle sector and observation range interval are set, and the lower limit of the range interval, Dmin, is specified. The appearance of signal marks from the radiation sources is monitored on the survey indicator. After selecting the SO, the direction is determined and the first virtual reference point is designated: D оп (t0)=D0*, α оп(t0)=α0*, after which tracking begins. Two sets of signal propagation times to the centers of the PS A antennas are calculated (block 10) m , (m=1,…,M):- from the expected (real) point S oni - [t s moni ] and - from the virtual point R oni - [t R moni ]. At each frequency ƒ k for each mn-th pair of antennas, the PS organizes (block 7) two complex phase vectors: for the expected (real) point S oni :

[0027] [exp(j2πƒ k t s moni )]=>[e s moni ] and for the virtual point R oni ,: [exp(j2πƒ k t R moni )]=>[e R moni ], m=1, …, M. These vectors are multiplied (block 6) and the “matrix inversion” transformation is applied: [e s moni ]⋅[e R moni ]* T ] -1 , obtaining the transformation table H SR xi . At the next FFT cycle, the input sample signal arrays [ ] from block 1 in block 2 is multiplied by H SR xi , obtaining the transformed signal samples [H SR mi ]. The PVSPM of the signals of the transformed input sample are measured (block 3). The differences in the signal arrival times from the virtual reference point to the centers of each mn-th pair of antennas τ are calculated (block 10). R mni (α oni ,r oni )=t R mi -t R ni , as well as weighting coefficients and w R αmni =∂τ R mni / ∂α and w R Dmni =∂τ R mni / ∂D For each mn-th pair of antennas, discriminatory values ​​of DH are obtained c 1mni i-th FFT cycle (block 4). Then the normalizing values ​​of the DH are determined (block 8): c 2αmni and c 2Dmni for each mn-th pair of antennas. Determine (blocks 5 and 9) estimates of the directional deviations δ αi (δ αi =c 1αi / c 2αi ) and range δ Di (δ Di =c 1Di / c 2Di), for which they sum up the values ​​by index c 1mni with weighting factors: w R αmni and w R Dmni accordingly, obtaining the values ​​c 1αi , and c 1Di , which are divided into sums by index mn of quantities c 2mni with the same weighting coefficient squared (w R αmni ) 2 and (w R Dmni ) 2 , receiving accordingly from 2αi With 2Di. After which (block 12) estimates of α are determined i *, D i *, adjusting α oni , D oni by the values ​​δ αi , δ Di accordingly, and the current coordinates of the radiation source α s (t i ), D s (t i ) at time t i and forecast of reference values ​​α oni+1 D oni+1 for the next (i+1)-th step, it is determined by known methods of dynamic smoothing.

[0028] Increasing the accuracy of determining the range of a passive sonar using the PS focusing at a virtual point R on a plane with a nonlinear range scale over the entire range interval of the near zone is achieved by taking into account the difference in the phase vectors of signals arriving at the PS antennas from two reference points: a real one with the expected coordinates and a virtual one with coordinates coinciding in direction with the real point, but differing in range. As follows from Fig. 3, the difference in signal delays τ mnIn the mnth pair of antennas, the PGL PS is practically independent of the range to the IGI D, unlike the prototype. The delay difference is stabilized by transforming the input sample from point S, which in turn allows the sample from point S to be adapted to the geometry of the virtual point R. Phase transformation of the received signal provides the passive sonar with high resolution and accuracy in determining the range to the radiation source at all distances in the near field. Current refinement of the radiation source coordinates a s (t i ), D s (t i ) at time t i are determined by known methods of dynamic smoothing.

Claims

A method for passively determining the coordinates of a moving radiation source, comprising: receiving a signal in combination with interference by a wide-aperture receiving system (RS) of M antennas (M≥3) in the Fresnel zone, preliminary processing including cyclic FFT and orientation of the RS antennas in the direction of the hydroacoustic radiation source (HES), indication within a given sector of directions α and a range interval D of the observation field with a coordinate grid (CG), observation of signal marks (SM), selection of the SM for automatic tracking (AT) of the actual location point of the radiation source S with coordinates <α i S ,D i S >, where i is the number of the next FFT cycle, and at each next i-th signal sample, a virtual reference focusing point of the receiving system R is assigned опi with coordinates <α опi , r опi > and determine current estimates of the direction to IGI α i * and range D i* with automatic tracking, characterized in that the virtual reference point R опi placed in coordinates <α i S ,r i S >, which coincide in direction, but differ in range from the actual location of the radiation source S: α i S ,r i S = , where γ ф - a scale factor equal to the square of the lower limit of the observation interval D assigned by the observer min , organize a conversion table for the input M-dimensional sample, for this purpose calculate two sets of signal propagation times to the centers of antennas A m (m=1, …, M): [t S moпi ] - from the expected (reference) point S опi and [t R moпi ] - from the virtual point R oпi , using the coordinates of the centers of the PS antennas x m and the speed of sound in the medium C, for each mn-th pair of antennas the PS organizes two complex phase vectors at each frequency ƒk : for the expected point S oпi [exp(j2πƒ k t s moпi )]⇒[e s moпi ] and for the virtual point R oпi [exp(j2πƒ k t R moпi )]⇒[e R moпi ], where ƒ k - discrete frequency in the reception range, ⇒ - designation sign (for brevity), m=1, …, M, these vectors are multiplied and subjected to the standard transformation “matrix inversion”: [e s moпi ]⋅[e R moпi ] *T ] -1 ⇒H Xi SR , obtaining the transformation table H Xi SR , at the next FFT cycle, the signal arrays of the input sample X i from antenna combiners [ ], containing the "useful" signal s m (ƒ k ) and interference n m (ƒ k ), multiplied by H SR Xi , obtaining the transformed signal samples [X SR mi ], to determine the direction of α sand range D s use a virtual reference point R опi with coordinates <α i S ,r i S > and phase vector [e R moпi ], determining the deviations δ α and δ D based on the signals of the input M-dimensional sample by measuring the parameters of the discriminatory characteristic (DC): its “discriminatory” <c 1α , c 1R > and "normalizing" <c 2α , c 2R > parts, for this purpose the differences in the signal arrival times from R are calculated опi to the centers of each mn-th pair of antennas τ R mni (α oпi , r oпi )=t R mi -t R ni , as well as weighting coefficients w R αmni and w R Dmni , taking into account the difference in the geometry of the relative position of the PS antennas and the virtual reference point R опi , for direction and range respectively, equal to the derivatives ∂τ R mni / ∂α and ∂τR mni / ∂r, respectively, determine the pairwise complex mutual power density spectra (PCMPD) of signals of all mn-x pairs of antennas, then obtain for each mn-th pair of antennas the discriminatory values ​​of DH c 1mni , summing over frequency ƒ k imaginary parts of the frequency response function of each mn-th pair of antennas in the reception band with the multiplication of each term by the frequency response coefficient h k , at frequency ƒ k and by the coefficient e kmni compensation of the time difference τ mni signal propagation from R oпi to the centers of the antennas of each mn-th pair of antennas: e kmni (τ R mni (α опi , r опi ))=exp(j2πƒ k (t mопi -t noпi )), then the normalizing values ​​of DH c are determined 2mni for each mn-th pair of antennas, summing over frequency ƒ k real parts of the WSPM with each term multiplied by h k , by the compensation coefficient e kmni and per square of frequency ƒ k2 , then the estimates of the deviations in direction δ are determined αi as δ αi =c 1ai / c 2αi and range δ Di as δ Di =c 1Di / c 2Di , for which the values ​​of c are summed by index mn 1mni with weighting factors: w R αmni and w R Dmni accordingly, obtaining the values ​​c 1ai and c 1Di , which are divided into sums by index mn of quantities C 2mni with the same weighting factor squared (w R αmni ) 2 , (w R Dmni ) 2 , obtaining accordingly c 2αi , and c 2Di , then determine the estimates of α i *, D i *, adjusting α опi , D опi by the values ​​δ αi , δ Di accordingly, and the current coordinates of the radiation source a s (t i ), D s (t i ) at time t i and forecast of reference values ​​α опi+1 , D опi+1for the next (i+1)-th step, it is determined by known methods of dynamic smoothing.

Images