Radar track classification system into false tracks (angels), corresponding to radar angels, and other tracks corresponding to targets of interest such as aircraft or missiles.

Abstract

Radar track classification system, into false angel tracks, corresponding to radar angels, and other tracks, corresponding to targets of interest, comprising three stages: - a first stage (1) of source classification, receiving as input a track update, configured to determine if a track has a secondary radar source and output a classification as other track if an updated track has a secondary radar source, or transmit the track update to a second stage (2); - the second stage (2) of detection and characterization of an angel phenomenon, configured to detect an angel phenomenon and output a status of detection or not of an angel phenomenon, to a first status test module (3), and, in case of detection of an angel phenomenon, output a characterization of the detected angel phenomenon, to a third stage (4);- the first status test module (3), receiving as input the track update from the first stage and the angel phenomenon detection status of the second stage (2), configured to test the angel phenomenon detection status by the second stage (2), and, if the angel phenomenon is not detected, output a classification of the updated track as another track, and if the angel phenomenon is detected, transmit the track update to the third stage (4); and - the third stage (4), configured to classify the track update as a false angel track or another track, using two pre-trained classifiers arranged in series, each employing a random forest algorithm. Figure for the abbreviation: [Fig.2];

Description

Title of the invention: System for classifying radar tracks into false angel tracks, corresponding to radar angels, and other tracks corresponding to targets of interest such as aircraft or missiles.

[0001] The present invention relates to a system for classifying radar tracks into false track angles, corresponding to radar angels, and other tracks, corresponding to targets of interest such as aircraft or missiles.

[0002] The invention relates to radar angels, which are radar echoes that cannot be attributed to reflections on targets of interest, such as aircraft in the field of aerial surveillance, but can be attributed to atmospheric phenomena such as the variation of the refractive index of the air or the presence of particles suspended in the air. When the atmospheric conditions are met, the radar angels constitute sufficiently coherent sets for the radars to initialize and confirm tracks. By extension, these false tracks, that is to say tracks that are not on targets of interest, are called false angel tracks. These false angel tracks are added to the tracks, are propagated throughout the surveillance chain and disrupt the aerial surveillance mission.

[0003] An aim of the invention is to prevent these false trails from appearing in the situation presented to the operator or from being identified as such.

[0004] Existing solutions are often based on minimum speed criteria for track confirmation because false angel tracks often have much lower speeds than those of aircraft. Increasing the minimum speed threshold has the disadvantage of also suppressing tracks on slow-moving drones, a particularly interesting type of target. Increasing the minimum speed threshold is sometimes limited to a restricted area in distance around the radar and in an altitude range. However, the area in which false angel tracks are observed extends over large distances and very wide altitude ranges. In practice, this is inapplicable because the area in which drone detection is affected is too large. A minimum radial speed threshold is sometimes defined per angular sector.In practice, this is ineffective and very difficult to adjust because false angel tracks rarely have a radial trajectory relative to the radar. Finally, these palliatives are adjusted manually, but since the presence and characteristics of the angel tracks depend on atmospheric conditions, they require frequent adjustments.

[0005] When atmospheric conditions are met, in particular when the Angel phenomena originate from particles carried by the wind, radar angels constitute sufficiently coherent sets for radars to initialize and confirm tracks. These tracks generally have common characteristics. First, they move at relatively low speeds (less than 50 ms-1) and are similar in terms of heading and speed. They are distributed on a disk centered on the radar and when observed over a long period, their straight trajectories appear to have been traced by the passage of a toothless comb through the sand.

[0006] [Fi.l] schematically represents, in thicker lines, representative aircraft tracks, and in thinner lines, radar angles.

[0007] The phenomenon of the appearance of radar angels is not permanent because it depends on atmospheric conditions. The order of magnitude of the duration of the phenomenon goes from a few tens of minutes to a few hours. When it is in progress, the stationarity of the phenomenon exceeds ten minutes.

[0008] These false angel leads are added to the other leads, are propagated throughout the surveillance chain and disrupt the surveillance mission.

[0009] It is known to do signal processing and computer processing to process angels. The techniques used are: - Elimination by following only moving echoes; - Filtering echoes using their Doppler velocity; - Correlation with secondary surveillance radars; - The constant false alarm rate; - Counting tracks with predefined characteristics of angel phenomena and deleting these tracks when their number exceeds a threshold in a given sector.

[0010] These types of known solutions have the following limitations respectively: - Elimination by following only moving echoes involves determining a minimum threshold from which movement is considered to exist. The massive emergence of slow targets (such as slow drones) no longer allows thresholds to be applied that eliminate angels without also eliminating drones. The threshold definition is at best manual, at worst a factory setting that cannot be adjusted by a radar operator, which compromises timely adaptation to current conditions. - Filtering echoes using their Doppler velocity also requires setting a minimum threshold that applies to the radial velocity relative to the radar, i.e. the speed of an object measured in the direction of the beam (or line of sight) towards or from the radar. Applying this threshold also eliminates true slow targets. In addition, the distribution of the angels all around radar with non-radial trajectories to the radar compromises the setting of a relevant threshold. The threshold definition is at best manual, at worst a factory setting not adjustable by a radar operator, which compromises timely adaptation to current conditions. - Correlation with secondary surveillance radars is not possible to distinguish between angels and non-cooperative targets, which by definition do not respond to secondary radars. - The constant false alarm rate method is certainly adaptive, but it leads to a hardening of detection confirmation and therefore affects the detection of targets with a low radar equivalent surface, which includes drones. - Counting tracks with predefined "angel" characteristics and deleting these tracks when their number exceeds a threshold in a given sector has two major drawbacks: the presumed characteristics of the angels are assumed to be known, which is poorly suited to evolving phenomena; counting by sector instead of non-sectorized counting is poorly suited to a non-sectorized phenomenon and leads either to delaying the application of the deletion process or to triggering it untimely.

[0011] A primary radar is a radar equipped with a primary antenna that transmits a signal and receives the signal reflected by a target, and a secondary radar is a radar equipped with a secondary antenna that receives an interrogation signal transmitted by a target in response to the reception of a signal transmitted by this antenna. A primary / secondary radar is a radar equipped with a primary antenna and a secondary antenna.

[0012] A track is represented by a set of parameter values ​​representative of the fusion of successive detections of a target by a radar. When a track is maintained by primary detections, this track has a primary source. When a track is maintained by secondary detections, this track has a secondary source. A track can have a primary source and a secondary source.

[0013] An aim of the invention is that these false trails can be identified automatically so as to either eliminate them or allow their filtering on the aerial situation presented to a surveillance operator. This filtering must not be done to the detriment of the true trails on slow targets and with a small radar equivalent surface, and must adapt automatically to the current phenomenon.

[0014] According to one aspect of the invention, there is proposed a system for classifying radar tracks of aircraft-type targets of interest, into false track angles, corresponding to radar angles, and into other tracks, corresponding to targets of interest:

[0015] - a first source classification stage, receiving as input an update of track, configured to determine whether the updated track has a secondary source and output a classification as another track if the updated track has a secondary source, or transmit the track update to a second stage;

[0016] - the second stage of detection and characterization of an angel phenomenon, configured to detect an angel phenomenon and output a status of detection or not of an angel phenomenon, to a first status test module, and, in the event of detection of an angel phenomenon, output a characterization of the detected angel phenomenon, to a third stage;

[0017] - the first status test module, receiving as input the track update of the first stage and the status of detection or not of angel phenomenon of the second stage, configured to test the value of the status of detection or not of an angel phenomenon by the second stage, and, in case of status of not detection of an angel phenomenon, output a classification as another track of the updated track, and in case of status of detection of angel track phenomenon, transmit the track update to the third stage; and

[0018] - the third stage, configured to perform a classification of the updated track as a false angel lead or other lead, by two pre-trained classifiers, arranged in series, each using a random forest type algorithm.

[0019] In one embodiment, the first source classification stage comprises: - a second test module configured to test whether an identifier of the updated track matches a track identifier having a secondary radar source already listed as a track with a secondary radar source, by accessing a database of track histories having a secondary source, to output a classification as another track of the updated track if the identifier matches a track having a secondary radar source in the database of track histories having a secondary radar source, and to transmit the track update to a third test module otherwise; - the third test module configured to test whether the track update comes from a secondary radar source, to store in memory the updated track in the track history database having a secondary radar source if the updated track is considered to come from a secondary radar source and output a classification as another track of the updated track, and to transmit to the second stage the track update otherwise.

[0020] According to one embodiment, the second stage of detection and characterization of an angel phenomenon comprises: - a fourth module for testing whether or not the end of an interval has been reached

[0021]

[0022]

[0023]

[0024]

[0025]

[0026] repeating time configured to, after storing the track update in a current time interval track update history database, if the end of the current time interval is reached to retrieve the contents of the current time interval track update history database, reset the current time interval track update history database, and transmit current time interval track updates to a partitioning module; - the partitioning module configured to perform a partitioning of the track updates of the elapsed time interval, by studying the similarity of attributes of the track update including the heading and the speed of the track, detect the presence of an angel phenomenon, output a status of detection or not of an angel phenomenon, and, in the event of detection of an angel phenomenon, output a characterization of the detected angel phenomenon, at the third stage. In one embodiment, the third stage of classifying the track update, for a time interval, as a false angel track or other track comprises: - a first pre-trained classifier using a random forest type algorithm making it possible to output a current classification of the track, when the current time interval has not elapsed; - a second pre-trained classifier using a random forest type algorithm; - a module for calculating the classification likelihood of the second classifier; and - a rule base for determining the final classification from the likelihood vector, configured to be applied at the output of the likelihood calculation module and to deliver as output a classification of the track as a false angel track or other track, at the end of the current time interval. According to one embodiment, the system includes a track update database configured to train the first and second classifiers of the third stage. There is also provided, according to another aspect of the invention, an aerial surveillance center, equipped with a false trail detection system, as previously described. The invention will be better understood by studying a few embodiments described as non-limiting examples and illustrated by the appended drawings in which: [Fig.l] schematically illustrates an angel phenomenon, according to the state of the art; and [Fig.2] schematically illustrates an angel track detection system comprising an angel phenomenon detection device and a radar track classification device into a false angel track or another track, according to one aspect of the invention;

[0027] [Fig.3] schematically illustrates a first stage of a system of [Fig.2], according to an aspect of the invention;

[0028] [Fig.4] schematically illustrates a second stage of a system of [Fig.2], according to an aspect of the invention;

[0029] [Fig.5] schematically illustrates a third stage of a system of [Fig.2], according to an aspect of the invention; and

[0030] [Fig.6] schematically illustrates a chronology of operation of the system of [Fig.l], according to one aspect of the invention.

[0031] Throughout the figures, elements having identical references are similar.

[0032] [Fig. 2] schematically illustrates an angel track detection system comprising an angel phenomenon detection device and a radar track classification device into a false angel track or another track, according to one aspect of the invention.

[0033] A track is a history of successive updates of the status, according to a set of characteristics, of the tracking of a supposed target. A track is designated by a unique identifier, called a track identifier.

[0034] The false track detection system takes as input a track update and provides as output a classification of "false track" or "other track".

[0035] The radar track classification system, into false track angels, corresponding to radar angels, and other tracks, corresponding to targets of interest, comprises three stages:

[0036] - a first stage 1 of source classification, receiving as input an update of runway, configured to determine whether a runway has a secondary radar source and output a classification as another runway if an updated runway has a secondary radar source, or transmit the runway update to a second stage 2;

[0037] - the second stage 2 of detection and characterization of an angel phenomenon, configured to detect an angel phenomenon and output a status of detection or not of an angel phenomenon, to a first status test module 3, and, in the event of detection of an angel phenomenon, output a characterization of the detected angel phenomenon, to a third stage 4; - the first status test module 3, receiving as input the track update from the first stage 1 and the status of detection or not of an angel phenomenon from the second stage 2, configured to test the value of the status of detection or not of an angel phenomenon by the second stage 2, and, in the event of a status of non-detection of an angel phenomenon, output a classification as another track of the updated track, and in case of angel phenomenon detection status, transmit the track update to the third stage 4; and

[0038] - the third stage 4, configured to perform a classification of the update of track as false angel track or other track, by two pre-trained classifiers, arranged in series, each using a random forest type algorithm.

[0039] The characterization of the detected angel phenomenon includes the parameter values ​​including the direction and speed of the detected angel phenomenon.

[0040] The present invention has the following advantages: - automatic detection of the angel phenomenon; - automatic extraction of the characteristics of the phenomenon; - automatic activation of the pre-trained false angels vs other track classifier; - automatic consideration by the classification stage of the parameters of the phenomenon; - the periodic re-evaluation of the classification of false leads “angels” vs. other leads; - significant filtering of false angel leads; and - a low risk of deleting drone tracks.

[0041] Two phases are to be distinguished, classic to a "Machine Learning" algorithm in English, or Automatic Learning, in the operation of the system:

[0042] - a learning phase which must be carried out before any exploitation but only needs to be executed once by providing a set of labeled track updates (i.e. whose actual classification is known) as input, the different classifiers of the system based on supervised Machine Learning are self-parameterized; and

[0043] - an inference phase which takes track updates as input, and returns a classification indicating whether it is likely a false angel lead or another lead.

[0044] [Fig.3] schematically illustrates the first stage 1 of source classification. The first stage 1 comprises: - a second test module 5 configured to test whether an identifier of the updated track corresponds to a track having a secondary radar source already identified, by access 6 to a database 7 of tracks having a secondary radar source, to output a classification as another track of the updated track if the identifier corresponds to a track identifier having a secondary radar source in the database 7 of track histories having a secondary radar source, and to transmit the track update to a third test module 8 otherwise; - the third test module 8 configured to test whether the track update includes a secondary radar source, to store 9 in memory the updated track in the database 6 of tracks having a secondary radar source if the track update is considered to have a secondary radar source and output a characterization of the false track detected, and to transmit to the second stage 2 the track update otherwise.

[0045] [Fig.4] schematically illustrates the second stage 2 of detection and characterization rization of an angel phenomenon. The second floor 2 includes: - a fourth module 10 for testing whether or not the end of a repetitive AT time interval has been reached, configured to, after storing 11 the track update in a database 12 of the history of track updates of the current AT time interval, if the end of the current AT time interval is reached, extract 13 the content of the database of the history of track updates of the current time interval, reset 13 the database 12 of the history of track updates of the current AT time interval, and transmit track updates of the current AT time interval to a partitioning module 14; - the partitioning module 14 configured to perform a partitioning of the track updates of the elapsed time interval AT, by studying the similarity of attributes of the track update including the heading and the speed of the track, detecting 15 the presence of an angel phenomenon, delivering as output a status of detection or not of an angel phenomenon, and, in the event of detection 15 of an angel phenomenon, delivering as output a characterization of the detected angel phenomenon, to the third stage 4.

[0046] [Fig.5] schematically illustrates the third stage 4 of classification of the update track day, for a time interval At, as a false track angel or other track. The time is divided into successive time intervals At, as illustrated in [Fig.6]. The third stage 4 includes: - a first pre-trained classifier 18 using a random forest type algorithm to output a current classification of the track, when the current time interval At has not elapsed; - a second pre-trained classifier 19 using a random forest type algorithm; - a classification likelihood calculation module 20 of the second classifier 19; and - a rule base 21 for determining the final classification from the likelihood vector, configured to be applied at the output of the module 20 of likelihood calculation and output a classification of the track as a false angel track or other track, at the end of the current time interval AT.

[0047] From the track update and the characterization of the detected angel phenomenon, a module 22 for extracting and calculating characteristics of interest performs a data extraction of the track update and the characterization of the detected angel phenomenon to provide the first classifier 18 with characteristics of the track update, i.e. values ​​of parameters representative of the track update comprising its kinematic parameters and its radar classification.

[0048] The first classifier outputs a classified track update which is stored in memory by a memory storage module 23 in a database 24 of track update history of the segment (i.e. all the updates of a track having taken place during the current time interval AT).

[0049] A fifth test module 25 tests whether the current time interval At has elapsed, and if this is not the case, as illustrated in [Fig.6], without being at the end of the current time interval At, a memory extraction module 6 extracts from a database 27 of track classification histories, a current classification of the track delivered as output.

[0050] If the fifth test module 25 determines that the current time interval At has elapsed, a module 28 for extracting data and resetting memory outputs the data from a database 29 of track update histories of the segment to a module 30 for extracting and calculating characteristics of interest, and reinitializes the latter 29.

[0051] The module 30 for extracting and calculating characteristics of interest also receives as input the characterization of a false track detected by the second stage 2, and delivers as output characteristics of the track segment, i.e. values ​​of parameters representative of the track segment, including the statistics of the kinematic characteristics and the radar classification of the track updates of the segment, intended for the second classifier 19.

[0052] The second classifier 19 delivers as output a classification of the track segment as a false angel track or other track, to the classification likelihood calculation module 20 which also receives as input a memory extraction carried out by a memory extraction module 31 which extracts the data from a likelihood vector history database 32.

[0053] The classification likelihood calculation module 20 calculates, for example, by matrix calculation involving the classification of the second classifier 19, the memory extraction of the last likelihood vector of the track delivered by the module 31 and a matrix of fixed parameters then a likelihood vector, which it delivers as output to a module 33 for storing in memory said likelihood vector in the likelihood vector history database 32.

[0054] From the likelihood vector, the base 21 of determination rules 21 makes it possible to evaluate the values ​​of the likelihood vector with respect to fixed thresholds and to deliver as output a classification of the track among a false radar track or another track, a memory storage module 34 stores in a history database 35. Track classification

[0055] The system according to the invention also comprises a track update database configured to train by learning the first and second classifiers of the third stage 4.

[0056] According to one aspect of the invention, an air surveillance center can be provided with a system for classifying radar tracks into false track angles, corresponding to radar angels, and into other tracks, as described previously.

[0057] The specificity of the present invention, compared to a classic Machine Learning algorithm, is based on several aspects which follow.

[0058] First of all, the operation is adaptive; in fact, the first stage 1 determines whether the track includes updates from a secondary source and, if not, seeks to detect an angel phenomenon by extracting its kinematic characteristics. This detection uses a partitioning or "clustering" algorithm on determined attributes of the track updates. This requires storing the track updates in memory over a time interval AT. The present invention is innovative and better suited than a study by radar sector because the angel phenomenon is more naturally formalizable in a Cartesian or geographic reference frame than in a polar reference frame traditionally used by radar operators. Depending on the response to the secondary source detection and the angel phenomenon detection, the track updates pass or not into the third classification stage 4.This reduces the risk of misclassifying real tracks as angels.

[0059] Furthermore, the structure of the third stage 4 is novel, comprising two classifiers 18, 19 arranged in series, each using a random forest algorithm. The first classifier 18 processes the track updates individually according to determined attributes, and the second classifier 19 processes the tracks as a whole, based on the calculated statistical characteristics, to which the predictions of the first classifier 18 are added. The invention makes it possible both to capture local variations and to take into account the known globality of the track over a time interval. In addition, random forests are much less expensive in terms of resources and execution time than neural networks and are based on a decision tree structure, similar to a human decision-making process, thus providing a gain in explainability (capacity to relate and make understandable possible the elements taken into account by the AI ​​system for the production of a result). The classification is given for a track segment and therefore requires here to keep in memory the track updates over a time interval.

[0060] Furthermore, the invention presents another novel aspect, the classification given to each track and the detection of angel phenomenon is reversible because the track is not systematically deleted when it is classified as a "false angel track" as in certain existing treatments but marked as such. This allows a periodic re-evaluation, according to fixed time intervals. This is also allowed by the speed of execution of the classifiers seen previously). The detection of angel phenomenon is re-evaluated over wide time intervals, making it possible to follow the local evolution of the atmospheric situation in real time, to adapt / inhibit the angel classification if the meteorological phenomenon evolves / interrupts. In practice, a unit time interval At is determined, used for the short interval, the long interval AT corresponding to a multiple of this unit interval as in the example of [Fig.6].

[0061] Finally, another new aspect of the solution is to construct a likelihood vector taking into account the history of past predictions. This vector is specific to each track and indicates for each possible label the associated probability, at a given time. It is therefore updated at each interval, taking into account the values ​​of the previous vector, the new predictions of the classifiers, as well as a transition matrix from one state to another (adapted to each case), by weighting the contribution of the new values.

[0062] Among the main advantages of this solution, we find: - automatic detection of the angel phenomenon, which allows automatic adaptation of treatments to its intermittent nature; - automatic extraction of the characteristics of the phenomenon, essential characteristics for the treatment of angels; - automatic activation of the false track angel vs other track classification stage. The false track angel / other track classification is only activated when necessary; - automatic consideration by the classification stage of the parameters of the phenomenon; - periodic re-evaluation of the classification of false angels vs. other leads, taking into account the history of the lead; - a low risk of removing runways on slow runways with low radar equivalent area. No risk of erroneous angel classification when angel detection is not in progress; a risk that remains low when angel classification is activated; — the characteristics of the angels phenomenon are expressed in a universal reference system, adapted to the physical phenomenon and understandable for an operator; does not require human intervention to determine the treatments adapted to the intermittent phenomenon; the classification stage is based on acquired learning and does not require relearning.

[0063]

Claims

Claims

1. System for classifying radar tracks into false track angles, corresponding to radar angels, and other tracks, corresponding to targets of interest, comprising three stages: - a first source classification stage (1), receiving as input a track update, configured to determine whether a track has a secondary radar source and output a classification as another track if an updated track has a secondary radar source, or transmit the track update to a second stage (2); - the second stage (2) for detecting and characterizing an angel phenomenon, configured to detect an angel phenomenon and output a status of detection or not of an angel phenomenon, to a first status test module (3), and, in the event of detection of an angel phenomenon, output a characterization of the detected angel phenomenon, to a third stage (4); - the first status test module (3), receiving as input the track update from the first stage and the status of detection or not of an angel phenomenon from the second stage (2), configured to test the value of the status of detection or not of an angel phenomenon by the second stage (2), and, in the event of a status of non-detection of an angel phenomenon, output a classification as another track of the updated track, and in the event of a status of detection of an angel phenomenon, transmit the track update to the third stage (4); and - the third stage (4), configured to perform a classification of the updated track as a false angel track or other track, by two pre-trained classifiers, arranged in series, each using a random forest type algorithm.

2. The system of claim 1, wherein the first source classification stage comprises: - a second test module (5) configured to test whether an identifier of the updated runway matches a runway identifier having a secondary radar source already listed as a runway with a secondary radar source, by accessing a database (7) of secondary radar track histories, to output a classification as another runway of the updated runway if the identifier matches a runway identifier having a secondary radar source in the track history database (7) having a secondary radar source, and to transmit the track update to a third test module (8) otherwise; - the third test module (8) configured to test whether the track update comes from a secondary radar source, to store (9) in memory the updated track in the database (6) of track histories having a secondary radar source if the updated track is considered as a track coming from a secondary radar source and output a classification as another track of the updated track, and to transmit to the second stage (2) the track update otherwise.

3. System according to claim 1 or 2, in which the second stage (2) of detection and characterization of an angel phenomenon comprises: - a fourth module (10) for testing whether or not the end of a repetitive time interval has been reached, configured to, after storing (11) the track update in a database (12) of the history of track updates of the current time interval, if the end of the current time interval is reached, extract (13) the content of the database of the history of track updates of the current time interval, reset (13) the database (12) of the history of track updates of the current time interval, and transmit track updates of the current time interval to a partitioning module (14); - the partitioning module (14) configured to perform a partitioning of the track updates of the elapsed time interval, by studying the similarity of attributes of the track update including the heading and the speed of the track, detect (15) the presence of an angel phenomenon, deliver as output a status of detection or not of an angel phenomenon, and, in the event of detection (15) of an angel phenomenon, deliver as output a characterization of the detected angel phenomenon, to the third stage (4).

4. System according to one of the preceding claims, in which the third stage (4) of classification of the track update, for a time interval (AT), as a false angel track or other track comprises: - a first pre-trained classifier (18) using a random forest type algorithm making it possible to output a current classification of the track, when the current time interval (AT) has not elapsed; - a second pre-trained classifier (19) using a random forest type algorithm; - a classification likelihood calculation module (20) of the second classifier (19);and - a rule base (21) for determining the final classification from the likelihood vector, configured to be applied at the output of the likelihood calculation module (20) and to deliver at the output a classification of the track as a false track or other track, at the end of the current time interval (AT).;

5. System according to one of the preceding claims, comprising a database (24) of track updates configured to train by learning the first and second classifiers of the third stage (4).

6. Air surveillance center, provided with a system for classifying radar tracks into false track angles, corresponding to radar angels, and into other tracks, according to one of the preceding claims.

Images