RADAR SIGNAL DISENTING METHOD

专利摘要:
The present invention relates to a method for de-interlacing radar signals, the method comprising: - receiving electromagnetic signals from a receiver (12) and extracting pulses from the received signals, and - forming pulse trains comprising at least three pulses spaced from the same pulse repetition interval, each pulse train being defined by the repetition interval of the pulses. The method further comprises: - grouping the pulse trains having the same pulse repetition interval according to a predefined clustering law to form pulse levels, and - the association of the pulse levels according to at least one predefined association law for obtaining deinterleaved radar signals formed by the concatenation of the pulse trains of the associated pulse stages.
公开号:FR3031257A1
申请号:FR1403062
申请日:2014-12-31
公开日:2016-07-01
发明作者:Daniel Stofer；Jean Francois Grandin；Jean Marie Lemoine
申请人:Thales SA；
IPC主号:

专利说明:

[0001] The present invention relates to a method for deinterleaving radar signals, the method comprising: receiving electromagnetic signals from a receiver and extracting the pulses from the received signals; and training pulse trains comprising at least three pulses spaced from the same pulse repetition interval, each pulse train being defined by the repetition interval of the pulses. The present invention also relates to an associated deinterleaving device. One of the challenges of electronic warfare is to intercept radio emissions from detection systems such as radar transmitters. The presence of many transmitters causes the intercepted signals to be interleaved, i.e. the signals emitted by a radar transmitter of interest are scrambled by other signals emanating from the ambient electromagnetic environment. It is therefore necessary to deinterlace the intercepted signals to separate the different signals emitted by the different radar transmitters. However, signals emitted by the same radar transmitter can have characteristics, defining a waveform, which are variable, in particular in terms of the repetition interval of the signal pulses or the carrier frequency of the signal pulses, which makes the process complex deinterlacing. The great wealth of waveforms of the electromagnetic world corresponds to a wide variety of deinterlacing processes for extracting pulses of the same waveform from the ambient electromagnetic environment. More particularly, the technical field, object of this method, relates to the deinterlacing of the waveforms whose pulse repetition interval is medium or short (up to a few hundred microseconds). These waveforms generally consist of several pulse trains. It is known to use radar signal extractors implementing a process of deinterlacing radar signals in two stages. The first step is the formation of pulse trains from all the impulses intercepted. The second step is to group the trains of pulses formed to obtain deinterleaved radar signals. The first pulse train formation step exploits the statistical information of the intercepted signals such as the carrier frequencies of the signal pulses, the pulse repetition intervals and the arrival directions of the pulses. The second stage of grouping the pulse trains groups the trains of pulses formed according to their proximity to form de-interlaced signals.
[0002] Nevertheless, the existing pulse train grouping algorithms do not offer the same level of maturity as those for training pulse trains. In particular, the step of grouping the pulse trains is generally approached as a problem of "clustering" or "classification of data", where each train is compared to another according to a single distance criterion. However, the waveforms formed by the pulse trains are of great diversity, some families of waveforms may have characteristics totally antagonistic to each other. Therefore, a single distance criterion can lead to the reconstruction of erroneous signals. The technical problem relates to the grouping of pulse trains originating from the same radar signal in a dense electromagnetic environment in which several distinct waveforms can appear simultaneously, the difficulty of not performing poor pulse train groupings. . There is therefore a need for a method of deinterleaving radar signals for grouping pulse trains with better reliability, by limiting the risks of obtaining an erroneous deinterleaved signal, while being of rapid implementation. For this purpose, the subject of the invention is a method for de-interlacing radar signals of the aforementioned type, in which the method furthermore comprises: the grouping of the pulse trains having the same repetition interval 25 of the pulses according to a law predefined grouping for forming pulse stages, and - the association of the pulse stages according to at least one predefined association law to obtain deinterleaved radar signals formed by the concatenation of the pulse-level pulse trains. associates.
[0003] According to particular embodiments, the deinterleaving method comprises one or more of the following characteristics, taken individually or in any technically possible combination: each pulse train is also defined by at least one element selected from a group consisting of: the arrival time of the first pulse of the pulse train, the arrival time of the last pulse of the pulse train, the pulse frequency of the pulse train, the pulse duration of the pulse train, pulse train and the direction of arrival of the pulses of the pulse train. the method comprises, before the grouping step, a step of classifying the pulse trains according to their carrier frequency so as to obtain two classes of pulse trains: a class grouping the fixed carrier frequency pulse trains; and another class grouping the pulse trains of variable carrier frequency, the grouping step being implemented for each of the two classes of pulse trains and making it possible to obtain single-frequency pulse levels from the class single-frequency pulse trains and frequency-agile pulse bearings from the class of frequency-agile pulse trains. the association step comprises a phase of grouping the pulse steps of repetition intervals of the different pulses and which are linked in time to obtain groups of switching pulse stages. the association step comprises a grouping phase of fixed carrier frequency pulse stages, having repetition intervals of identical pulses and superimposed in time to obtain overlapping pulse bearing groups. each of the grouping and association laws is implemented by at least one algorithm making it possible to obtain groups from elements, the elements designating pulse trains during the grouping and landing step of pulses in the association step, the groups designating pulse steps in the grouping step and the groups of pulse stages in the associating step, the algorithm comprising: choosing a reference element from among a set of elements, deleting the reference element from the set of elements and adding, in a set of groups, a reference group comprising the reference element, o the selection in the set of elements of elements compatible with the reference group according to a set of criteria to obtain a set of candidate elements, o the evaluation of the distance between the group reference and each elem nt of the set of candidate elements, o the annexation of the element of the set of candidate elements minimizing a distance to the reference group and the deletion of the annexed element of the set of elements, O the repetition of the selection, evaluation and annexation phases as long as the set of candidate elements contains elements, and o the repetition of all the preceding phases as long as the set of elements includes elements. 5 - the reference element is the element of the set of elements whose arrival time of the first pulse is the smallest. the set of criteria evaluates the compatibility of the elements of the set of elements with the reference group according to one or more characteristics, the characteristics being chosen from a group comprising: the direction of arrival of the elements , the time superposition of the elements, the carrier frequency of the elements, the pulse width of the elements, the repetition interval of the element pulses, the phase of the elements and the number of pulses of the elements. the criteria are chosen from statistics on the characteristics of the radar waveforms of a database. For the grouping law, the distance is the difference in time between the last pulse of the reference group of the first pulse of the candidate element of the set of candidate elements, and for the law of association, the distance is a recovery rate between the reference group and the candidate element of the set of candidate elements or a score making it possible to select from the set of candidate elements the element sharing the most common characteristics with the reference group. the method comprises, before the grouping step, a step of rejecting the incoherent pulse trains in terms of pulse repetition interval. the method comprises, before the grouping step, a step of rejection of the pulse trains whose pulse repetition interval is greater than a pulse repetition interval threshold and whose number of pulses is less than at a threshold of number of pulses. The invention also relates to a device for de-interlacing radar signals comprising: a receiver capable of receiving electromagnetic signals; a digital signal processing unit able to extract the pulses from the signals received by the receiver; and a readable medium. The computer program is stored on a data processing unit and adapted to drive a process according to the invention. any of the preceding claims when the computer program is implemented on the data processing unit. Other features and advantages of the invention will appear on reading the following description of embodiments of the invention, given by way of example only and with reference to the drawings which are: FIG. schematic of a device for deinterlacing radar signals enabling the implementation of a deinterlacing process according to the invention; FIG. 2, a flowchart of an exemplary implementation of a deinterleaving method according to FIG. 3, a schematic representation of three pulse trains, FIG. 4, a schematic representation of a pulse train and the quantities defining this pulse train, FIG. 5, a diagrammatic representation of a pulse train, FIG. signal and a pulse stage 15 resulting from the grouping of three pulse trains of this signal, FIG. 6, a schematic representation illustrating the grouping of two pulse trains during a FIG. 7 is a flow diagram of the operation of an algorithm implemented during the deinterleaving process according to the invention, FIG. 8, a diagrammatic representation of a phase for eliminating elements incompatible with a reference group, FIG. 9, a schematic representation of a meeting phase of a component that is compatible with a reference group when the distance between the element and the reference group is FIG. 10 is a diagrammatic representation of a group of pulse bearings formed from pulse stages of a FMICW waveform, FIG. 11, a diagrammatic representation. illustrating the combination of three pulse stages to form groups of pulse stages in another association step of the de-interleaving method according to the invention, FIG. 12, a schematic representation illustrating the association of eight pulse levels to form two groups of pulse levels during a combination step of the deinterlacing method according to the invention, and FIG. 13; a schematic representation of the process of deinterleaving radar signals from receiving signals from a receiver to obtaining de-interlaced signals. A device 10 for deinterleaving radar signals is shown in FIG. 1. The device 10 for de-interlacing radar signals is adapted to implement a method of deinterleaving radar signals. The input data of the deinterleaver 10 is a measure of the characteristics of the pulses received by the device 10. The characteristics are, for example, the pulse carrier frequency, the pulse width, the pulse power, the direction of the pulses. arrival of impulses or periods of repetition of impulses. Such characteristic pulse measurements are derived, for example: from the reception of radar signals by a receiver, then from the digitization of the signals and the extraction of the pulses of the signals by a digital signal processing unit the reception of analogue signals created in the laboratory simulating the reception of radar signals, then the digitization of the signals and the extraction of the pulses of the signals by a digital signal processing unit of the generation, via software, digital data simulating the reception and digitization of radar signals, and the extraction of the signal pulses by a digital signal processing unit, or the generation, via software, of pulses which are then recorded on an interface reception.
[0004] As illustrated in FIG. 1, the device 10 comprises an electromagnetic wave receiver 12, a computer 14 and a readable information medium 16 interacting with the computer 14. The receiver 12 is able to receive electromagnetic signals emanating from, for example, radio detection systems such as radars. The electromagnetic signals are, for example, derived from radar transmitters or are analog signals created in the laboratory and simulating radar signals. The receiver 12 is connected to the computer 14 and is able to send the signals received by this receiver 12 to the computer 14. The receiver 12 is, for example, an antenna. For example, the receiver 12 is an elementary antenna, a network antenna, a reflector antenna, a circularly polarized antenna, a waveguide antenna, an active antenna, a shortened antenna, a broadband antenna, patch antenna, loop antenna or loop antenna, or system consisting of one or more of the preceding antennas.
[0005] The computer 14 is adapted to receive signals from, for example, the receiver 12 or digital data from simulation software. The computer 14 is a computer having a processor 18 and, optionally, a man-machine interface 20 and a display unit 22. The computer 14 further comprises a digital signal processing unit 23 and optionally an interface 32. The processor 18 comprises a data processing unit 24, memories 26 and an information carrier reader 28. The information carrier reader 28 is adapted to receive and read the readable information medium 16. The information carrier reader 28 is connected to the data processing unit 24. The readable medium of information 16 is a support readable by the information carrier reader 28. The readable information medium 16 is a medium adapted to store electronic instructions and capable of being coupled to a bus of a computer system. By way of example, the readable information medium 16 is a floppy disk (floppy disk), an optical disk, a CD-ROM, a magneto-optical disk, a ROM memory, a memory RAM, an Erasable Programmable Read-Only Memory (EPROM), an EEPROM (Electrically-Erasable Programmable ReadOnly Memory), a magnetic card or an optical card. On the readable information medium 16 is stored a computer program product including program instructions. The computer program is loadable on the data processing unit 24 and is adapted to cause the implementation of a method of deinterleaving radar signals according to the invention. The human-machine interface 20 is, for example, a keyboard. The display unit 22 is, for example, a screen.
[0006] The digital signal processing unit 23 is configured to digitally process the signals received by the computer 14. More specifically, the digital signal processing unit 23 is configured to digitize the received signals, extract the pulses from the signals, and measure the characteristics of each pulse extracted.
[0007] In a variant, the digital signal processing unit 23 is also configured to generate digital data simulating the reception and digitization of radar signals or to generate pulses directly. The interface 32 makes it possible, on the one hand, to store pulses resulting from the digital signal processing carried out by the digital signal processing unit 23 and, on the other hand, to receive previously memorized pulses or to receive pulses generated by simulation software. The operation of the radar signal deinterleaving device is now described with reference to FIG. 2 which is a flowchart of an exemplary implementation of a method of deinterleaving radar signals according to the invention. The deinterleaving method comprises a step of receiving 100 of a plurality of electromagnetic signals by the receiver 12, digitizing the received signals and extracting the electromagnetic pulses from the digitized signals by digital signal processing. The digitization of the signals and the extraction of the pulses I ,, ..., lm are carried out by the digital signal processing unit 23 and are optionally recorded in the interface 32. Such pulses I ,, ... Im electromagnetic emanate or are representative, in particular, of signals emitted by radio detection systems such as radars. In a variant, the pulses 1 'are already recorded in the interface 32.
[0008] The steps of the deinterleaving method described hereinafter are carried out by the computer 14 interacting with the readable information medium 16. The deinterleaving method then comprises a step 110 of forming pulse trains T1, Tn from the pulses lm received by the receiver 12 or recorded in the interface 32.
[0009] During the forming step 110, at least three pulses 12, 13 are grouped together to form a pulse train T. The criterion for forming the T1, Tri pulse trains consists of grouping together in a single pulse train TX, the pulses lm having the same repetition interval of the PRI pulses. It is understood by the term "pulse repetition interval" the period separating two consecutive pulses. In other words, the pulses..., I, grouped together in the same pulse train TX are such that any two consecutive pulses of the pulse train T are spaced apart by the same repetition interval of the pulses PRI as two other any consecutive pulses of the pulse train T. The pulse trains T1,..., T, formed constitute an AND set of pulse trains T1, Tn. The trains of pulses T1, Tn formed are defined by the interval 3031257 9 of repetition of the pulses PRI separating two consecutive pulses of the train of pulses T1, ..., Tn. 3, three pulse trains T1, T2 and T3 having pulse repetition intervals PRI, PRI2 and PRIS, possibly different from each other, are illustrated in FIG. As illustrated in FIG. 4, a pulse train Tx is also defined by the arrival time of the first pulse of the pulse train Tx, called TOAdéb, and the arrival time of the last pulse 1m of the train of In addition, each pulse train Tx is optionally defined by at least one element selected from a group consisting of: the pulse frequency ..., lm of the pulse train Tx, the duration of the pulses Tx, pulses I ,, ..., lm of the pulse train Tx and the direction of arrival of the pulses ..., lm of the pulse train Tx. The deinterleaving method then comprises a step 120 of rejection of pulse trains T1, ..., T, which are incoherent.
[0010] The rejection step 120 checks the coherence in the repetition interval of the pulses PRI of the pulse trains T1,..., T, formed during the formation step 110. Indeed, certain pulse trains Tx can have been formed by mixing at least two repetitive intervals of the near but still different PRI pulses. An incoherent pulse train Tx in terms of the repetition interval of the PRI pulses is, for example, detected by means of a statistical test based on the intervals separating all the pulses of the train Tx. For example, a Chi-square test may be used. The pulse trains Tx judged to be incoherent in terms of the repetition interval of the pulses PRI are eliminated from the set ET of pulse trains T1, - - -, Tn. During this rejection step 120, optionally, the coherence of the pulse trains T1,..., Tri in terms of the carrier frequency of the pulses is also verified. It is understood by the term "frequency carrying a pulse", the frequency of the carrier of the pulse, the carrier being a wave modulated by an input signal. Indeed, some pulse trains Tx may have been formed by the mixture of at least two near but different carrier frequencies. An incoherent pulse train Tx in terms of carrier frequency is, for example, detected by means of a statistical test based on the frequencies of the pulses of the train Tx. For example, a Chi-square test may be used. Tx pulse trains deemed incoherent in terms of pulse carrier frequency are eliminated from the set ET of pulse trains T1, Tn.
[0011] In the case where the repetition intervals of the pulses PRI, respectively the carrier frequencies of the pulses, are modeled as Gaussian variables and independent of each other, the coherence in repetition interval of the pulses PRI, respectively in carrier frequency of the pulses, is evaluated by a statistical test 5 of the chi-square. It is understood by the expression "chi-square test", abbreviated as "x2 test" or "chi2 test", a statistical test to test the adequacy of a series of data to a family of laws probabilities or to test the independence between two random variables. The deinterleaving method then comprises a step 130 for rejecting the 10 Tx pulse trains that do not belong to the category of high or medium frequency waveforms of recurrence. A waveform gathers T1, Tn pulse trains having common characteristics in terms, in particular, of repetition interval of the PRI pulses, of the pulse carrier frequency and of the number of pulses. For example, all pulse trains T1,..., Tri having short pulse repetition intervals PRI and a large number of pulses constitutes a high or medium frequency recurrence waveform. Conversely, the set ET pulse trains T1,..., T ,, having high pulse repetition intervals PRI and a low number of pulses constitutes another low frequency type waveform. of recurrence.
[0012] The rejection step 130 includes defining characteristic thresholds of a waveform: a pulse repetition interval threshold SpRi and a pulse number threshold Pulse. The pulse repetition interval threshold SpRi and the threshold number of Simulation pulses are defined from a database representative of a number of radar waveforms of interest. In an alternative embodiment, this waveform database is stored in the memory 26 of the processor 18. In another variant embodiment, only the thresholds determined from the database are stored in the memory 26 of the computer. processor 18. The threshold SpRi is chosen to be greater than the values predominantly observed in the waveform database. The pulse repetition interval threshold SpRI is, for example, in the broad sense between 1 microseconds (ps) and 1 milliseconds (ms) to test the membership of high or medium frequency recurrence waveforms. The pulse threshold is chosen to be lower than the values mostly observed in the waveform database. For example, the pulse number threshold is understood in the broad sense between 1 pulse and 100 pulses to test the membership of the high or medium frequency recurrence waveforms. In this case, such thresholds of SARI pulse repetition interval and number of Simpuls pulses, allow to exclude pulse trains T1, Tn of low frequency recurrence type. The rejection step 130 then comprises the rejection of the pulse trains Ti, Tn whose pulse repetition interval PRI is greater than the repetition interval threshold of the SARI pulses. The rejection step 130 also includes the rejection of the pulse trains T1,..., Tn whose number of pulses is less than the threshold of a number of pulse pulses. The pulse repetition interval threshold SARI and the pulse number pulse threshold thus form a template so that any pulse train TX coming out of this template is rejected. Thus, rejected TX pulse trains are removed from the set ET of T1, Tn pulse trains and are subsequently processed by another specific de-interleaving method. The deinterleave method then comprises a step 140 of classification of pulse trains T1, Tn. The classification step 140 consists of sorting the pulse trains T1,..., T, of the set AND of pulse trains T1,..., I-, according to the carrier frequency of the pulse pulses. each pulse train T1, Tn. During this classification step 140, the pulse trains T1, Tn are divided into two classes C1, C2 of pulse trains T1, Tn. The first class C1 groups the pulse trains T1 of fixed carrier frequency, that is to say the pulse trains T1,..., T, formed of pulses having the same carrier frequency at an uncertainty near measure. The measurement uncertainty is, for example, equal to plus or minus 5 percent (%) of the carrier frequency. Pulse trains T1, Tn of fixed carrier frequency are also called pulse trains T1, Tn single frequency. The second class C2 of pulse trains T1,..., T ,, groups together the pulse trains T1,..., Tr, of variable carrier frequency, that is to say the pulse trains. T1, ..., Tr, formed of pulses of different carrier frequencies. Pulse trains T1,..., 1 -0 of variable carrier frequency are also called pulse trains T1,..., T0 agile in frequency. The carrier frequency of pulses T1, ..., T pulse trains, agile frequency is generally random or pseudo random.
[0013] The deinterleaving method then comprises a step 150 of grouping the pulse trains T1, T, of the set ET of pulse trains T1, .- -, Tn- During the grouping step 150 the pulse trains T1, Tn having the same repetition interval PRI pulses are grouped according to at least one predefined clustering law to form pulse levels P1, - - Pp. It is understood by the expression " pulse stage ", a set of pulses belonging to at least one pulse train formed during the grouping step 150. It is illustrated in FIG. 5, an example of a received signal R by the receiver 12 and a 10 P1 pulse stage formed by the grouping of pulse trains T4, T5 and T6 of this signal. Similarly, it is illustrated in Figure 6, a P2 pulse stage from the grouping of T6 and T7 pulse trains. More precisely, the step 150 of grouping the pulse trains T1, Tn is carried out, on the one hand, for the pulse trains T1, Tn of fixed carrier frequency during a first substep 160 to obtain pulse levels P1, Pp of fixed carrier frequency. The law of regrouping used is called "law of fixed regrouping". The step 150 of grouping the pulse trains T1, Tn is implemented, on the other hand, for the pulse trains T1, ..., Tn of variable carrier frequency during a second substep 170 to obtain pulse levels P1, Pp of variable carrier frequency. The law of regrouping used is called "variable law of regrouping". The pulse trains T1, Tn grouped during the grouping step 150 have not been put together to form a single pulse train T1, Tn during the formation step 110 because such pulse trains T1, Tn exhibited, for example, mitings M, i.e., missing pulses, due to technical reasons. These technical reasons depend in particular on the listening limitations of the receiver 12, the quality of the measurements by the digital signal processing unit 23, the proximity of the receiver 12 to the transmitter, the positioning of the receiver 12 relative to to the transmitter and electromagnetic disturbances.
[0014] The bindings are, for example, represented on the pulse trains T6 and T7. The missing pulses are reconstituted during the formation of the P2 pulse stage. In general, a grouping law consists of an algorithm whose general structure is described with reference to FIG. 7. Initially, the algorithm comprises a phase 150A for selecting an erf reference element out of a set of elements. E.
[0015] The algorithm then comprises a phase 150B for deleting the reference element erf from the set of elements E and for adding, in a set of groups Eg, a gapped reference group comprising l The algorithm then comprises a selection phase 150C in the set of elements E, elements compatible with the gaffed reference group according to a set of criteria C. The elements compatible with the group reference code form a set of candidate elements Ec. The criteria C are, for example, chosen from statistics on the radar waveform characteristics of the waveform database.
[0016] For example, Figure 8 illustrates an already formed group G3 and a group G4 being formed which is at this stage the gopher reference group of the algorithm. The elements el, e2, e3 and e4 are the elements of the set of elements E of the algorithm. As illustrated by a cross in this figure 8, the elements e1 and e3 are incompatible with the group G4 in formation. Conversely, as illustrated by a check mark in FIG. 8, the elements e2 and e4 are compatible to be joined with the group G4 in formation. The set of candidate elements Ec of the algorithm therefore comprises two elements: e2 and e4. The algorithm then comprises a phase 150D for evaluating the distance between the gaffed reference group and each element of the set of candidate elements Ec.
[0017] Then, the algorithm comprises a phase 150E of annexing an element of the set of candidate elements E, to the gated reference group. The appended element is removed from the set of elements E. The element of the set of candidate elements E, joined with the reference group gef is the element having the smallest distance d. It is understood by the term "annexation of an element to a group", the attachment of the element to the group: at the end of the annexation the final group is formed of the initial group and the annexed element which is therefore joined to the initial group. For example, FIG. 9 illustrates the determination of the element to be joined to the gree reference group, that is to say to the group G4 in formation, among the set of candidate elements Ec. Figure 9 includes the same groups and elements as those of Figure 8, but shown in different ways. Thus, the already formed group G3 comprises three elements represented by three circles. The G4 group being formed already has four elements represented by four circles. As can be seen in FIG. 9, the distance d2 separating the element e2 from the group G4 is smaller than the distance d4 separating the element e4 from the group G4. Therefore, the element of the set of candidate elements that minimizes the distance to the gaff reference group is element e4. The e4 element will therefore be brought together with the G4 group in formation. The selection phases of compatible elements 150C, distance evaluation 150D and annexation 150E are then repeated with the remaining elements of the set of elements E, as long as the set of candidate elements In calculated in the phase of selection of compatible elements 150C, comprises elements. Finally, the selection phase 150A, the deletion and addition phase 150B, the selection of compatible elements 150C, the distance evaluation 150D, and the annexation 150E are repeated as long as the set of elements E has elements .
[0018] The fixed clustering law applied in the first substep 160 is specifically described in the following. Each element of the algorithm is a pulse train T1, T ,, of the first class C1 of the pulse trains T1,..., T, of fixed carrier frequency and each group of the algorithm is a step of pulses P1, Pp of fixed carrier frequency. The gde reference group of the algorithm is a pulse stage P1, Pp. In addition, the initial set of elements E of the algorithm is the set of pulse trains T1, Tr, of the first class C1 pulse trains T1, Tn of fixed carrier frequency. During the selection phase 150A, the element chosen from the set of elements E is the element whose arrival time of the first pulse is the smallest. In other words, the selected element is the element of the set of elements E arriving first on the receiver 12 or the element recorded on the interface 32 with the arrival time of the first TOA pulse the smallest. The set of criteria C comprises the following criteria: an arrival direction criterion, a time criterion, a frequency criterion, a pulse width criterion, a pulse repetition interval criterion and a phase criterion . Only the elements of the set of elements E satisfying all the criteria of the set of criteria C are compatible with the reference group and are added to the set of candidate elements In. Alternatively, the set of criteria C does not include all of the stated criteria or includes criteria different from those stated. The arrival direction criterion makes it possible to test the compatibility of each element of the set of elements E with the reference group as a function of their arrival direction DOA. The arrival direction criterion stipulates that in order to be compatible, the ga reference group and the element must be formed of pulses having the same direction of arrival DOA.
[0019] The arrival direction criterion is checked for an element of the set of elements E when the element verifies a comparison test of the arrival directions DOA of the element and the reference group gef. The comparison test is based on a statistical model of the direction of arrival DOA and is adapted to take into account the proportion of outliers. To verify the comparison test, the element of the set of elements E must verify the following equation: 2 Ir (1 1 (D0Ai .-- DOA2) 2 5 2 * * * - + * erf-1 ( P (1/011/0)) 1- 2r 2 n1 n2 where DOA1 is the direction of arrival of the gated reference group, DOA2 is the direction of arrival of the element of the set of elements E to test, 62 is the variance of the arrival direction of the pulses 10 r is the proportion of degraded arrival direction measurements, n1 is the number of pulses present in the reference group gef, n2 is the number of pulses present in the element of the set of elements E to be tested, errl is the reciprocal of the error function, the error function being given by the equation erf (x) = 712 g e-t2 dt , and P (1101110) is the probability of detection of equality between DOA1 and DOA2, H, designating the assumption that the values DOA1 and DOA2 are equal, and 1) (1101H0) designating the probability of making the choice H, knowing that one is in the case H ,. Alternatively, another comparison test is to calculate, then compare, the average arrival direction DOA of the pulses of the gaff reference group with respect to the mean arrival direction DOA of the pulses of each element of the set. As a variant, yet another comparison test consists of calculating, then comparing, the median arrival direction DOA of the pulses of the gaffed reference group 25 with respect to the median DOA arrival direction of the pulses of each element of the set of elements E. The use of the median and not the average makes it possible to avoid taking into account aberrant data strongly present in the DOA arrival direction measurements. The time criterion makes it possible to test the temporal compatibility of each element 30 of the set of elements E with the reference group gef. The temporal criterion states that an element of the set of elements E superimposed in time with the reference group gef is incompatible with the gaff reference group. The time criterion further states that an excessive time difference between an element of the set of elements E and the ga reference group also causes an incompatibility of the element with the gree reference group. Indeed, a significant time difference between the gaffed reference group and the element may mean that there have been no pulses emitted. In this case, the reference group and the element, even if they belong to the same issuer, are distinct. The maximum time difference separating the gob reference group of a compatible element from the set of element E is, for example, a multiple of the average repetition interval of the average pulses of the gaff reference group. The maximum time difference is, for example, equal to twenty times the average value of the repetition interval of the PRI pulses of the gaff reference group. The value of the multiple is determined empirically or from the waveform database. Alternatively, other values of multiples or other threshold values that do not depend on the repetition interval of the PRI pulses can still be considered. The frequency criterion makes it possible to test the frequency compatibility of the elements 15 of the set of elements E with the reference group. The frequency criterion states that to be compatible two elements must be formed of pulses having the same carrier frequency f. The frequency criterion is thus verified for an element of the set of elements E when the element verifies a comparison test of the carrier frequencies of the pulses of the element with the carrier frequencies of the pulses of the gaffed reference group. The comparison test compares the average carrier frequency of the pulses of the ga reference group with the average carrier frequency of the pulses of each element of the set of elements E. Such a comparison test is based on the fact that a frequency measurement follows a Gaussian model of known variance. Nevertheless, assuming that the carrier frequency measurements are obtained by means of frequency windows, the possibilities of exclusion of certain frequency values are to be expected. In this case, the distribution of the measured carrier frequencies is no longer Gaussian, because partially truncated, and therefore the comparison of averages is biased.
[0020] As a variant, especially in the case where the frequency distribution is no longer Gaussian, an x2 test is used. To verify the frequency criterion, the element of the set of elements E must verify the following equation eq2: nl n2 2 2 N2 f1, i f2, i <2 2 f12, i f2 i + n2> * Xni + n2-1, P (Ho 'Ho) i = oi = o 3031257 17 where hi is the carrier frequency of the order pulse i of the reference group g, f2, i is the carrier frequency of the pulse of order i of the element of the set of elements E to be compared, n1 is the number of pulses of the reference group gef, n2 is the number of pulses of the element of the set of elements E to compare, cry is the variance of the distribution of the measurements of the carrier frequency of the pulses, 10 ep is the quantile xp for the probability p of the distribution of chi2 at y degrees of freedom, ie if a random variable X follows a chi-square law with y degrees of freedom, the probability of having) (xp is p, and P (1-101110) is the probability of detection of equality between ri / 10A2, i and E720f2i, H, designating l 'hypot assume that the two frequency values are equal, and P (1101110) designating the probability of making the choice Ho knowing that it is in the case Ho. The pulse width criterion makes it possible to test the compatibility in width of LI pulse elements of the set of elements E with the reference group gree. The pulse width criterion states that in order to be compatible, the reference group and the element must be formed of pulses having the same pulse width.
[0021] As illustrated in FIG. 4, the pulse width LI of a pulse is the duration of the pulse. The pulse width criterion is verified for an element of the set of elements E when the pulse widths LI of the pulses of the element are of the same order of magnitude as the pulse widths LI of the pulses of the group. reference gree. It is understood by the term "same order of magnitude", a relative deviation less than a threshold determined from the database of waveforms, or equal for example 50%. The PRI pulse repetition interval criterion makes it possible to test the compatibility of the pulse repetition interval PRI of each element of the set of elements E with the gated reference group. The PRI repetition interval criterion stipulates that the pulses of each element of the same group must have the same repetition interval of the PRI pulses. The pulse repetition interval criterion PRI is checked for an element of the set of elements E when the element checks a comparison test of the repetition interval of the element's pulses PRI with the element E. repetition interval of the PRI pulses of the gaff reference group. The comparison test consists first of all in estimating, for each element of the set of elements E, the repetition interval of the theoretical PRI pulses resulting from the union of the pulses of the element with the pulses of the reference group. The repetition interval of the theoretical PRI pulses is estimated, for example, by the Fourier method or the least squares method. The comparison test then consists in performing a suitability test to check, for example, whether the sum of the squared errors is greater than or less than a threshold. The adequacy test is for example an x2 test. This comparison test has the advantage of determining, in addition, the repetition interval of the PRI pulses of the group obtained when the element of the set of elements E is joined with the reference group. However, such a comparison test is relatively slow and resource consuming. Alternatively, a shorter and less resource consuming comparison test is presented. To verify the PRI repetition interval criterion, the element of element set E must verify the following equation eq3: (PRI1 - PRI2) 2 5 4 * oloa * (erf-1 (Pd) ) Where PR / 1 is the repetition interval of the gated reference group pulses, PR / 2 is the pulse repetition interval of the element of the element set E to be compared, 20 '' TOA is the variance of the TOA pulse arrival time distribution, erf-1 is the reciprocal of the error function, and Pd is the equality detection probability of two PRI values. Alternatively, if the accuracies of the repetition intervals of the pulses of each element have been estimated, yet another test is proposed. To verify the PRI pulse interval criterion, the element of element set E must verify the following equation eq4: (PRI1 - PRI2) 2 5 2 * (var (PRI1) + var (PRI2 )) * (erf-1 (Pd)) 2 where var (PRI1) is the variance of the pulse repetition interval of the ga reference group, 30 var (PRI2) is the variance of the pulse repetition interval of the element of the set of elements E to be compared, and the other notations are those of the equation eq3.
[0022] The pulse repetition interval criterion also takes into account that some steps may have missing pulses and therefore the measured pulse pulse repetition interval is likely to be a multiple of the repetition interval. real PRI pulses.
[0023] Thus, the pulse repetition interval criterion PRI also compares the multiples of the pulse repetition intervals of each element of the set of elements E with the repetition interval of the PRI pulses of the gapped reference group, and Conversely. For example, the comparison test compares the repetition interval of the reference pulses GFR of the reference group with respectively two times, three times, four times, the repetition interval of the pulses PRI of each element of the set of elements E or the repetition interval of the pulses PRI of each element of the set of elements E with respectively twice, three times, four times, the repetition interval of the pulses PRI of the reference group. The number of multiples to be compared depends on the listening conditions and the percentage of the M elements. Nevertheless, three is a reasonable multiple limit. It is, in fact, unlikely that a mite element retains one pulse out of four at regular intervals. The phase criterion makes it possible to test the phase compatibility of the arrival times of the pulses of each element of the set of elements E with the reference group. The phases in time of arrival of the pulses of two elements were previously calculated with respect to a reference time tref. However, if the reference time tref is far from the temporal position of the elements to be compared, this can be problematic for the comparison. From a mathematical point of view, the accuracy of the arrival phase of the pulses is obtained by the least squares method or by the Fourier method. The accuracy of the phase in arrival time of the pulses satisfies the following equation eq5: E Icl 2 var esp = n E ki2 '' TOA - (Eki) 2 where rp is the estimate of the phase as a function of time d pulse arrival, t, 11-K is the rank or order of pulse i with respect to zero time such that ki = [PRI 30 with ti the arrival time of pulse i and PRI the pulse repetition interval and [] denoting the integer part, n is the number of pulses of the obtained group, and 3031257 is the variance of the measurement distribution of the arrival time of the TOA pulses. From the equation eq6, it is possible to show that the accuracy of the phase is better, ie, smaller, when the orders of the pulses are weak, that is to say when the time reference tref is close to the time position of the element. The phase criterion assumes that the two elements to be compared first satisfy the PRI repetition interval criterion. If this is not the case, the reference group and the element are incompatible in phase. The phase criterion is checked for an element of the set of elements E when the element verifies a comparison test of the phases of the element and the reference group. The comparison test comprises first of all the determination of a new reference time tref, which will be positioned at the center of the temporal position of the two elements to be compared. The comparison test then comprises the calculation of the phases of the reference group and the element in this new reference time tref. Finally, the element of the set of elements E is compatible with the gree reference group when it satisfies the following equation eq6: (çPioa, 1 4α, 2) 2 5 2 (varCeto., I) + varFPioa, 2)) * (erf-1 (P (H011-10))) 2 WHERE (pitoaa = ((Ptoa, i tref) * mod (PRIi) and ePtoa, 2 = (ePtoa, 2 tref) * mod ( PRI2), toafin, i + toadeb, 2 where tref = is the average between the arrival time of the last pulse of the reference group and the arrival time of the first pulse of the element of the set of elements E to compare, (x) mod (y) indicates that one computes a modulo by deducting y as many times as necessary to x so that (x) mod (y) is between 0 and y, ePtoa , 1 is the arrival time phase of the TOA pulses of the gated reference group, ePtoa, 2 is the arrival time phase of the TOA pulses of the element of the set of elements E to be compared, var denotes the variance, where el-ri is the reciprocal of the error function, 3031257 21 and where P (1101 Ho) is the probability of detection of equality between the phases, H, designating the hypothesis that the phases are equal, and P (1-101H0) designating the probability of making the choice H, knowing that we are in the case H ,. Still in a variant, especially when the variance data are not accessible, the element of the set of elements E is declared compatible with the reference group when it checks the equation eq7 Pioaa (Ptoa, 21 5 2 * Utoa * erf 1 (Kilo I / 10)) where the notations are identical to the notation of equations eq5 and eq6. Only the elements of the set of elements E satisfying all the criteria of the set of criteria C are compatible with the reference group. The possibility of having at least two elements compatible with the greference group is infrequent. However, if this is the case, the distance d is the time difference separating the last impulse 1m from the gaff reference group of the first impulse If of the candidate element of the candidate element set Ec. The element joined with the reference group gef is then the element which minimizes the distance d, that is to say the element for which the time difference between the first pulse of the element and the last I pulse, the reference group is minimal. When an element of the set of candidate elements E, is appended to the reference group, the characteristic data of the group obtained are updated. Thus, the arrival time of the first pulse of the group obtained is the minimum of the arrival time of the first pulse of the gaffed reference group and the attached element. The arrival time of the last pulse of the group obtained is the maximum of the arrival times of the last pulse of the gaffed reference group and the attached element. The direction of arrival DOA of the group obtained is the average of the arrival directions of the gaffed reference group and the appended element. The carrier frequency of the group obtained is given by the following equation eq5: fgroup - N1 * f1 + N2 * f2 Ni + N2 where fgroupe denotes the carrier frequency of the group obtained, N1 denotes the number of pulses of the reference group, F1 denotes the carrier frequency of the gaff reference group, f2 denotes the carrier frequency of the attached element, and N2 denotes the number of pulses of the attached element. The sum of the carrier frequencies of the group obtained is the addition of the sum of the carrier frequencies of the gaffed reference group and the sum of the frequencies 3031257 carrying the attached element. The sum of the squared carrier frequencies of the obtained group is the addition of the sum of the squared carrier frequencies of the gaffed reference group and the sum of the squared carrier frequencies of the attached element. The number of pulses of the obtained group is the sum of the number of pulses of the gaffed reference group and the number of pulses of the appended element. The pulse width LI of the group obtained is given by the following equation eq9: Llg''pe N1 + N2 where L1 group denotes the length of pulses LI of the group obtained, N1 denotes the number of pulses of the reference group. , Lli denotes the pulse length LI of the gated reference group, LI2 is the pulse length LI of the appended element, and N2 is the number of pulses of the appended element. The repetition interval of the PRI pulses of the obtained group is calculated by means of a least squares linear regression on all 15 pulses of the new group. This linear regression is given by the following echo equation: ## EQU1 ## where k, the order of the pulses, t1 the arrival time of pulses and n the number of pulses of the group obtained. Alternatively, an approximation of the repetition interval of the PRI pulses is calculated by assuming the estimates of the repetition interval of the gp reference group pulses PRI and the pulse repetition interval PRI of the appended element. independent of each other. The approximation consists, firstly, in correcting the repetition intervals of the PRI pulses of the elements of the group and the repetition interval variances of the PRI pulses of the elements of the group when the repetition interval of the PRI pulses of the group. one of the elements of the group is a multiple of the repetition interval of the PRI pulses of the other element of the group. For example, if the repetition interval of the gp reference group pulses is equal to twice the repetition interval of the pulses PRI of the appended element, the repetition interval of the gp reference group pulses PRI is halved and the repetition interval variance of the gp reference group PRI pulses is divided by four. Then, the repetition interval of the PRI pulses of the resulting group is calculated from the following echi equation: var (PRI2) * PRI1 + var (PRIi) * PRI2 PRI = var (PRIi) + var (PRI2) where PRI is the estimate of the repetition interval of the pulses of the group obtained, PR / 1 is the estimate of the repetition interval of the pulses of the reference group g ref, var (PRIi) is the variance of PRII, PRI2 is the estimate of the repetition interval of the pulses of the annexed element, and var (PRI2) is the variance of PR / 2.
[0024] In addition, the variance of the estimate of the pulse repetition interval PRI obtained with the equation eqiverifies the following equation eq12: var (PRIi) * var (PRI2) var (PRI) = var (PRIi) + var (PRI2) where var (PRI) is the variance of the estimate of the repetition interval of the pulses of the group obtained and the other notations are identical to those of the equation eqii. Alternatively, in the case where the variances of the repetition intervals of the PRI pulses have not been estimated for each element of the obtained group, another calculation of the repetition interval of the PRI pulses consists in correcting the repetition intervals of the pulses. multiples as in the previous variant and to calculate the estimate of the repetition interval of the PRI pulses of the group obtained by using the following equation eq13: PRI = (T0Afin - TOAdeb + 2) 20 where TOAdeb is the arrival time of the first impulse of the group obtained, TOAfin is the arrival time of the last impulse of the group obtained, and the other notations are identical to the notations of the equation eqii. The variance of the estimate of the pulse repetition interval PRI obtained with the equation verifies the following equation eq14: 2 (TOAfin - TOAdeb1) 2 PRI + 2) where TQA is the variance of the time measurement of pulse arrival, and the other notations are identical to the notations of equation eq13. TOAfin - TOAdeb var (PRI) = 2 3031257 24 The phase in arrival time TOA of the pulses of the group obtained is calculated following the following steps. In the first place, a new reference time tref is taken centered between the first element and the second element forming the group. Then, the phases of each of the elements forming the group are calculated with this new reference time tref. Then, the phase of the group obtained with respect to this reference time tref satisfies the following equation eq16: ePTOA = + N2 where (pj, 0A is the phase in time of arrival of the pulses of the group obtained calculated with respect to a time reference number tref, (p-0,4,1 is the phase of the reference group gef recalculated with respect to a reference time tref, N1 is the number of pulses of the gaff reference group, and N2 is the number of Finally, the phase of the obtained group calculated with respect to a time of origin satisfies the following equation eq16: (Proa = ((P.0,4 + tref) * mod (Fru) 15 where (proA is the phase in time of arrival of the pulses of the obtained group computed with respect to the time of origin, (x) mod (y) denotes the modulo as in equation eq6, tref is the reference time relative to to which the cproA phase has been calculated, and is the estimate of the repetition interval of the pulses of the group obtained, and notations are identical to the notations of the equation eq16, the original time is the time of reception of the first pulse on the receiver 12 or the smallest time of the pulses recorded on the interface 32. The law of variable grouping applied in the second substep 170 is specifically described in the following.
[0025] Each element of the algorithm is a pulse train T1, ..., Tri of the second class C2 of pulse trains T1,..., Tr, of variable carrier frequency and each group of the algorithm is a pulse bearing P1, Pp of variable carrier frequency. In addition, the set of elements E of the algorithm is the set of pulse trains T1, Tn of the second class C2 of pulse trains T1, Tn of variable carrier frequency. In what follows, only the differences of the second substep 170 with respect to the first substep 160 are highlighted. The frequency criterion stipulates that the difference between the maximum carrier frequency of the pulses of an element and the minimum carrier frequency of the pulses of an element is determined by means of the frequency criterion. the threshold must be below a threshold This threshold is related to the frequency scanning technology limits of the radar transmitters of interest This threshold is chosen from the 5 waveform statistics of the database used. This threshold is, for example, understood in the broad sense between 100 Megahertz (MHz) and 10 Gigahertz (GHz) .The deinterleaving method then comprises a step 180 of association in overlapping the pulse levels P1, Pp according to a In the overlapping association step 180, the pulse stages P1, Pp having the same repetition interval of the PRI pulses and which overlap in time are grouped together to form groups of steps. of pulses G1, ..., G9 in overlap More specifically, step 180 of association of the pulse bearings Pl, Pp is carried out solely for the fixed carrier frequency pulses P1,. grouping step 150. The overlapping association step 180 makes it possible, in particular, to group frequency-modulated waveforms called FMICW (abbreviation of the English Frequency Modulated Interrupted Continuous Wave). The FMICW waves are formed of pulses spaced from the same repetition interval of the PRI pulses and whose carrier frequency increases continuously in time for a given duration. The frequency profile of a FMICW wave is illustrated as a function of time in FIG. 10. The pulses of the same frequency and having the same repetition period of the PRI pulses form a pulse train or a pulse stage. As can be seen in FIG. 10, each pulse belongs to a pulse stage P3. The pulses of the same level of pulses P3 have the same frequency. The P3 pulse bearings of FIG. 10 have not been grouped to form a single pulse level during the grouping step 150 because the time criterion of the grouping step 150 eliminates the elements that are superimposed on each other. the weather. However, the pulse levels or pulse trains of a FMICW wave are superimposed over time.
[0026] The overlapping association step 180 therefore makes it possible, in the case of FMICW waves, to assemble incrementally overlapping pulse levels so as to reconstruct a FMICW signal. Thus, in this FIG. 10, the assembly of the pulse bearings P3 makes it possible to form a group of pulse bearings G1.
[0027] The association law used to associate in overlap the pulse levels P1, Pp of fixed carrier frequency is subsequently called "overlapping association law". The overlapping association law is implemented by an algorithm of the same general structure as the general structure of the fixed recovery law algorithm described in the grouping step 150. Each element of the algorithm is a pulse bearing P1, Pp fixed carrier frequency and each group of the algorithm is a group of pulse bearings G1, ..., Gg overlap. The gure reference group of the algorithm is a group of 10 levels of pulses G1, ..., Gg. In addition, the set of elements E of the algorithm is the set of pulse levels P1, Pp of fixed carrier frequency. In what follows, the characteristics of the law of association in recovery identical to the characteristics of the law of fixed grouping are not described again. Only the differences of the overlapping association law with respect to the fixed clustering law are described in the following. The temporal criterion states that to be compatible, two elements must be superimposed over time. When the gated reference group is a group of pulse stages, the frequency criterion is identical to the frequency criterion of the second sub-step 170 of the grouping step 150. In this case, the phase criterion is not identical. not applied. Indeed, there is no constraint on the phases in arrival time of the pulses at the level of groups of bearings. When the gated reference group is a pulse stage P1, Pp, the phase criterion is identical to the phase criterion of the first substep 160 of the grouping stage 150. In this case, if the criterion of phase is checked and if the criteria of direction of arrival, time, pulse width and pulse repetition interval are also verified, then the frequency criterion is evaluated according to two cases of figures. In the first case, the carrier frequencies of the pulses of the two P1, Pp pulse levels to be compared are identical and the theoretical number of pulses of the two levels together is greater than or equal to the actual number of pulses of the two levels. pulses P1, Pp joined together. In this first case, both levels are compatible. In the second case, the carrier frequencies of the pulses of the two pulse levels P1, Pp to be compared are different or the theoretical number of pulses of the two levels together is less than the actual number of pulses of the two levels. pulses P1, Pp joined together. In this second case, the two levels are incompatible. Between several compatible elements to be joined with a gated reference group, the overlap association step 180 gives priority to the element whose phase is already present in the gaff reference group. This first priority rule makes it possible to first gather the pulse levels P1, Pp that are broken up and have the same repetition interval of the PRI pulses. In addition, if a group corresponding to an FMICW is already in training, a second priority rule applies. According to this second priority rule, the information of the steps already present in the group corresponding to a FMICW are consolidated, before adding a new pulse level P1,..., Pp to the group. Once the precedence rules are applied, preference is given to the compatible element that has the best recovery rate with the gref reference group. The recovery rate is given by the equation eqi, as follows: AT recovery rate Min (ATi, AT2) where recovery rate is the recovery rate between a first element and a second element, this recovery rate is between 0 and 100% LTrovement is the time difference between the reception time of the first impulse of the compatible element and the reception time of the last impulse of the reference group grf, LiTi is the duration of the reference group grf, AT2 is the duration of the compatible element, and Min (AT1, AT2) is the minimum of ATiet AT2. The distance d that one element of the set of candidate elements 25 must be minimized to be joined to the gaff reference group is given by the following equation eq18: d = a + (1 - recovery rate) where a takes the value zero if the phases are equal, and takes the value 1 if the phases are not equal, the other notations being identical to the notations of the equation eq17. The choice of such a distance d makes it possible to join to the reference group, the element that makes the rate of explanation grow as fast as possible. The explanation rate of a group is the ratio of the number of real pulses of the group to the number of theoretical pulses of the group.
[0028] When an element is appended to the gree reference group, the characteristics of the group obtained in terms of direction of arrival DOA of the pulses, time of arrival of the first pulse, time of arrival of the last pulse, of pulse frequency, pulse number, pulse repetition interval PRI, pulse width LI, pulse arrival phase are updated in the same manner as in the case of the first substep 160 of the grouping step 150. The deinterleaving method then comprises a step 190 of switching association of the pulse stages.
[0029] During the switching association step 190, the pulse stages P1, Pp having a different pulse repetition interval PRI and which follow each other in time are grouped according to at least one predefined association law to form groups of pulse bearings G1,..., Gg in commutation. For example, as illustrated in FIG. 12, the pulse bearings P7, P3, Pg, and P10 are connected by arrows in fine lines to form a group of pulse bearings G1,..., Gg in switching mode. . Similarly, the pulse bearings 11 P tP -, P 12, P 13 and P 14 are connected by arrows in bold lines to form another group of pulse bearings G1, ..., Gg in switching. As shown in FIG. 12, each bearing group G1,... Gg is formed of bearings which do not overlap in time and which have repetition intervals of the different PRI pulses. More precisely, the step 190 of association in switching of the pulse bearings P1, Pp is carried out, on the one hand, during a first substep 200, for the pulse levels P1, Pp of fixed carrier frequency not having been grouped together during the overlapping association step 180. At the end of this first substep 200, groups of pulse bearings G1,..., Gg in overlap fixed carrier frequency are obtained. In this case, the association law used is called "fixed switching association law". The step 190 of switching association of the pulse bearings P1, Pp is implemented, on the other hand, for the pulse stages P1, Pp of variable carrier frequency during a second substep 210 to obtain pulse bearing groups G1, ..., Gg in variable frequency carrier switching. In this case, the association law implemented is called "variable switching association law".
[0030] Each of the switching association laws is implemented by an algorithm whose general structure is identical to the general structure of the algorithm described in the grouping step 150. Each element of the algorithm is a step pulses P1, Pp and each group of the algorithm is a group of pulse stages G1, ..., G9. The reference group is a group of pulse stages G1, ..., G9. In the following, the characteristics of the fixed switching association law identical to the characteristics of the fixed grouping law are not described again. Only the differences of the law of association in fixed commutation with respect to the law of fixed grouping are described in what follows. Each element of the algorithm is a pulse stage P1, Pp of variable carrier frequency that has not been gathered during the overlapping association step 180. Each group of the algorithm is a group of steps of pulses G1, ..., G9 in fixed carrier frequency switching. In addition, the set of elements E of the algorithm is the set of fixed carrier frequency pulses P1, Pp that have not been gathered during the grouping step 150. The width criterion The pulse rate is used to determine the steps having near pulse widths LI or near shape factors. The shape factor of a bearing is defined as the ratio between the bearing pulse width LI and the pulse repetition interval PRI. Indeed, the observation of specific waveforms, particularly high frequency waveforms of recurrence and average recurrence frequency, shows a constant in the values of the pulse widths of the same signal and / or in the values of the shape factors of the bearings of the same signal. As the pulse width measurements LI are relatively unreliable, the pulse width criterion consists of a comparison test of the orders of magnitude of the pulse widths of two elements or the shape factors of two elements. The pulse repetition interval criterion PRI determines the steps having repetition intervals of the near pulses. In fact, the observation of different waveforms makes it possible to define an upper limit in the intervals of repetition intervals of the pulses of the bearings of the same signal. Indeed, for the same signal, the repetition intervals of the pulses of the different stages forming the signal must not be spaced by more than a certain threshold value. Therefore, two steps are compatible when the repetition intervals of the PRI pulses of each of the two levels satisfy the following equation: 19 Max (PRI1, PRI2) -Min (PRI1, PRI2) <s Max (PRI1, PRI2) + MIN (PRIi.PRI2) 1 Where PR / 1 is the repetition interval of the pulses of the gaff reference group, PRI2 is the repetition interval of the pulses of the annexed element, Max (PRI1, PRI2) is the maximum PR / let PRI2, 5 Min (PRI1, PRI2) is the minimum of PRI1 and PRI2 and S1 is a relative difference threshold chosen from the waveform statistics of the database used, or equal to example 60%. Equation eq19 is valid for high and medium frequency recurrence waveforms.
[0031] The frequency criterion makes it possible to determine the bearings having frequencies carrying close pulses. Indeed, the observation of different waveforms makes it possible to set an upper limit in the carrier frequency deviations of the pulses of the bearings of the same signal. Such an upper limit is related to the technological limitations of issuers. According to the frequency criterion, two compatible elements first satisfy the following equation eq20: If1-fmoyl e ç + fmoy -2 where f1 is the carrier frequency of one of the elements to be tested, fmoy is the average of the frequencies carriers resulting from the union of the two elements to be tested, and S2 is a relative difference threshold chosen from the statistics of the waveforms 20 of the database used, or equal for example 10%. Moreover, according to the frequency criterion, two compatible elements each also satisfy the following equation eq21 Ifl fmoyl <S3 where the notations are identical to the notations of the equation eq20, and S3 is a threshold chosen from the statistics on the forms of the data base used, or being, for example, broadly defined between 100 MHz and 10 GHz. The criterion of number of stepwise impulses makes it possible to determine the steps having numbers of close pulses. Indeed, the observation of different waveforms makes it possible to determine a threshold beyond which two levels are incompatible. According to the criterion of number of pulses, two compatible elements each satisfy the following equation eq22: INthreshold Nthmoyl Nth + Nth <54 average pitch 3031257 31 Where Nthpati 'is the theoretical number of pulses of one of the elements to be tested , Nthmoy is the theoretical mean number of pulses resulting from the union of the two elements to be tested, and S4 is a threshold set as a function of the waveform database used or understood in the broad sense, for example between 5% and 30% for waveforms of high or medium frequency of recurrence type. During the first substep 200 of the switching association step 190, the distance measurement d makes it possible to select among the compatible elements of the algorithm the one that shares the most common characteristics with the reference group. of the algorithm. The common characteristics sought are as follows: the reference group gree has one of its pulse widths LI close to the pulse width LI of a pulse of a candidate element, the gree reference group has a form factor Near the form factor of a candidate element, the gree reference group shares one of its carrier frequencies with a candidate element or the gree reference group has the same number of theoretical pulses as one of the candidate elements. The distance d is defined by a score given by the following equation eq23: d = 1 * b + 1 * c + 1 e + 1 * g + 1 * h + score where b is a number equal to 1 if the two elements to compare do not present a common form factor and equal to 0 otherwise, 20 c is a number equal to 1 if the two elements to be compared do not have common pulse widths and equal to 0 otherwise, e is an equal number at 1 if the two elements to be compared do not have a common theoretical number of pulses and equal to 0 otherwise, g is a number equal to 1 if the two elements to be compared do not have a repetition interval of the common PRI pulses and equal to 0 otherwise, h is a number equal to 1 in the case of single-frequency bearings where the two elements to be compared do not have a common pulse carrier frequency and equal to 0 in the other cases, and score is a threshold between 0 and 1 and given by the following equation eq24: score = 1 (TOAdeb, 2-T0Afm, i f2 fmoy , 1) fife f 2 Atref 30 where TOAdeb, 2 is the arrival time of the first pulse of the element, TOAfin, i is the arrival time of the last pulse of the reference group, 3031257 32 fmoya is the the average carrier frequency of the pulses of the reference group grf, f2 is the carrier frequency of the pulses of the element, At'f is a value chosen from the statistics on the waveforms of the database used, or value in the broad sense between 1 ms and 10 ms, Af'f is a value chosen from the statistics on the waveforms of the database used or could for example be equal to the threshold S3 of the equation eq21 . Alternatively, another distance d can still be used. The updating of the characteristics of each group obtained at the end of the first substep 200 of the switching association step 190 is identical to the update performed during the overlapping association step. 180. The second substep 210 of the switching association step 190 is strictly identical to the first substep 200 of the switching association step 190 except that each element of the algorithm is a pulse bearing P1, Pp of variable carrier frequency and each group of the algorithm is a group of pulse bearings G1, ..., Gg in variable carrier frequency switching. In addition, in the case of the second substep 210, the set of elements E of the algorithm is the set of pulse stages P1, Pp of variable carrier frequency.
[0032] Furthermore, the common feature that the gree reference group shares one of its carrier frequencies with a candidate element is not sought during this second substep 210 and the number h of the equation eq23 is still zero in this second substep 210. Thus, at the end of the overlap association step 180 and the switching association step 190, three types of pulse stagger groups Gg are obtained: a first type of pulse bearing groups G1, ..., Gg overlapping, a second type of pulse bearing groups G1, ..., Gg in fixed carrier frequency switching and a third type of pulse bearing groups G1,..., Gg in variable carrier frequency switching. Each group type groups radar signals with common characteristics. Each group results from the concatenation of the pulse trains T1,..., Tr, forming the associated pulse stages P1, Pp during the association steps 180 and 190. Each group is therefore a de-interlaced radar signal formed of the concatenation of pulses according to a regrouping law and a law of association.
[0033] Figure 13 illustrates the deinterleave chain of radar signals. In this figure 13, three Eradar radar transmitters emit radar signals that are intercepted by the receiver 12. The signals received by the receiver 12 and from different radar transmitters are mixed and superimposed between them. The implementation of the method 5 according to the invention by the computer 14 makes it possible to group together the trains of pulses coming from the same radar transmitter and thus to obtain three groups of deinterleaved radar signals. Thus, the deinterlacing method according to the invention makes it possible to separately group the pulse trains T1,..., Tn of different radar signals whose pulse repetition interval is medium or short, in a dense electromagnetic medium. The grouping laws and the association laws used to group the pulse trains T1,..., Tri of the same radar signal, implement a first phase of selection of the compatible elements and a second phase of minimization. a distance d. The first selection phase makes it possible, in particular, to eliminate the pulse trains T1, Tn or pulse stages P1, Pp which do not have the same structural characteristics and thus to overcome the obtaining of de-interlaced signals. wrong. Thus, the set of criteria C allows a rapid elimination of the elements of the algorithm incompatible with each other. The speed of the algorithm is, thus, better than 0 (N2), with N the initial number of elements of the set of elements E of the algorithm, that is to say that the time of calculation of the algorithm grows slower than the square of the number of elements to be processed. The algorithm for grouping pulse trains generally requires less computing time than a few milliseconds or tens of milliseconds on a current processor. In addition, the distance d to be minimized is specific to each of the grouping and association laws so as to take into account the structural characteristics of each of the group types in particular. The classification step 140 classifies and then processes the pulse trains T1, Tn according to the waveform at which the pulse trains T1, Tn belong. The waveform databases are therefore exploited so as to implement a treatment adapted to the different types of waveforms. For example, the overlapping association step 180 is specific for fixed carrier frequency waveforms having time-superimposed pulse stages P1, Pp, especially FMICW waveforms. In the absence of the overlapping association step 180, such waveforms would not have been optimally grouped together during the switching association step 190.
[0034] Moreover, step 120 of rejection of the pulse trains 1-1, Tn which are incoherent in the repetition interval of the pulses PRI and optionally at the carrier frequency of the pulses makes it possible to perform a first sorting of the T1 pulse trains. , T ,, not viable. Thus, pulse trains T1, ..., T, incoherent are eliminated.
[0035] Likewise, the step 130 for rejecting pulse trains T1, Tn that do not belong to predefined waveforms makes it possible to increase the accuracy of the grouping of pulse trains T1, T, at the time of transmission. grouping step 150 and the accuracy of the association of the Pp pulse levels during the association steps 180 and 190. In addition, the merging is also taken into account since the grouping and associating laws of the method include the repetition interval hypothesis of double or triple PRI pulses. Thus, the deinterleaving method is innovative because of its better control of the input data, by its prioritization of the deinterleaving process steps which allows a faster and closer to the data classification and the use of rules. grouping based on databases of real waveforms. The deinterleaving method allows the grouping of the pulse trains with greater reliability than in the state of the art, while not degrading the speed performance of the algorithms used.
[0036] Moreover, the algorithms of the laws of grouping and association are adaptable, insofar as each of the criteria of the set of criteria C and each of the tests of comparison of these criteria are modifiable without requiring to modify the all the algorithms. Similarly, the distance d is also locally modifiable for each of these algorithms. Thus, the method according to the invention is quite capable of following the evolution over time of the radar transmitters.

权利要求:
Claims (13)
[0001]
CLAIMS 1.- A method of deinterleaving radar signals, the method comprising: - receiving (100) electromagnetic signals from a receiver (12) and extracting the pulses (li, Im) from the received signals, - forming (110) of pulse trains (T1, Tn) comprising at least three pulses spaced from the same pulse repetition interval (PRI), each pulse train (T1, ..., Te) being defined by the pulse interval pulse repetition (PRI), characterized in that the method further comprises: - grouping (150) the pulse trains (T1, Tn) having the same pulse repetition interval (PRI) according to a predefined regrouping law to form pulse levels (P1, Pp), and - the association (180, 190) of the pulse levels (P1, Pp) according to at least one predefined association law to obtain deinterleaved radar signals formed from the concatenation of pulse trains (T1, ..., Tn ) associated pulse steps (P1, ..., Pp).
[0002]
The method of claim 1, wherein each pulse train (T1, Tn) is also defined by at least one element selected from a group consisting of: the arrival time (TOA) of the first pulse (11) ) of the pulse train (T1, ..., Ta), the arrival time (TOA) of the last pulse (lm) of the pulse train (T1, ..., Te), the pulse frequency (I1, ..., Im) of the pulse train (T1, ..., Te), the pulse duration (Il, lm) of the pulse train (T1, Tn) and the direction of arrival of the pulses (Il, ..., lm) of the pulse train (T1, .- Ta).
[0003]
3. A method according to claim 1 or 2, wherein the method comprises before the grouping step (150), a classification step (140) pulse trains (T1, ..., Te) according to their frequency carrier to obtain two classes (C1, C2) of pulse trains (T1, Tn): a class (C1) grouping the pulse trains (T1, ..., Te) of fixed carrier frequency and the other class (C2) grouping the pulse trains (T1, ..., Te) of variable carrier frequency, the grouping step (150) being implemented for each of the two classes (C1, C2) of pulse trains (T1, ..., Tn) and making it possible to obtain single-frequency pulses (P1, ..., Pp) from the class (C1) of the pulse trains (T1, Tn) mono Frequency and pulse levels (P1, Pp) frequency-agile from the class (C2) of the pulse trains (T1,..., Tr,) that are frequency-agile.
[0004]
4. A method according to any one of the preceding claims, wherein the associating step (180, 190) comprises a step of grouping the pulse steps (P1, Pp) of pulse repetition intervals (PRIs). ) and which are linked in time to obtain groups of pulse bearings (G1, ..., G9) switching. 10
[0005]
5. A method according to any one of the preceding claims, wherein the association step (180, 190) comprises a step of grouping the fixed carrier frequency pulses (P1, Pp) with repetitions of the pulses (PRI) identical and superimposed over time to obtain groups of pulse bearings (G1, ..., G9) recovery. 15
[0006]
6. A method according to any one of the preceding claims, wherein each of the grouping and association laws is implemented by at least one algorithm for obtaining groups from elements, the elements designating trains. pulses (T1, Tn) during the grouping stage and the pulse stages (P1, Pp) during the association step, the groups designating pulse levels (P1, Pp) during of the grouping step and the groups of pulse levels (G1, ..., G9) during the association step, the algorithm comprising: o the choice (150A) of a reference element ( one of a set of elements (E), the deletion (150B) of the reference element (eef) from the set of elements (E) and the addition, in a set of groups (E9 ), a reference group (greef) including the reference element (eef), o the selection (150C) in the set of elements (E), elements compatible with with the reference group (greef) according to a set of 30 criteria (C) to obtain a set of candidate elements (Es), o the evaluation (150D) of the distance between the reference group (greef) and each element of the set of candidate elements (Es), o the annexation (150E) of the element of the set of candidate elements (Es) minimizing a distance (d) to the reference group (gef) and the deletion 35 of the annexed element of the set of elements (E), the repetition of the selection (150C), evaluation (150D) and annexation (150E) phases as long as the set of candidate elements (Es) comprises elements, and the repetition of all of the preceding phases (150A, 150B, 150C, 150D, 150E) as long as the set of elements (E) comprises elements.
[0007]
7. The method according to claim 6, wherein the reference element (eref) is the element of the set of elements (E) whose arrival time (TOA) of the first pulse (11) is the smallest. 10
[0008]
The method of any one of claims 6 or 7, wherein the set of criteria (C) evaluates the compatibility of the elements of the set of elements (E) with the reference group (gaff) according to one or more characteristics, the characteristics being chosen from a group comprising: the arrival direction (DOA) of the elements, the temporal superimposition of the elements, the carrier frequency (f) of the elements, the pulse width ( LI) elements, the pulse repetition interval (PRI) of the elements, the phase (1) of the elements and the number of pulses of the elements. 20
[0009]
The method of any one of claims 6 to 8, wherein the criteria (C) are selected from statistics on the characteristics of the radar waveforms of a database.
[0010]
10. A method according to any one of claims 6 to 9, wherein for the clustering law, the distance (d) is the time difference separating the last pulse (lu) from the reference group (greek) of the first impulse (If) of the candidate element of the set of candidate elements (Es), and in which for the law of association, the distance (d) is a recovery rate between the reference group (gree ) and the candidate element of the set of candidate elements (Ec) or a score for selecting among the set of candidate elements (Ec) the element sharing the most common characteristics with the reference group ( GREF).
[0011]
11. A method according to any one of the preceding claims, wherein the method comprises, prior to the grouping step (150), a step (120) of rejecting the incoherent pulse trains in terms of the repetition interval of pulses (PRI). 3031257 38
[0012]
12. A method according to any one of the preceding claims, wherein the method comprises before the grouping step (150), a step of rejecting (130) pulse trains whose pulse repetition interval (PRI) ) is greater than a pulse repetition interval threshold (SARI) and whose pulse number (Ii, ..., lm) is less than a pulse number threshold (Sims) -
[0013]
13. A device for deinterleaving radar signals comprising: a receiver (12) capable of receiving electromagnetic signals; a digital signal processing unit (23) able to extract the pulses (li, l) from the signals received; by the receiver (12), and - a readable information medium (26), on which is stored a computer program including program instructions, the computer program being loadable on a data processing unit (24). and adapted to carry out the implementation of a method according to any one of the preceding claims when the computer program is implemented on the data processing unit (24).

类似技术:

---

公开号 | 公开日 | 专利标题

---

EP3241281B1|2018-09-26|Method for the deinterlacing of radar signals

---

EP3049822B1|2017-05-03|Method for non-supervised deinterleaving by n-dimensional enrichment

---

EP2105760A1|2009-09-30|Method and system for tracking emitters

---

FR2999297A1|2014-06-13|PASSIVE LOCATION METHOD OF RADAR TRANSMITTERS

---

EP3047298B1|2017-08-09|Method for detecting targets and associated multifunction radar

---

EP3505957B1|2020-08-05|Treatment method for multi-target detection, characterisation and tracking and associated device

---

WO2014140495A1|2014-09-18|Method for predicting the level of radio field and/or the duration of propagation of cellular telephone networks using high-resolution terrain databases

---

EP3709049B1|2021-06-23|Radar processing system and related noise reduction method

---

EP3491411B1|2021-07-21|Method and device for identifying radar transmission modes

---

EP1229348A1|2002-08-07|Method for detecting and characterising obstacles in a radar system

---

US20210364670A1|2021-11-25|Device, method and computer program product for validating data provided by a rain sensor

---

EP3657200A1|2020-05-27|Detection method by long integration and corresponding radar system

---

WO2020245380A1|2020-12-10|Method for grouping pulses intercepted by a listening system; associated computer program product and readable information medium

---

EP2339764B1|2012-09-12|Method and system for automatic non-cooperative identification of an RF transmission system

---

FR3105441A1|2021-06-25|Method and system for auto-localization from radio waves, corresponding program and program medium.

---

FR3070768A1|2019-03-08|METHOD FOR AUTOMATIC CLASSIFICATION OF A TARGET VESSEL AND RADAR SYSTEM THEREOF

---

WO2020136258A1|2020-07-02|Device for generating a simulated sea-clutter data set, and associated method and computer program

---

FR3081053A1|2019-11-15|METHOD FOR TRACKING TARGETS, IN PARTICULAR MARITIMES, AND RADAR USING SUCH A METHOD

---

FR3050537A1|2017-10-27|PRINCIPLE OF POST-PROCESSING RADAR SIGNALS WITH LOW POWER TRANSMISSION

---

FR3058529A1|2018-05-11|METHOD FOR CONTROLLING THE ELECTROMAGNETIC COMPATIBILITY OF A RADAR DETECTOR WITH AT LEAST ONE IMPULSE SIGNAL EDGE TRANSMITTER

---

FR2892203A1|2007-04-20|Precipitations` accumulation spatialized estimating method for e.g. climatic phenomena study field, involves estimating accumulation of precipitation parts corresponding to small and large scale components from respective data

---

FR2948197A1|2011-01-21|DETECTION OF FAST TARGETS OF SMALL SIZES BY NON-COHERENT POST INTEGRATION USING A SLIDING CLASSIFICATION ALGORITHM

---

FR3000558A1|2014-07-04|METHOD FOR OPTIMIZING THE MONITORING OR SCREENING OF SIGNALS BY A SENSOR

同族专利:

---

公开号 | 公开日

---

US10135488B2|2018-11-20|

---

WO2016107905A1|2016-07-07|

---

FR3031257B1|2018-03-30|

---

US20170366223A1|2017-12-21|

---

EP3241281B1|2018-09-26|

---

EP3241281A1|2017-11-08|

---

ES2703684T3|2019-03-12|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

US20110150053A1|2009-12-17|2011-06-23|Ji Tae Kim|Method of detecting a radar signal, radar detection module, and wireless transceiver including the same|

---

FR2893792B1|2005-11-23|2008-02-15|Commissariat Energie Atomique|ULTRA WIDEBAND TYPE SIGNAL IMPULSE RECEIVER AND ASSOCIATED METHOD|

---

FR3011338A1|2013-09-27|2015-04-03|Thales Sa|METHOD OF UNINTERLACING NOT SUPERVISED BY N DIMENSIONAL ENRICHMENT|

---

US9229095B2|2013-10-28|2016-01-05|Raytheon Company|Adaptive dynamic cluster deinterleaving|FR3054671B1|2016-07-26|2018-07-13|Thales|METHOD AND DEVICE FOR IDENTIFYING RADAR TRANSMISSION MODES|

---

US11156708B2|2017-08-15|2021-10-26|Aviation Communication & Surveillance Systems, Llc|Systems and methods for enhanced reception|

---

US10901065B2|2017-10-11|2021-01-26|The Boeing Company|Location determination of an emitter using frequency-of-arrivalmeasured from a single moving platform|

---

FR3075971B1|2017-12-27|2020-05-29|Avantix|RADAR DETECTION METHOD AND ASSOCIATED DEVICE|

法律状态:
2015-12-31| PLFP| Fee payment|Year of fee payment: 2 |

---

2016-07-01| PLSC| Publication of the preliminary search report|Effective date: 20160701 |

---

2016-12-29| PLFP| Fee payment|Year of fee payment: 3 |

---

2018-01-02| PLFP| Fee payment|Year of fee payment: 4 |

---

2018-12-31| PLFP| Fee payment|Year of fee payment: 5 |

---

2020-10-16| ST| Notification of lapse|Effective date: 20200906 |

优先权:

---

申请号 | 申请日 | 专利标题

---

FR1403062A|FR3031257B1|2014-12-31|2014-12-31|RADAR SIGNAL DISENTING METHOD|

---

FR1403062|2014-12-31|FR1403062A| FR3031257B1|2014-12-31|2014-12-31|RADAR SIGNAL DISENTING METHOD|

---

ES15817426T| ES2703684T3|2014-12-31|2015-12-30|Method of deinterlacing radar signals|

---

EP15817426.8A| EP3241281B1|2014-12-31|2015-12-30|Method for the deinterlacing of radar signals|

---

PCT/EP2015/081423| WO2016107905A1|2014-12-31|2015-12-30|Method for deinterleaving radar signals|

---

US15/540,078| US10135488B2|2014-12-31|2015-12-30|Method for deinterleaving radar signals|