RU2558674C1 - Method for automated detection of compact groups of interacting aerial objects - Google Patents

Abstract

FIELD: radio engineering, communication.

SUBSTANCE: radar station measures slant range to observed aerial objects, azimuths, elevation angles and radial velocities thereof. Interval estimates of the measured coordinates of the aerial objects are generated taking into account measurement errors thereof. A plurality of aerial objects is presented in the form of a graph of interaction thereof, the vertices of which correspond to objects and the edges reflect satisfaction of conditions of possible or reliable interaction thereof. Reliable and possible groups of objects are selected. Longer tracking of groups of aerial objects is achieved owing to that possible groups are detected earlier than reliable groups, and in case of short-term divergence of interacting aerial objects, reliable groups are not dropped from tracking right away but are transferred to the category of possible groups.

EFFECT: longer tracking of groups of aerial objects owing to earlier detection thereof.

2 cl, 6 dwg

Description

The invention relates to radar and can be used for radar tracking of air objects (IN).

Currently, radar systems (radar) are the main means of detection and tracking of HE. All modern radars have limited bandwidth. In this regard, in cases of massive use of HE, characteristic of military conflicts of recent decades, the number of HE found can exceed the specified limit, as a result of which the radar will not be able to fully provide a solution to its task. This phenomenon is known as the “swarm effect” [1].

Modern methods of using aviation provide for the widespread use of group actions of HE, including as part of groups with a small distance between objects. In this regard, it becomes necessary to identify such HE groups in the process of analyzing the air situation. In addition, tracking of single objects in such groups, performed in modern radars using identification gates, is significantly complicated due to the multiple intersection of gates [2].

These problems can be overcome by summarizing (enlarging) the processed and displayed information about the air situation. The generalization of information may consist in identifying compact groups (groupings) of military personnel that solve the general problem during the course of hostilities. The identified group of interacting HE can be accompanied and displayed as one goal, and not as several goals, which will reduce the load on the radar and its operators [3].

Currently, the grouping of interacting HEs is usually performed manually by the radar operator [3], which creates an additional load for it.

Thus, the task of automating the identification of compact groups of interacting HEs is very urgent.

Currently known methods of automated grouping of HE are designed to achieve the following goals:

- reducing the amount of information displayed to decision makers (DM);

- facilitating the determination of the composition of the group and the identification of the intent of the air strike;

- reducing the amount of information processed by computing and transmitted over the data channel.

Grouping is usually carried out using the following signs of the interaction of HE in the group [4, 5]:

- a relatively small distance between objects in space, providing the possibility of visual or radar contact between them;

- coincidence of courses and speeds of objects in the group.

Thus, for grouping, such coordinates of the state (CS) of HE can be used, such as ranges to HE, their azimuths, elevation angles, radial velocities, rectangular coordinates, velocity modules, and courses. Grouping is distinguished on the basis of primary radar information (RLI), which includes oblique ranges, azimuths, elevation angles and radial velocities of HE [3]. Grouping HE and the formation of traces of group objects directly according to the primary radar data allows to solve the following problems:

- preliminary analysis of the air situation;

- elimination of the time-consuming operation of selecting marks of individual HEs performing flight as part of a group with a small distance between objects obtained on different surveys of radar airspace necessary for individual tracking of HE.

There are various approaches to automating the grouping of HE. In [4], it is proposed to group them in such a way as to ensure a minimum of the ratio r _sr / R _sr , where r _sr is the average distance between the objects in the group, R _cp is the average distance between the groups in the space, including the coordinates of the objects, their courses, speeds, and also signs of nationality, etc. The automatic solution to this problem is based on an orderly enumeration of options for dividing the set of observed VOs into groups and finding an option that provides a minimum of r _cp / R _cf. The main disadvantage of this approach is the need for multiple generation of options for splitting the set of VO into groups.

A simpler method of grouping from the point of view of implementation was proposed in [5]. According to this method, the set of detected VOs is represented as a graph of their interaction. The vertices of this graph correspond to the observed VO, and the edges reflect the possibility of interaction between a pair of VO. The connected components of this graph [6] are groups of interacting VOs. In this case, two types of interaction in the group can be distinguished:

- direct interaction when the VO pair is connected by an edge in the interaction graph;

- indirect interaction, when between a pair of HE there are at least one path in the interaction graph and all paths between contain more than one edge.

и $${\langle x_1^q,x_2^q,\dots ,x_n^q\rangle }^T$$

соответственно. Компонентами данных векторов могут быть наклонная дальность, азимут, угол места и радиальная скорость ВО. Тогда непосредственное взаимодействие между ВО p и q возможно при выполнении следующего условия:The condition for the direct interaction of a VO pair is formally expressed as follows. Let the vectors of their CS be given for HE p and q $${〈x_{\mathrm{one}}^p,x_2^p,...,x_n^p〉}^T$$

and $${〈x_{\mathrm{one}}^q,x_2^q,...,x_n^q〉}^T$$

respectively. The components of these vectors can be oblique range, azimuth, elevation and radial velocity VO. Then a direct interaction between VO p and q is possible under the following condition:

where ε _i is the a priori specified maximum allowable for interacting HE differences in the values of ix CS. The value of ε _i can be determined a priori based on the study of the experience of group actions of aviation [7].

There is currently no single approach to choosing ε _i for various CSs. So, for example, in [3], the values of ε _i for the spatial coordinates of the HE are determined on the basis that the HE interact in the group based on visual or radar observation of each other. Moreover, the choice of ε _i depends on the illumination, weather conditions, characteristics of the airborne radar and other factors. Given these circumstances, in [3] for rectangular coordinates of VO in the plane, it is proposed to use the values of ε _i in the range from 200 m to 5000 m.

Thus, as an approach for determining the size of the strobe of grouping HEs for various CSs, one can take an expert assessment of the values of ε _i .

The task of grouping HE is solved as the task of identifying the connected components of the graph of interaction of HE. In [5], it was shown that the best algorithm for solving this problem from the point of view of time expenditures is the algorithm for traversing the graph in depth [8].

A significant drawback of the described method for grouping VOs is that the result of checking conditions (1) can change with random changes in the differences in the CS of the VO pair near the threshold value caused by errors in their estimation, while there is no real change in the composition of VO groups. At the same time, since from the tactical point of view it is precisely the groups of interacting HEs that constitute the greatest danger, it is important, on the one hand, to identify the formation of such groups as early as possible, and on the other hand, to prevent their premature removal from accompaniment as a result of a short-term difference in HE or random changes in the ratings of their COP.

To overcome this drawback, it is necessary to use milder conditions for the belonging of the HE pair to the same group, which allows assessing the reliability of the decision on the belonging of HE to the identified groups. At the same time, the use of several degrees of confidence will mitigate the effect of random changes in the estimates of the CS of the HE on the grouping result. For this, it is necessary to express in an explicit form the uncertainty of the values of the CS of the HE that arise due to errors in their estimation, and take it into account when checking pairs of HE for belonging to one group.

Currently, sources are unknown in which the methods of grouping HE are published that implement the uncertainty of the CS of HE and the assessment of the reliability of belonging of HE to the selected groups.

The closest analogue (prototype) is the above method of grouping VO, using expression (1) as a condition for a VO pair to belong to one group, representing the VO set in the form of a graph of their interaction and the depth search algorithm to identify VO groups.

The aim of the claimed invention is to develop a method for identifying compact groups of interacting HEs, which takes into account the uncertainty of their CS values in an explicit form and assesses the reliability of HE belonging to the selected groups.

The technical result of the invention consists in increasing the time of tracking VO groups due to their earlier identification and preventing premature removal from escort, reducing the burden on radar operators, obtaining explicit estimates of the degree of reliability of the grouping results, taking into account the uncertainty of the CS VO.

The technical result is achieved due to the fact that, in contrast to the prototype, the values of the numerical CS of VO are represented by intervals. Three types of relationships between HE pairs are introduced (lack of interaction, possible interaction, reliable interaction). To specify these relations, use arithmetic operations and relations over intervals [9, 10]. The entire set of observed VOs is represented in the form of a graph of their interaction with two types of edges corresponding to the possible and reliable belonging of a pair of objects to one group. Using a modified depth search algorithm, possible and reliable VO groups are identified. The composition of reliable and possible VO groups is displayed for decision makers.

, где для всех $$i=\overline{\mathrm{1..}n}$$

$$x_i^p$$

- фактическое значение i-й КС ВО p, $${\xi }_i^p$$

- ошибка ее измерения. Координата $$x_1^p$$

соответствует наклонной дальности до ВО, $$x_2^p$$

- азимуту ВО, $$x_3^p$$

- углу места ВО, $$x_4^p$$

- радиальной скорости ВО.The initial data for the grouping of HE are estimates of the SC of HE obtained as a result of primary processing of radar images. Information about each observed HE p comes in the form of a vector of measured values of its CS $$x_{\mathrm{AND}}^p={〈x_{\mathrm{one}}^p+{\xi }_{\mathrm{one}}^p,x_2^p+{\xi }_2^p,x_3^p+{\xi }_3^p,x_{\mathrm{four}}^p+{\xi }_{\mathrm{four}}^p〉}^T$$

where for everyone $$i=\overline{\mathrm{one..}n}$$

$$x_i^p$$

- the actual value of the i-th KS VO p, $${\xi }_i^p$$

- the error of its measurement. Coordinate $$x_{\mathrm{one}}^p$$

corresponds to an inclined range to VO, $$x_2^p$$

- azimuth of VO, $$x_3^p$$

- the corner of the VO place, $$x_{\mathrm{four}}^p$$

- radial velocity VO.

гауссовским. Тогда отклонение измеренного значения i-й КС ВО р от фактического с вероятностью, близкой к единице, не превышает $$3{\sigma }_i^p$$

, где $${\sigma }_i^p$$

- среднеквадратичное отклонение (СКО) величины $${\xi }_i^p$$

.Since the combination of a large number of various random processes affects the result of measurements of CS, by virtue of the central limit theorem, we can consider the distribution of values of $${\xi }_i^p$$

Gaussian. Then the deviation of the measured value of the i-th KS VO p from the actual one with a probability close to unity does not exceed $$3{\sigma }_i^p$$

where $${\sigma }_i^p$$

- standard deviation (RMS) of the value $${\xi }_i^p$$

.

Thus, each CS can be represented by an interval with lower and upper boundaries

- текущая оценка i-й КС ВО p, $$\Delta x_i^p=3{\sigma }_i^p$$

.respectively, where $${\stackrel{\wedge }{x}}_i^p=x_i^p+{\xi }_i^p$$

- the current assessment of the i-th COP in p, $$\Delta x_i^p=3{\sigma }_i^p$$

.

The uncertainty of the compared interval CS of the HE leads to the uncertainty of the results of their pairwise comparison and, as a result, to the uncertainty of the result of checking the conditions of interaction of the pair of HE proposed in [5]. To account for this uncertainty, three types of interaction relations between pairs of HE are introduced on the set of HE:

- lack of interaction;

- possible interaction;

- reliable interaction.

As in the prototype of this invention, to represent information about VO and the relationships between them, the VO interaction graph G = 〈V, E〉 is used, where V is the set of vertices of the graph, each of which corresponds to a certain VO, E is the set of edges reflecting the presence of interaction between VO.

However, unlike the prototype, the edge of the graph can already display one of two types of direct interaction, namely:

- possible direct interaction;

- reliable direct interaction.

In contrast to the prototype, the following rule is applied to check for the presence of direct interaction between a pair of VO p and q, obtained by modifying condition (1) using arithmetic operations on intervals [9] and the relation of inequality of intervals [10].

(см. фиг.1а), то ВО p и q не взаимодействуют.If there exists i∈ {1, 2, ..., n} such that $$\left|{\stackrel{\wedge }{x}}_i^p-{\stackrel{\wedge }{x}}_i^q\right|>{\epsilon }_i+\Delta x_i^p+\Delta x_i^q$$

(see figa), then IN p and q do not interact.

и $$\left|{\stackrel{\wedge }{x}}_i^p-{\stackrel{\wedge }{x}}_i^q\right|>{\epsilon }_i-\Delta x_i^p-\Delta x_i^q$$

(см. фиг.1б), то ВО р и q возможно непосредственно взаимодействуют.Otherwise, if there exists i∈ {1, 2, ..., n} such that $$\left|{\stackrel{\wedge }{x}}_i^p-{\stackrel{\wedge }{x}}_i^q\right|\le {\epsilon }_i+\Delta x_i^p+\Delta x_i^q$$

and $$\left|{\stackrel{\wedge }{x}}_i^p-{\stackrel{\wedge }{x}}_i^q\right|>{\epsilon }_i-\Delta x_i^p-\Delta x_i^q$$

(see figb), then p and q possibly directly interact.

$$\left|{\stackrel{\wedge }{x}}_i^p-{\stackrel{\wedge }{x}}_i^q\right|\le {\epsilon }_i-\Delta x_i^p-\Delta x_i^q$$

(см. фиг.1в), ВО р и q достоверно непосредственно взаимодействуют.Otherwise, when for all $$i=\overline{\mathrm{one..}n}$$

$$\left|{\stackrel{\wedge }{x}}_i^p-{\stackrel{\wedge }{x}}_i^q\right|\le {\epsilon }_i-\Delta x_i^p-\Delta x_i^q$$

(see figv), IN p and q reliably directly interact.

In addition, in contrast to the prototype, two types of indirect interaction are introduced, namely:

- possible indirect interaction;

- reliable indirect interaction.

A possible indirect interaction between a pair of VO p and q exists when there is at least one path between the vertices of the interaction graph corresponding to the objects p and q, on any path between these vertices there is at least one edge representing the possible direct interaction of VO pairs, each the path between these peaks includes more than one edge.

A reliable indirect interaction between the vertices of the interaction graph corresponding to the objects p and q exists when there is at least one path between these vertices of the graph, all edges of which represent reliable direct interaction of the VO pairs, and each such path includes more than one edge.

In contrast to the prototype, two types of VO groups are distinguished: possible groups and reliable groups.

A possible group is a set of VOs in which there is at least one pair of VOs that are in relation to a possible direct interaction, but are not in a state of reliable indirect interaction.

A trustworthy group is a set of VOs in which any pair of objects is in relation to reliable direct or reliable indirect interaction.

To identify reliable and possible groups of HE, in contrast to the prototype, a modified algorithm for traversing the graph in depth is proposed, the scheme of which is presented in figure 2.

At the input of the algorithm, a list of BOs with their serial numbers and interval values of the CS is fed.

The first VO from the input list is assigned to the current VO. If the current VO is not included in any reliable group, then the recursive procedure of forming a new reliable group is started, the scheme of which is shown in Fig. 3. The current VO is the first to be included in this group. Upon completion of the group formation procedure, or if the current VO has already been included in any group, it is checked whether the current VO is the last in the input list. If so, then the operation of the algorithm ends. Otherwise, the next VO from the input list is assigned to the current VO, and the described procedure is repeated for it.

At the input of the recursive procedure for the formation of the group of VO, presented in figure 3, the number of the current generated reliable group and the current ungrouped VO.

During the implementation of the recursive group formation procedure, HE pairs are checked for direct interaction. In this case, reliable groups are formed, which, when the conditions of direct possible interaction are met for pairs of HE from different reliable groups are combined into possible groups of HE. If it is found that for a pair of HEs that were previously included in one possible group, the condition of direct reliable interaction is fulfilled, then the possible group of HEs is converted to reliable.

, азимута $${\stackrel{\wedge }{x}}_2^p$$

, угла места $${\stackrel{\wedge }{x}}_3^p$$

и радиальной скорости ВО $${\stackrel{\wedge }{x}}_4^p$$

соответственно. БИОД, БИОА, БИОУМ, БИОРС с учетом полученных оценок КС ВО и их известных СКО формируют интервальные оценки дальности до ВО $$x_1^p$$

, азимута ВО $$x_2^p$$

, угла места ВО $$x_3^p$$

и радиальной скорости ВО $$x_4^p$$

по формулам (2) соответственно. Интервальные оценки КС ВО поступают в блок формирования списка ВО (БСВО), который составляет список ВО, присваивает им уникальные номера и передает список блоку выявления групп ВО (БВГ). БВГ выполняет процедуру выявления достоверных и возможных групп ВО и приписывает ВО номера достоверных и возможных групп, в которые они входят. БВГ передает информацию о выявленных группах ВО в блок сопровождения групп ВО (БСГ). БСГ формирует трассы групп ВО и передает информацию о них в индикатор РЛС. Индикатор выполняет отображение информации о трассах групп ВО для оператора РЛС.To implement the proposed method for the automated identification of compact groups of interacting HEs, taking into account the uncertainty of their CS values, a typical radar is used [11], a generalized functional diagram of which is shown in Fig. 4. The radar includes a range meter (ID), an azimuth meter (IA), a position angle meter (IUM) and a radial velocity meter (IRS), which measure inclined ranges to HE, their azimuths, elevation angles and radial velocities, respectively. Compared to a typical radar, the following units were added: a unit for generating an interval estimate of the oblique range to AO (BIOD), a unit for forming an interval estimation of an azimuth of VO (BIOA), a block for forming an interval estimation of an elevation angle of a VO (BIOUM), and a block for generating an interval estimation of the radial velocity of a VO (BIORS) to which the measured values of the range to VO come from ID, IA, IUM and IMS $${\stackrel{\wedge }{x}}_{\mathrm{one}}^p$$

azimuth $${\stackrel{\wedge }{x}}_2^p$$

elevation $${\stackrel{\wedge }{x}}_3^p$$

and radial velocity VO $${\stackrel{\wedge }{x}}_{\mathrm{four}}^p$$

respectively. BIOD, BIOA, BIOUM, BIORS, taking into account the obtained estimates of the SC of the HE and their known standard deviations, form interval estimates of the range to the HE $$x_{\mathrm{one}}^p$$

azimuth $$x_2^p$$

, elevation angle $$x_3^p$$

and radial velocity VO $$x_{\mathrm{four}}^p$$

by formulas (2), respectively. Interval assessments of the CS of the HE come into the block of forming the list of HE (BSW), which compiles the list of HE, assigns them unique numbers and transmits the list to the block identifying groups of HE (BHG). BVG performs the procedure for identifying reliable and possible groups of HE and assigns the numbers of reliable and possible groups to which they belong. BHG transmits information on identified VO groups to the VO group escort unit (BSG). The BSG forms the tracks of the VO groups and transmits information about them to the radar indicator. The indicator displays information about the tracks of the VO groups for the radar operator.

In the course of computer simulation, the possibility of using the claimed method with the support of VO groups was investigated. In particular, accompaniment of VO groups in the process of their merger and separation was modeled. The results of computer simulation of the claimed method are presented in figure 5 and figure 6. Identified reliable and possible groups are circled by ellipses. The distances along the coordinate axes in Fig. 5 and Fig. 6 are indicated in conventional values.

Figure 5 shows an example of the identification and tracking of two groups of VO in the process of their merger into one group. In this case, a single group of VOs, also considered a reliable group, joined the reliable group of three VOs. In Fig. 5, reliable VO groups are indicated by codes of the form "D_k.i", where k is the airspace survey number on which this group is identified, and i is the unique number among the reliable groups assigned to it during tracking. Possible groups are indicated by codes of the form “B_k.i”, where k is the airspace survey number on which this group is identified, and i is the unique number among possible groups assigned to it during tracking.

Figure 5 shows that in the process of tracking four VOs according to reviews 1-7, a reliable group 1 consisting of one VO (codes of group D_1.1-D_7.1) was identified and accompanied, and a reliable group 2 consisting of three VOs (group codes D_1.2-D_7.2). According to reviews 5-7, a merger of reliable groups 1 and 2 into a new possible group with number 1 was found (group codes B_5.1-B_7.1). In the process of further rapprochement of HE from the reliable groups 1 and 2 (possible group 1) according to the review 8, their merger into a new reliable group D_8.3 was revealed. Subsequently, this group was accompanied as one reliable group (group codes D_9.3-D_20.3), and reliable groups 1 and 2 were taken out of escort.

The above example shows that tracking VO groups using the claimed method for identifying them makes it possible to detect the association of VO groups first by a possible criterion, and then by a reliable one. Due to this, the possibility of earlier than the prototype detection of the association of VO into groups is achieved.

Figure 6 presents an example of tracking a group of VOs in the process of separating one VO from a group. The designations of the BO groups are similar to those used in FIG.

As shown in Fig.6, according to reviews 1-13, a reliable group of BO 1 (codes of group D_1.1-D_13.1), consisting of four BOs, was identified and accompanied. However, in the process of tracking one of the objects gradually moved away from the other three, as a result of which review 14 revealed the separation of reliable group 1 by reliable criterion. New reliable groups 2 (1 VO) and 3 (3 VO) (codes D_14.2, D_14.3) were identified, which formed a possible group 1 (code B_14.1). Reliable group 1 was removed from the escort, however, all the objects included in it continued to be accompanied as a possible group 1.

During VO tracking, at reviews 14-17, there was a further discrepancy between the reliable groups 2 and 3 (codes of groups D_14.2-D_17.2 and D_14.3-D_17.3, respectively), while the tracking of possible group 1 (codes of group B_14 continued). 1-B_17.1). At review 18, the collapse of possible group 1 was discovered and it was removed from escort. Subsequently, according to reviews 18-20, reliable groups 2 and 3 were identified and accompanied (codes of groups D_18.2-D_20.2 and D_18.3-D_20.3, respectively).

The fact that after separation of the HE group according to reliable criteria, it continues to be accompanied as a possible group for some time, allows us to prevent premature withdrawal of groups from accompaniment as a result of a short-term discrepancy in HE or accidental changes in their SC estimates.

Since from the point of view of tactics the groups of interacting HE represent the greatest danger, the possibility of increasing the time of their accompaniment due to earlier detection and prevention of premature removal from escort is a significant advantage of the claimed method compared to the prototype.

In addition, the claimed method, in contrast to the prototype, allows in the process of grouping HE to take into account explicitly the uncertainty of the estimates of the CS of HE and to assess the degree of reliability of the objects belonging to the selected groups. This makes it possible to reduce the burden on radar operators by automating airborne grouping, reduce the amount of information about the air situation displayed for radar operators and process them in the radar, and provide operators and decision-makers with information about the degree of reliability of grouping results, taking into account the uncertainty of the airborne CS.

Literature

1. Kanaschenkov A.I., Merkulov V.I., Samarin O.F. The appearance of promising airborne radar systems. Features and limitations. - M .: IPRZhR, 2002.

2. Farina A., Studer F. Digital processing of radar information. - M .: Radio and communication. 1993.

3. Willow B.C., Merkulov V.I., Drogalin V.V. and others. Estimation of range and speed in radar systems. Part 3. / Under. ed. B.C. Willows and V.I. Merkulova - Moscow: Radio Engineering, 2010 .-- 427 p.

4. Goroshchenko LB Methods for coordinated guidance and attacks by several fighters of a group of enemy aircraft. // Flight. - 2000, No. 6. S.32-36.

5. Kirsanov A.P., Sorvenkov O.S., Suzansky D.N. Methods for processing radar information accompanied by compact groups of airborne objects // Radio Engineering. - 1996, No. 10. S.102-108.

6. Tatt W. Graph Theory: Per. from English - M .: Mir, 1988 .-- 424 p.

7. Fedosov EA Russian Air Force aviation and scientific and technological progress. Combat systems and systems yesterday, today, tomorrow. - M.: Bustard, 2005.

8. Aho A., Hopcroft J., Ulman J. Construction and analysis of computational algorithms: Per. from English - M .: Mir, 1979.- 536 p.

9. Alefeld G., Herzberger Yu. Introduction to interval computing: Per. from English - M.: Mir, 1987 .-- 360 p.

10. Narignani A.S. Underdetermination in the system of representation and processing of knowledge // Izv. USSR Academy of Sciences. Tech. cybernetics, 1986. No. 5. - S. 3-28.

11. Aviation radar systems and systems: a textbook for students and cadets of the Air Force / PI. Dudnik, G.S. Kondratenkov, B.G. Tatarsky, A.R. Ilchuk, A.A. Gerasimov. Ed. P.I. Angelica. - M.: Publishing. VVIA them. prof. NOT. Zhukovsky, 2006 .-- 1112 p.

Claims (2)

1. A method for automatically detecting compact groups of interacting airborne objects (AT), which includes dividing the observed HE into groups according to the totality of the signs of HE interaction based on the received radar information about the distance between the HE in space, the possibility of visual or radar contact between them, providing for the presentation of the set of VO in the form of a graph whose vertices correspond to the observed VO, and the edges indicate the presence of interaction between a pair of VO, different by the fact that the edges of the graph can reflect both reliable direct interaction of VO and possible direct interaction of VO, and detected groups are defined as groups of possible interaction (possible groups) and groups of reliable interaction (reliable groups), values of state coordinates (CS) of VO obtained in a radar station as a result of primary processing of radar data, including the inclined range to the VO in the range meter, azimuth in the azimuth meter, elevation in the elevation meter and rad the total velocity of the HE in the radial velocity meter, are represented by intervals with the lower and upper boundaries of the form

respectively, where:
p - identifier BO, for all

- the current assessment of the i-th COP in p,

- the actual value of the i-th COP,

- error of its measurement,

- triple standard deviation

moreover, if for the pair BO p and q there

such that

then VO p and q do not interact, otherwise if exists

such that

and

then VO p and q possibly directly interact, otherwise, when for all

IN p and q reliably directly interact; here ε _i is the a priori specified maximum allowable difference for the interacting VOs of i-CS, moreover, a set of VOs for which between each pair of corresponding VO vertices in a graph there is a path consisting only of edges of reliable interaction, a possible group is a set of VOs for which there exists a path between each pair of corresponding VO vertices in the graph, and there is at least one pair of vertices, each path between which includes at least one edge of a possible interaction, for ohm VO classification for a possible or reliable group is performed according to the rule of the graph traversal in depth and comprises: an input list compiled VO with their serial numbers and interval values of COP; for each VO from the input list, starting from the first, in the order of increasing numbers, if it was not previously included in any reliable group, a new reliable group is created, it is assigned a unique number and the recursive procedure for generating a reliable group is launched, starting from the current VO consisting in the fact that the formed group is designated as the current group, the current VO is included in the current group, sequentially, starting from the first, all VOs are selected from the input list if the selected VO does not coincide with the current VO and was not previously included in the current reliable group, then if the current VO and the selected VO reliably directly interact, then if the selected VO was not previously included in any reliable group, then the selected VO is included in the current reliable group and the recursive procedure of forming the current reliable group is launched starting from the selected VO, upon completion of this procedure, if the selected VO was the last in the input list, the procedure ends, otherwise the next VO is selected from the input list; if the selected VO was previously included in a reliable group that is different from the current one, then the current reliable group includes a reliable group of the selected VO, if the selected VO was the last in the input list, the procedure ends, otherwise the next VO is selected from the input list; if the current VO and the selected VO possibly directly interact, then a new reliable group is created, a unique number is assigned to it, and the recursive procedure of forming a new reliable group starts, starting with the selected VO, at the end of this procedure, the current and new valid groups are combined into one possible group, if the selected VO was the last in the input list, the procedure ends, otherwise the next VO is selected from the input list; if the current and the selected VO do not interact, then if the selected VO was the last in the input list, the procedure ends; otherwise, the next VO is selected from the input list. 2. A method for automatically detecting compact groups of VOs according to claim 1, characterized in that the radar further comprises a unit for generating an interval estimate of the range to the VO, a block for generating an interval estimate of the azimuth of VO, a block for generating an interval estimate of the elevation angle of the VO, a block for generating an interval estimation of the radial velocity of VO block for forming a list of HE, block for identifying groups of HE and block tracking groups of HE.

Images