RU2806448C1 - Method for detecting manoeuvring small-sized air objects using parametric transformations and device for its implementation - Google Patents

Abstract

FIELD: radar.

SUBSTANCE: invention can be used in surveillance and sector radar stations (radars) to solve problems of detecting manoeuvring small air objects (targets) in conditions of low signal-to-noise ratios. The claimed method involves lowering the primary signal detection threshold, accumulating marks over several radar surveys, grouping marks, using parametric Hough and Radon transformations to the formed groups, restoring trajectory segments and combining them according to spatio-temporal parameters. After which the most probable trajectories are selected and at the same time a decision is made to detect targets. A device for implementing the method is also claimed.

EFFECT: increase in the detection range of manoeuvring small-sized air objects and increase in the probability of detecting true trajectories in conditions of low signal-to-noise ratios.

3 cl, 4 dwg

Description

Field of technology to which the invention relates

The invention relates to the field of radar and can be used in surveillance and sector radar stations (radars) to solve problems of detecting maneuvering small air objects (targets) in conditions of low signal-to-noise ratios.

State of the art

There is a known method for detecting the trajectory of a maneuvering object, which consists in probing the space with packs of coherent radio pulses, performing an analog-to-digital conversion of the complex signal reflected from the maneuvering object, represented by in-phase (real) and quadrature (imaginary) components, obtaining signal samples, and performing a discrete Fourier transform , determine the modules of the obtained values, calculate the mixed posterior probability distribution density at a set of points determined by amplitude, range, speed, acceleration, calculate the ratio of the number of points in which the object exists at the time of calculation and at the previous step to the total number of points, compare the resulting ratio with a threshold, based on the comparison, a decision is made to detect or not detect the trajectory of an object, characterized in that the discrete Fourier transform is carried out for a plurality of acceleration channels, in each of which, before calculating the discrete Fourier transform, a phase shift is carried out to compensate for the inter-period phase shift due to acceleration , in accordance with the phase shifts, acceleration estimates are formed, which are the coordinates of points for calculating the mixed posterior probability distribution density [1]. In the considered method, the problem of detecting a trajectory described by a 2nd order polynomial is solved; the jerk of the target during the observation interval is not taken into account. To implement this method, you must have a multi-channel device. In addition, calculating the posterior probability density function at each step significantly increases the computational cost.

There is a known method for radar detection of a target trajectory, providing for a regular survey of space by radar probing of each angular direction, subsequent reception of an echo signal, its processing and making a decision on detecting a target trajectory based on the results of several reviews, characterized in that the decision to detect a target trajectory is made according to a criterion exceeding the threshold for confirming the hypothesis by the value of sufficient statistics, which is calculated as the sum of the squares of the maximum signal values from the local area of the first review and from K-1 gates from subsequent reviews, the positions of the gates for subsequent reviews are formed based on the positions of the maximum signal values in the gates of previous reviews, in the case to detect the target trajectory, the trajectory points are chosen to be the coordinates of the resolution elements selected over K reviews according to the criterion of maximum signal amplitudes [2]. The disadvantage of the considered method is the need to calculate sufficient statistics, compare them with a threshold for each resolution element and select the strobe size from review to review. In this case, the method does not take into account changes in the parameters of the target movement model, which can lead to the detection of false trajectories and the omission of true trajectories.

The closest method of target detection, including coordinated filtering of the received signal, followed by isolating the signal envelope and comparing it in each discrete range with the detection threshold, according to the invention, an additional channel for processing radar information is introduced, including the process of comparing the signal with the detection threshold of the first detection stage, which less than the threshold level of the prototype channel, analog-to-digital conversion of all passed signals, their inter-observer accumulation, conversion and storage of the results in the Hough drive, further comparison of the conversion results with the threshold device of the second stage of detection, based on the results of which traces of all targets are formed, the signals of which have passed the threshold of the first stage detection of an additional channel, which is also the fact of target detection [3]. The inability to detect maneuvering air objects and the presence of an additional channel in the device are disadvantages of this method.

The closest device [4] contains a series-connected matched filter 1, a signal envelope detector 2, a threshold device 3, a random access memory device 4, the inputs of which are the outputs of the threshold devices from each channel, a computing device 5, the output of which is the input of the mark gating unit 6, Outputs 1 and 2 of which are the input of the channels for processing marks of the block for connecting rectilinear trajectories, each of which solves the problem of identifying a trajectory on a plane, and consists of a series-connected block for implementing the Hough algorithm, which includes a computing device 7, a discretizer 8 and a counter 9, an extraction block cells from cluster 10, speed mark selection block 11, the output of which is connected to the input of the trajectory combining block 12, the output of which is connected to the input of the threshold device 13, an additional block for linking curvilinear trajectories is introduced, the input of which is the 3 output of the mark gating block 6, consisting of sequentially connected block of implementation of the Radon algorithm, which includes a computing device 14, a discretizer 15, a counter 16, a block for selecting cells from a cluster 17, a mark selection block 18, a threshold device 19, the output of the straight-line trajectories tying blocks and the curvilinear trajectories tying block is the input of the trajectory smoothing block 20, the output of which is the output of the entire device. This device provides a solution to the problems of detecting maneuvering target trajectories in spaced radar systems, where the time of accumulation of marks is limited by the location of targets within the radar barrier. It is a priori specified in the device that during the accumulation time, targets can move rectilinearly or along an elliptical arc (circle). Thus, the presented device cannot be used to detect small-sized maneuvering targets in surveillance radars, due to the fact that during the accumulation time, targets can move along complex trajectories consisting of several straight segments and elliptical arcs (circles).

Disclosure of the Invention

The problem to be solved (technical result) is to increase the detection range of maneuvering small-sized air objects and increase the probability of detecting true trajectories in conditions of low signal-to-noise ratios.

This technical result is achieved by the fact that in the method of radar detection of targets, including coordinated filtering of received signals from targets with subsequent selection of their envelopes, analog-to-digital conversion, comparison in each resolution element with the detection threshold and their inter-survey accumulation, according to the invention, two channels are introduced processing of radar information, in contrast to the prototype, where parametric Hough and Radon transformations are carried out for the accumulated sample of marks, further comparison of the results of transformations in the threshold device of the second detection stage, based on the results of which trajectory segments are restored and then combined (identified) according to spatio-temporal parameters, after whereby the most probable (true) trajectories are selected by comparison with the third detection threshold and at the same time a decision is made to detect targets.

This result is also achieved by the fact that in a radar target detection device containing a series-connected matched filter 1, a signal envelope detector 2, a threshold device 3, an ADC unit 4, a random access memory device 5, a computing device 6, a mark gating unit 7, outputs 1 and 2 which, according to the invention, are the inputs of the channels for processing the marks of the block for tying straight segments of trajectories, each of which solves the problem of identifying a trajectory on the plane, and consists of a series-connected block for implementing the Hough algorithm, which includes a computing device 8, a discretizer 9 and a counter 10, block selecting cells from cluster 11, mark selection block by speed 12, the output of which is connected to the input of the trajectory combining block 13, the output of which is connected to the input of the threshold device 14, the curvilinear trajectory linking block whose input is the 3 output of the mark gating block 7, consisting of series-connected block for implementing the Radon algorithm, including a computing device 15, a discretizer 16, a counter 17, a block for selecting cells from a cluster 18, a block for selecting marks 19, a threshold device 20, the output of the blocks for linking rectilinear and curvilinear trajectory segments is the input of an additionally introduced block for identifying trajectory segments 21, the output of which is the input of the block for selecting true trajectories 22 and the input of which is connected to the block for storing trajectory parameters 23, the output of block 22 is the output of the entire device.

The essence of the proposed method is as follows.

The radar surveys the space with a set threshold that ensures the detection of weak echo signals from maneuvering small air objects [5]. Signals from targets and interference signals that exceed the detection threshold, in the form of marks corresponding to the coordinates of range, azimuth and elevation, are stored in the radar receiving device in a data array.

The resulting data array is subjected to the well-known parametric Hough and Radon transformations [6-8], which make it possible to identify characteristic segments of straight lines obtained during the accumulation of multi-view information from airborne objects moving in a straight line, and curves obtained during the accumulation of multi-view information from maneuvering air objects.

Detection occurs based on an a priori model of the movement of air objects. Marks of rectilinearly moving air objects, obtained in several periods of review, will line up in an almost straight line sequentially from review to review at relatively equal distances. Marks from maneuvering air objects obtained in several periods of review will be lined up in a curved line (ellipse arc) sequentially one after another from review to review at relatively equal distances. False interference marks may appear randomly in different resolution elements [9] (Fig. 1).

After applying parametric transformations, groups of marks belonging to rectilinear and (or) curvilinear segments of trajectories are obtained (Fig. 2). Next, the number of marks in the group is compared with the threshold according to the specified criterion for detecting trajectory segments “m of Where - detection threshold of straight trajectory segments, - detection threshold of curved trajectory segments, and on this basis a decision is made on detection or non-detection of target trajectory segments.

After detecting trajectory segments, the problem of segment identification is solved to detect the resulting (true) trajectory. To correctly identify and combine trajectory segments, similar marks are identified and excluded from the same trajectory segment. The fusion parameters are the values of the target motion parameters on the final (initial) segment and the time of detection of the first (last) mark of the trajectory segment.

The result of merging may be false trajectories. To eliminate them, the parameters of the linked trajectories are compared with the parameters corresponding to the true trajectories, and the most probable (true) ones are selected from all the linked ones. This allows us to solve the problem of detecting true trajectories and subsequently the problem of target detection.

Based on a parametric transformation, the method for detecting the trajectories of maneuvering air objects is highly resistant to a large number of false marks and missing true marks, which makes it possible to solve the problem of detecting the trajectories of maneuvering air objects with high efficiency in conditions of low signal-to-noise ratios.

Thus, lowering the primary detection threshold of echo signals, accumulating signal marks that exceeded the threshold, and further processing using the parametric Hough and Radon transformations makes it possible to increase the detection range and the probability of detecting maneuvering small-sized air objects.

The inventions are illustrated by the following drawings:

Fig. 3-the closest radar target detection device that implements the closest method.

Figure 4 - the claimed device for detecting maneuvering small-sized air objects (targets), implementing the claimed method.

A device for detecting maneuvering small-sized air objects (targets), implementing the claimed method, contains (Fig. 4) a series-connected matched filter 1, a signal envelope detector 2, a threshold device 3, an ADC unit 4, a random access memory device 5, a computing device 6, a gating unit marks 7, outputs 1 and 2 of which, according to the invention, are the inputs of the mark processing channels of the block for tying straight segments of trajectories, each of which solves the problem of identifying a trajectory on the plane, and consists of a series-connected block for implementing the Hough algorithm, which includes a computing device 8, a discretizer 9 and a counter 10, a block for selecting cells from a cluster 11, a block for selecting marks by speed 12, the output of which is connected to the input of the trajectory combining block 13, the output of which is connected to the input of the threshold device 14, a block for tying curvilinear segments of trajectories whose input is the 3rd output of the gating block marks 7, consisting of a series-connected block for implementing the Radon algorithm, including a computing device 15, a discretizer 16, a counter 17, a block for selecting cells from a cluster 18, a mark selection block 19, a threshold device 20, the output of the blocks for connecting rectilinear and curvilinear trajectories is the input an additionally introduced block for identifying trajectory segments 21, the output of which is the input of a block for selecting true trajectories 22 and the input of which is connected to a block for storing parameters of true trajectories 23, the output of block 22 is the output of the entire device.

Industrial applicability.

The inventive device can be implemented as part of a hardware and software complex. For example, the Pantsir-S1 anti-aircraft missile and gun target detection station includes information processing and control equipment (IPOU), which is a hardware-software complex [10]. The fiducial marks are processed in physical devices - programmable logic integrated circuits FPGA1 and FPGA2. The processing process uses memory chips (ZBT), random access memory (DDK) and signal processors (DSP). Data from the DSP is sent to the FPGA Sync synchronization chip. It synchronizes the operation of two processing FPGAs, as well as transmits data via an Ethernet channel to ETX Core 2 DUO modules. ETX modules carry out trajectory processing. Each ETX module has flash memory for loading the operating system. After secondary processing is completed, packages of targets and trajectories are transferred to the central computer system (CCS). The PCI bus controller is connected to the ETX modules to perform switching with the multiplex information exchange channel for communication with the central digital computer. These elements form the basis of the hardware.

The software part of the complex consists of an information processing algorithm that solves the following problems:

- processing of detected echo signals at an intermediate frequency in the receiving device and comparing them with the threshold level of the first stage of detection;

- analog-to-digital conversion of signals from the output of a threshold device and their inter-review accumulation;

- parallel implementation of parametric transformations of accumulation results using the Hough and Radon algorithms;

- comparison of the results after parametric transformations with the threshold at the second detection stage and the formation of trajectory segments;

- combining trajectory segments according to a given criterion and forming target trajectories;

- selection of the most probable (true) trajectories by comparison with the third detection threshold, making a decision on target detection, as well as transferring trajectories to the tracking device;

- storage of a priori parameters of true trajectories for the third stage of detection.

Carrying out the invention

The method is carried out as follows.

With the help of the radar, it surveys the space and detects targets, with a given threshold (depending on the required quality indicators), ensuring the detection of weak echo signals from maneuvering small-sized air objects. Signals from targets and interference signals that exceed the detection threshold are stored in an array of data corresponding to the coordinates of range, azimuth and elevation. The use of the Hough and Radon algorithms makes it possible to detect rectilinear and curvilinear segments of trajectories based on the results of the accumulation of radar information over several radar surveys. Echo signals of rectilinearly moving targets, received in several viewing periods, will line up in an almost straight line sequentially from review to review at relatively equal distances. Marks from maneuvering air objects obtained in several periods of review will be lined up in a curved line (ellipse arc) sequentially one after another from review to review at relatively equal distances. False interference marks may appear randomly in different resolution elements.

The results of comparison with the threshold of the first stage of detection of block 3, obtained in each review, are digitized in block 4, saved in block 5, in block 6 converted from a spherical coordinate system to a Cartesian one, and in block 7 subjected to grouping.

According to the accepted concept of the block construction of the trajectory detection algorithm, gating is carried out for all marks arriving during the accumulation time that are not identified with the detected trajectories. The center of the capture strobe is taken to be the initial mark of the detected trajectory, and its size is traditionally determined using a priori specified minimum and maximum speeds of movement of the intended targets. The groups of marks formed after gating are subject to parametric Hough and Radon transformations.

From the mark gating block, the data is supplied to the inputs of the mark processing channels of the rectilinear trajectories linking block, where segments of rectilinear trajectories are selected on the XY, YZ planes. The output of the Hough algorithm implementation block will be matrices with marks belonging to the same line, traces of which fall into a cell with an exceeded threshold.

The input of the speed mark selection block receives matrices corresponding to sets of marks belonging to the same line. Mark selection is performed to select several trajectories lying on the same line and filter out false marks that fall on the same line. All marks belonging to the detected line are grouped by time. Next, the distances between marks that belong to different time cells are calculated. Groups (proposed trajectories) are formed from all marks as follows: firstly, only one mark from each time cell can belong to a group (time selection), and secondly, the distances between cell marks should not exceed the maximum permissible value.

At the output of the block, a matrix is formed corresponding to the trajectory of the air object in space. If the number of marks on one trajectory exceeds the threshold value, a decision is made to detect the trajectory, and a straight segment of the trajectory is formed, composed of marks with coordinates x, y, z, t. The resulting data array is input to the trajectory segment identification block.

Also, from the mark gating unit 7, data is supplied to the input of the block for tying curvilinear trajectories into the computing device of the block for implementing the Radon algorithm, where curvilinear segments of the trajectories of air objects are detected.

After linking the trajectory segments, the problem of identifying them in block 21 with each other is solved in order to detect trajectories, for example, by combining the end of one segment with the beginning of another. The detected trajectory segments may contain identical marks. To correctly identify and combine trajectory segments, identical marks are found and excluded from one trajectory segment. The parameters of the combination are the speed of movement of targets on the final (initial) segment and the time of detection of the first (last) mark.

After combining the segments, the number of marks located along the trajectory is compared with the threshold using the “m out of n” criterion, and on this basis a decision is made to detect or not detect trajectories. The result of merging may be false trajectories. To eliminate them, the parameters of the trajectories are compared with the parameters corresponding to the true trajectories in block 23, and the most probable (true) ones are selected from all those detected in block 22. In the case when several trajectories (including false ones) are initiated, each of they will be assigned a certain weight. Based on the comparison results, the trajectory with the highest weight will be accepted as the true one. Simultaneously with the decision to detect the trajectory, a decision is made to detect the airborne object.

Thus, the problem of detecting maneuvering small-sized air objects is solved simultaneously with the construction of their trajectories using parametric transformations, which achieves the claimed technical result.

Bibliography

1. Pat. 2553459 Russian Federation, IPC G01S 13/06 (2006.01). Method for detecting the trajectory of a maneuvering object / V.A. Belokurov, D.N. Kozlov, V.I. Koshelev; applicant and patent holder of the Ryazan State Radio Engineering University. No. 2014104650/07; application 02/10/2014; publ. 06/20/2015, Bulletin. No. 17.

2. Pat. 2710202 Russian Federation, IPC G01S 13/58 (2006.01). Method for radar detection of target trajectory / I.V. Sausage; applicant and patent holder I.V. Sausage. - No. 2019113588; application 04/30/2019; publ. 12/25/2019, Bulletin. No. 36.

3. Pat. 2777652 Russian Federation, IPC G01S 13/52 (2006.01). A method for detecting small airborne objects and a device for its implementation / A.V. Golubev; applicant and patent holder A.V. Golubev. -No. 2021108285; appl. 03/26/2021; publ. 08/08/2022, Bulletin. No. 22.

4. Pat. 2776417 Russian Federation, IPC G01S 7/41 (2006.01), G01S 13/04 (2006.01). Complex detector of curvilinear trajectories of air objects using parametric transformations / N.A. Leshko, I.S. Ashurkov, A.V. Kadykov and others; applicant and patent holder FGKVOUVO Yaroslavl Higher Military School of Air Defense. - No. 2021134114; application 11/22/2021; publ. 07/19/2022, Bulletin. No. 20.

5. Signal detection theory. P.S. Akimov, P.A. Bakut, V.A. Bogdanovich et al.; Under. ed. Bakuta P.A. - M.: Radio and communication, 1984.

6. Konovalov A.A. Fundamentals of trajectory processing of radar information. St. Petersburg: Publishing house of St. Petersburg Electrotechnical University "LETI", 2013. pp. 145-148.

7. Ashurkov I.S., Zhitkov S.A., Zakharov I.N., Leshko N.A., Moroz A.V., Sakhno I.V., Simulation model of the process of detecting aerodynamic targets based on parametric transformations in a multi-position location system // Advances in modern radio electronics. 2020. T. 74. No. 12. pp. 2-44.

8. Pat. 27322916 Russian Federation, IPC G06G 7/78 (2006.01), G01S 13/00 (2006.01). Complex detector of the rectilinear trajectory of an air object in space using the Hough transform / S.A. Zhitkov, N.A. Leshko, I.S. Ashurkov, I.N. Zakharov, A.N. Tsybulnik, applicant and patent holder of the Yaroslavl Higher Military School of Air Defense. - No. 2019119734; application 06/24/2019; publ. 09/24/2020, Bulletin. No. 27.

9. Kuzmin S.Z. Digital radar. Introduction to theory. - Kyiv: KVShch, 2000.

10. Operation manual. Part 4. Target detection station. 72B6-E2.00.00.000-01RE3.

Claims (2)

1. A method for detecting maneuvering small-sized air objects, including coordinated filtering of signals received by a radar station (radar) with subsequent selection of their envelopes, analog-to-digital conversion, comparison in each resolution element with the detection threshold and their inter-review accumulation, characterized in that all marks arriving during the accumulation time, while the center of the capture strobe is taken to be the initial mark of the detected trajectory, and its size is determined using the specified minimum and maximum speeds of movement of the intended targets; the groups of marks formed after gating are transmitted to two channels for processing radar information, where for the accumulated samples of marks carry out parametric Hough and Radon transformations, then compare the results of transformations in the threshold device of the second stage of detection, based on the results of which rectilinear and curvilinear segments of the trajectories of the intended targets are restored based on the results of the accumulation of radar information over several radar surveys, combine the segments of the trajectories according to spatiotemporal parameters, identical marks are found, which are excluded from one segment of the trajectory, while the values of the speed of movement of targets on the final or initial segment and the time of detection of the first or last mark, respectively, are used as merging parameters; after merging the trajectory segments, the number of marks located along the trajectory is compared with a threshold according to the “m out of n” criterion and make a decision on detection or non-detection of the target trajectory, after which they select the most probable trajectories by comparison with the third threshold for detecting true trajectories and at the same time make a decision on target detection. 2. A device for detecting maneuvering small-sized air objects, containing a series-connected matched filter, a signal envelope detector, a threshold device, an ADC unit, a random access memory device, a computing device, a mark gating unit, characterized in that the outputs of the mark gating unit are the inputs of the mark processing channels of the block tying of rectilinear segments of trajectories and a block of tying of curved segments of trajectories, each of which solves the problem of identifying a trajectory on the plane, while the block of tying rectilinear segments of trajectories consists of a series-connected block for implementing the Hough algorithm, which includes a computing device, a discretizer and a counter, a block for selecting cells from a cluster, a block for selecting speed marks, the output of which is connected to the input of a block for combining trajectories, the output of which is connected to the input of a threshold device, and the block for linking curvilinear segments of trajectories consists of a series-connected block for implementing the Radon algorithm, which includes a computing device, a sampler, and a counter , a block for selecting cells from a cluster, a block for selecting marks by speed, a threshold device, the output of the blocks for linking rectilinear and curvilinear trajectory segments is the input of an additionally introduced block for identifying trajectory segments, the output of which is the input of a block for selecting true trajectories and the input of which is connected to a block for storing true parameters trajectories, the output of the block is the output of the entire device.