Method for detecting a drone and its remote pilot and system implementing it

Abstract

METHOD FOR DETECTING A DRONE AND ITS REMOTE PILOT AND SYSTEM IMPLEMENTING IT Method (500) for detecting a drone (21) and / or a remote pilot (22) in a monitored area (300), said method comprising a step of acquiring (510) on at least one receiver (111) a plurality of frames of a digital signal (yn, y'n) to be analyzed, said method comprising the following steps: (513) of digital frequency filtering of the digital signal (y'n) in narrow frequency bands (wk) producing a number K of filtered digital signals (zn,k); (521) of estimating the energy levels (E k) of the filtered digital signals (zn,k) obtained in the digital filtering step (513); (523) of constructing a time-frequency state matrix (ME) of the energy levels (E k);(524) of calculating a transition matrix (MT), based on a first order derivative of the state matrix (ME);(525) of calculating an average rate of variation (α) of the transition matrix (MT);and(526) detecting the presence of the drone (21) and / or the remote pilot (22) as a function of the value of the average variation rate (α). Abstract figure: Figure 1;

Description

Title of the invention: Method for detecting a drone and its remote pilot and system implementing it Technical field

[0001] The present invention belongs to the technical field of methods and systems for monitoring secure areas or areas to be secured against potential threats.

[0002] The invention relates more particularly to a method and a system for the detection and localization of small aerial drones, said system being easily transportable, deployable and mobile, in particular to meet specific needs in certain sites.

[0003] The invention finds a direct application in the sector of surveillance of sensitive sites (airports, nuclear power plants, military base camps, detention centers, etc.) or public events (sporting gatherings and open-air musical performances, etc.). State of the art

[0004] The surveillance of sensitive sites, such as airports, nuclear sites such as nuclear power plants, as well as military installations represents a growing challenge in terms of security of infrastructure and people in order to limit and control risks. The same is true during open-air public events in particular, which are increasingly exposed to risks of attack.

[0005] To address these risks, authorities and security forces are continually developing solutions.

[0006] Recently, the emergence of the use of aerial drones, linked to the fact that they are more accessible and more affordable, has opened the way to uses for malicious purposes. More specifically, in recent years, diverted uses (reconnaissance missions, delivery of explosives, targeted attacks, etc.) of consumer drones have been observed, which have the particularity of being small and more difficult to detect. These diverted uses have led to an increased awareness of the potential threats posed by malicious drones, as well as the implementation of reinforced security measures to detect, prevent and counter these attacks.

[0007] As examples, advanced detection systems have been developed to detect the presence of aerial drones in an observed area, such as radars, infrared sensors and surveillance cameras, as well as jamming and neutralization technologies to disable malicious drones.

[0008] On the other hand, and as mentioned previously, small drones are diffi easily detectable by current technologies. To address these threats, one solution would be to analyze the communication signals between drones and remote pilots.

[0009] Document WO2018170739A1 describes a method for monitoring at least one unmanned aerial vehicle, in a predefined area, which consists in particular of generating a datagram containing monitoring data indicating a position of the vehicle and / or a position of a control station in communication with the vehicle.

[0010] Document WO2018067296Al describes in particular a method for detecting, characterizing as well as taking control and / or intercepting unmanned vehicles.

[0011] This method comprises several steps, including in particular scanning a plurality of frequencies of one or more radio frequency (RF) signals, measuring the RF power in one or more frequency bands of the plurality of scanned frequencies, and generating a temporal RF signature based at least in part on the measured RF power. On the other hand, the described method is not effective when several unauthorized vehicles are present in the monitored area. This therefore represents a limitation for the monitoring of sensitive sites.

[0012] Document WO2021191625A1 describes in particular an aerial drone operator detector, comprising a network of radiofrequency antennas, said detector being designed to be installed on board an aircraft. Preferably, the detection method scans several frequency ranges to acquire the signal to be analyzed. The amplitude and the frequency of the detected radiofrequency background are then compared with a database containing frequency signatures of known aerial drones, thus making it possible to identify said detected drone. This method is not, however, capable of detecting and locating aerial drones which are not known in the database used by said method.

[0013] Efforts must then continue to develop detection technologies and strategies for monitoring airspace, for example around a monitored area, and which are particularly suited to new drone use cases. Presentation of the invention

[0014] The present invention aims to overcome all or part of the drawbacks presented previously, and then proposes a solution making it possible in particular to detect the intrusion of one or more small aerial drones in a monitored area. Furthermore, the invention also presents a functionality for locating small aerial drones as well as the remote pilots of said drones.

[0015] The present invention then proposes a method for detecting a drone and / or a remote pilot in a monitored area, said method comprising an acquisition step on at least one receiver of a plurality of frames of a digital signal to be analyzed.

[0016] This method is remarkable in that it comprises the following steps: - digital frequency filtering of the digital signal in narrow frequency bands producing a number K of filtered digital signals; - estimation of the energy levels of the filtered digital signals obtained at the digital filtering stage; - construction of a time-frequency state matrix of energy levels; - calculation of a transition matrix, based on a first-order derivative of the state matrix; - calculation of an average rate of variation of the transition matrix; and - detection of the presence of the drone and / or the remote pilot depending on the value of the average rate of change.

[0017] Advantageously, the method before the step of constructing a time-frequency state matrix of the energy levels, a step of comparing the energy levels with the value of an energy threshold.

[0018] According to a particular characteristic of the invention, the value of the energy threshold is evaluated beforehand and according to a Neyman-Pearson decision principle.

[0019] Advantageously, the step of detecting the presence of the drone and / or the remote pilot is carried out by comparing the average variation rate with respect to a detection threshold.

[0020] According to a particular characteristic of the invention, the value of the detection threshold is evaluated beforehand and according to the Neyman-Pearson decision principle.

[0021] According to another particular characteristic of the invention, the method further comprises, and after the step of detecting the presence of the drone and / or the remote pilot, a step of triggering an alert.

[0022] According to another particular characteristic of the invention, the method further comprises, and after the step of triggering an alert, a step of locating the drone and / or the remote pilot which have been detected.

[0023] Advantageously, the step of locating the drone and / or the remote pilot is based on a multiple signal classification algorithm (MUSIC).

[0024] Advantageously, the step of locating the drone and / or the remote pilot comprises the following consecutive steps: - estimation of a spectrogram of the yn signals; - estimation of the covariance matrix of the spectrogram of the yn signals; - calculation of eigenvalues ​​and eigenvectors of the spectrogram of yn signals; - construction of a vector subspace of the noise; - search for vectors orthogonal to the noise vector subspace.

[0025] Advantageously, the method further comprises a step of updating an operator interface of a system implementing said method.

[0026] Another object of the invention is a system implementing the method for detecting a drone and / or a remote pilot, said system comprising a network of receivers, a control unit.

[0027] Advantageously, the system further comprises at least one surveillance camera.

[0028] Advantageously, the control unit of the system comprises, in whole or in part, an electrical energy storage means, a charge and discharge controller, a transmitter / receiver module, a data acquisition module, a storage memory, a data processor, a digital content display module and a means of interaction with the displayed digital content.

[0029] Finally, the present invention relates to a computer program product comprising a set of program code instructions which, when executed by a processor of a control unit, implement the method for detecting a drone and / or a remote pilot.

[0030] The fundamental concepts of the invention having been set out above in their most elementary form, other details and characteristics will emerge more clearly on reading the description which follows and with reference to the appended drawings. Presentation of the drawings

[0031] The figures are given purely for illustrative purposes for a better understanding of the invention without limiting its scope. The various elements may be represented schematically and are not necessarily on the same scale. Throughout the figures, identical or equivalent elements bear the same numerical reference.

[0032] It is thus illustrated in:

[0033] [Fig.l]: a diagram of an architecture of a system for detecting and locating threats introduced into an area to be monitored, according to an embodiment of the invention;

[0034] [Fig.2]: the main steps of a method for detecting and locating threats present in an area to be monitored, according to one embodiment of the invention;

[0035] [Fig.3]: the main stages of the process of locating threats detected in the area to be monitored;

[0036] [Fig.4]: a diagram of signals analyzed by the threat detection and localization method;

[0037] [Fig.5]: a view of the main window of the operator interface representing the area to thus monitor the detected threats, according to one embodiment of the invention;

[0038] [Fig.6]: a diagram in the form of a diagram of a control unit allowing the implementation of the method, according to an embodiment of the invention. Detailed description of embodiments

[0039] It should be noted that certain technical elements well known to those skilled in the art are described herein to avoid any insufficiency or ambiguity in the understanding of the present invention.

[0040] In the embodiment described below, reference is made to a system for detecting and locating threats in a monitored area.

[0041] More particularly, the invention relates to a system for detecting and locating remote-controlled aerial drones, as well as their remote pilot.

[0042] To do this, the method is based on the analysis of communication signals between a remote pilot and the drone that he controls. Indeed, in their majority, communication between a remote pilot and a drone uses frequency hopping modulation (FHSS). This modulation will then cause amplitude variations in certain frequency bands and generate characteristic time signatures that the method will detect.

[0043] Advantageously, the present invention can be used to detect all communication systems which use frequency hopping modulation, whatever the frequency band used.

[0044] To facilitate the understanding of the present invention, the example of the detection and localization of potential threats (drones and remote pilots) in a sensitive area is used, without this presenting a limit as an application.

[0045] The system implementing the method for detecting and locating potential threats is, in a preferred embodiment, easily transportable, deployable and mobile. In addition, the system is designed to be energy efficient.

[0046] [Fig.l] represents a system 100 for detecting and locating at least one threat 200 present in a monitored zone 300.

[0047] The system 100 mainly comprises a network 11, composed of a plurality of receivers, as well as a control unit 10. The system 100 may also comprise at least one surveillance camera 12 in order to have visual contact with the monitored area 300 and facilitate the analysis of the type of threats 200 to coordinate interventions in response to intrusions.

[0048] In the preferred embodiment of the invention, the receiver network 11 comprises eight receivers 111 with omnidirectional directivities and sensitivities making it possible to measure signals in the frequency bands of 2.4 GHz and / or 5.8 GHz. The network 11 of receivers defines a surveillance perimeter 115, represented by dotted lines. In order to locate the threats 200 which are introduced and / or which are present within the surveillance perimeter 115, the network 11 comprises a number M of receivers 111 greater than or equal to a number D of threats 200.

[0049] In the example shown in [Fig. 1], the threats 200 are remote-controlled aerial drones 21 and their remote pilots 22. Indeed, the invention also relates to a method 500 for detecting and locating threats 200, said method being adapted to other signals emitted by said threats. The different steps of the method 500, according to one embodiment, are shown in [Fig. 2] as well as in [Fig. 3].

[0050] Furthermore, the system 100 and the method 500 are particularly suitable for the detection, but also for the localization, of drones 21 considered to be small, as defined for example in the document US Department of Defense, US Army Roadmap for UAS 2010-2035, Table 2-2, page 12, which classifies drones according to three criteria. First of all, a drone is considered to be small if its maximum takeoff weight is less than 9 kg. Then, these drones can fly up to a maximum altitude above the ground of 1200 feet. Finally, they move at a maximum speed of 100 knots.

[0051] To complete these three criteria, it may also be agreed to consider that a drone is said to be small when its dimensions (length, width, height) are all less than 50 cm. Small drones are currently difficult to detect because the methods and systems (active radar for example) used are not sufficiently precise when said drones have small dimensions. In other words, and for active radars, the radar equivalent surface of small drones is not large enough to be detected.

[0052] In [Fig.l], the network 11 of receivers 111, the surveillance cameras 12 and the control unit 10 are represented by means of symbols, which are respectively dots, crosses and a star.

[0053] In a particular embodiment of the invention, the signals coming from the receivers 111, and the images coming from the surveillance cameras 12, are transmitted to the control unit 10 located in the area to be monitored 300, or located close to said area, in other words on the ground, whether for example inside a building or a temporary installation.

[0054] The main functionality of the control unit 10 is to acquire and analyze the signals coming from the receivers 111 to detect, for example, drones 21.

[0055] In a particular operating mode, the control unit 10 also makes it possible to locate the drones 21 and their remote pilot 22.

[0056] In a particular embodiment, the control unit 10 is transportable. easily by an operator. The unit is for example integrated into a rigid case, which can be waterproof. This case protects its contents from shocks. Thus, the threat analysis is carried out as close as possible to the area to be monitored 300, sensitive buildings 30 as well as sensitive infrastructures 31.

[0057] In another embodiment of the invention, the signals from the receivers 111 and the images from the surveillance cameras 12 are transmitted and analyzed in a surveillance unit 900 remote from the monitored area 300. The transmission of the signals to be analyzed is carried out via a computer network 13, wired (optical fiber, etc.) or wireless (satellite, radio, etc.). The surveillance unit 900 may be a building, a mobile unit (automobile, ship, etc.), and is capable of implementing the steps of the method 500.

[0058] Furthermore, the control unit 10, when positioned in the monitoring zone 300, is able to communicate with operators located remotely, in the monitoring unit 900 for example.

[0059] [Fig.2] represents the main steps of the method 500 for detecting and locating intrusion of threats in the monitored zone 300, according to one embodiment.

[0060] The method 500 mainly comprises: - a step 510 of acquiring the signals y, to be analyzed; - a step 520 of analyzing the acquired signals zn,k; - a step 530 of triggering an alert of the presence of at least one threat 200; - a step 540 of locating the detected threat; and - a step 550 for updating the operator interface.

[0061] More precisely, step 510 of acquiring the signals y, to be analyzed, is carried out simultaneously on the M receivers 111 of the network 11, and comprises: - a step 511 of converting the analog signals y„ into digital signals yn, by means of an analog / digital converter of the control unit 10; - a step 512 of digital frequency filtering of the yn signals, by means of a wideband H filter IV; and - a step 513 of digital frequency filtering of the signals y by means of narrow band filters wk.

[0062] In the preferred embodiment of the invention, during the acquisition step 510, the signals yt are subdivided into a number N of one hundred frames, said frames having a duration T of the order of 20 ms, without these parameters presenting any limits to the invention.

[0063] Step 512 of digital frequency filtering, using the wideband H filter W, has for the purpose of selecting the frequency band of interest on which the analysis is carried out. In one embodiment of the invention, the filtering step 512 targets the frequency bands IV with minimum frequencies fmin of 2.4 GHz as well as 5.8 GHz and a width of 80 MHz.

[0064] During the digital frequency filtering step 512, the value of a detection threshold adetection is also calculated and updated at regular intervals as a function of a noise level of the monitored zone 300. The value of the detection threshold adetection is obtained by carrying out a statistical analysis of the signals y„, as well as by evaluating the energy level of the noise.

[0065] Step 513 of digital frequency filtering consists of filtering the signals on the frequency bands containing the pulses of the communications between the remote pilot 22 and the drone 21.

[0066] In one embodiment, the digital frequency filtering step 513 is carried out using narrow band filters hk wk with a width of 400 kHz. Thus, at the end of step 513, we then obtain for one of the N frames, a number K of two hundred frequency sub-bands, representing KxN time signals zn,k- which are analyzed in the following step of the method 500.

[0067] Obviously, according to other embodiments, the minimum frequencies fmin, the width IV of the frequency band of the wideband filter H, as well as the width wk of the frequency band of the narrowband filter hk, can be modified in order to adapt the steps 512 and 513 of digital frequency filtering to other modulations by frequency hopping and / or communication protocol.

[0068] Step 520 of analyzing the filtered zn,k signals mainly comprises: - a step 521 of estimating the energy levels Ek of the signals z„,k; - a step 522 of comparing the energy levels Ek with a threshold value kenergy energy; - a step 523 of constructing a time-frequency state matrix of the energy levels Ek; - a step 524 of calculating a transition matrix; - a step 525 of calculating an average variation rate of the acquired signals zn,k ; And - a step 526 of detecting the presence of a drone and / or its remote pilot, by comparing the average variation rate a with respect to the detection threshold a detection*

[0069] Step 521 of estimating the energy levels Ek is applied to each signal zn,k by means of an energy estimator E, the energy level Ek on the Nk samples of the signal zn,k can be evaluated by means of the following formula:

[0070] [Math.l]

[0071] The method for estimating the energy levels Ek described just now does not present a limit to the invention. Indeed, other methods such as the calculation of autocorrelation of the signal or the analysis of the cyclostationarity of the signals can be used during step 521 of estimating the energy levels Ek.

[0072] At the end of the estimation of the energy level Ek of the signals zn,t, these levels are compared during step 522 with the value of the energy threshold Xémrgie, said threshold being determined beforehand and being evaluated according to the Neyman-Pearson decision principle, as a function of a probability of false alarm. Advantageously, the Neyman-Pearson principle (notably described in the document KD Tocher, Extension of the Neyman-Pearson Theory of Tests to Discontinuons Variâtes, Biometrika, 37, June 1950, pages 130-144) facilitates decision-making, in comparison with other techniques which may be more complex to use, without however improving the reliability of the prediction.

[0073] Other decision principles can also be used, such as Fisher's principle.

[0074] Furthermore, the value of the energy threshold ^energy is re-evaluated at a frequency defined according to the characteristics specific to the monitored zone 300 (sensitivity of the zone, type and time slots of activity, etc.).

[0075] In a particular embodiment of the invention, the Constant False Alarm Rate (CFAR) algorithm is used to estimate the value of the energy threshold ^energy-

[0076] Thus, if the energy level Ek is greater than the value of the energy threshold, this means that the sub-band wk is used in the communication channel between the remote pilot 22 and the drone 21. A binary value of 1 is then assigned to the sub-band wk.

[0077] On the other hand, if the energy level Ek is less than or equal to the value of the energy threshold, this means that the sub-band wk is not used in the communication channel between the remote pilot 22 and the drone 21. A binary value of 0 is then assigned to the sub-band wk.

[0078] Step 523 then consists of constructing a time-frequency state matrix ME of the energy levels Ek obtained at the end of steps 521 and 522 which were carried out for each of the N frames of the signals zn,k.

[0079] The time-frequency state matrix ME is then a binary matrix of size KxN, each column of which is called state vector Xk.

[0080] From the state matrix ME, a transition matrix MT is calculated during step 524. Step 524 of calculating the transition matrix MT consists of evaluating a norm (pk of the difference between the state vector Xk+] and the state vector Xk, as follows:

[0081] [Math.2] ^4-PU+i ' Xj

[0082] At the end of step 524, the transition matrix MT obtained is of dimension Kx (Nl).

[0083] Finally, step 525 of calculating the average variation rate a consists of evaluating a probability of the presence of communication between a remote pilot 22 and a drone 21.

[0084] The average rate of change a is evaluated using the following formula:

[0085] [Math.3]

[0086] Finally, during step 526, the average variation rate a is compared to a detection threshold a, determined beforehand according to the Neyman-Pearson decision principle.

[0087] Advantageously, the value of the detection threshold is constantly re-evaluated at a frequency, which, as for the energy threshold Xenergy, is a function of the characteristics of the zone to be monitored 300.

[0088] Furthermore, when the system 100 is powered up, all of the threshold values ​​used to determine the presence of drones 21 or remote pilots 22 in the monitored zone 300 are updated.

[0089] When the average variation rate a is less than or equal to the value of the detection threshold adetection, no threat is present in the monitored area 300, and the method 500 restarts signal acquisition to continue monitoring.

[0090] When the average variation rate a is greater than the value of the detection threshold a detection, an alert of the presence of a threat is triggered during step 530.

[0091] In both cases, an operator interface 600, as shown in [Fig.5], is updated during step 550.

[0092] Once a threat 200 has been detected in the monitored area, it is located during step 540.

[0093] [Fig.3] represents the main steps of the method 500, the aim of which is to locate the threats that were detected at the end of steps 510 to 530. These steps are based on the MUSIC (Multiple Signal Classification) algorithm described in the document Statistical Digital Signal Processing and Modeling (Hayes, 1996), with modifications to use the algorithm in the time-frequency domain, and not only in the time domain, as was initially the case.

[0094] Indeed, analysis in the time and frequency domain provides greater precision complementary to locate the drones 21 and the remote pilots 22. The temporal analysis generates two possible results to locate the same and unique drone 21 or remote pilot 22, due to the complementarity of the angles.

[0095] In other words, in time, we obtain two locations for a target, whereas in time-frequency we obtain only one location.

[0096] Thus, it is possible to know the positions of drones and remote pilots in real time.

[0097] Step 540 of locating a threat that has been detected mainly comprises: - a step 541 of estimating a spectrogram of the yn signals; - a step 542 of estimating the covariance matrix of the spectrogram of the yn signals; - a step 543 of calculating the eigenvalues ​​and vectors of the spectrogram of the yn signals; - a step 544 of construction of the noise vector subspace; - a step 545 of searching for vectors orthogonal to the vector subspace noise.

[0098] At the end of steps 541 to 545, the coordinates of the source of transmission of the communication signal are determined. In other words, step 540 of locating a threat 200 determines its azimuth (p), its zenith 0 as well as its radius q, relative to a reference center of a spherical coordinate system.

[0099] The reference center is defined by a receiver 111 of the network 11 or by a particular point located within the perimeter defined by the network 11 of receivers.

[0100] From the estimation of the azimuth (p, of the zenith 0 as well as the radius q, and in relation to the geolocated reference center, the latitude l, the longitude L and the altitude A of the threat are calculated.

[0101] Finally, the threat 200 is classified according to at least two types of threat, namely a drone 21 or a remote pilot 22.

[0102] This classification is carried out based on the altitude as well as the speed of movement of the threat.

[0103] For example, if the altitude A of the threat is more than 3 m from the ground, said threat is considered to be a drone 21. On the other hand, if the altitude of the threat is close to the ground (in other words at human height), and said threat only moves at a low speed (for example slow / fast walking), said threat is considered to be a remote pilot 22.

[0104] As mentioned in the introduction to the description of the invention, the presence of an aerial drone 21 in the surveillance perimeter 115 is deduced from the identification in the acquired signals y(t), of the presence of a radiofrequency communication between said drone and a remote pilot 22, said remote pilot being able to be present or not within said perimeter.

[0105] In other embodiments of the method 500, the presence of a threat 200 may be based on variations in the energy Ek contained in frequency bands of interest, these variations being able to be due, for example, to modulations in amplitude and phase (scintillation) or in frequency (Doppler effect) of the acoustic or electromagnetic emission signal emitted by said threat.

[0106] [Fig.4] schematically represents an example of data 400 processed by the method for detecting and locating threats, and this at different stages of said method.

[0107] The data 400a represents the time signal yn acquired by a receiver 111 of the system 100, said signal coming from the step 511 of analog / digital conversion of the analog signals yt. In addition, it is possible to see the segmentation into N frames of the signal yn, each of the frames having a duration T.

[0108] The data 400b and 400c respectively represent the frequency compositions 410b and 410c of the frames t1 and t5 on the frequency band W, after the wideband digital filtering step 512 H has been carried out. As mentioned previously, a second narrowband digital frequency filtering step 513 hk is then applied to the signal in order to obtain the energy Ek contained in each of the sub-bands wk during the step 521.

[0109] Once step 522 of comparing the energy levels Ek with the value of the energy threshold Xenergy has been carried out, the data 400d, representing the time-frequency state matrix ME of the energy levels Ek is obtained at the end of step 523, said matrix containing N state vectors Xk. From the time-frequency state matrix ME, the transition matrix MT is calculated during step 524, as described previously, and we obtain the data 400e which is then used to calculate, during step 525, the average variation rate a represented by the data 400f.

[0110] Once the threat detection and localization method 500 has been carried out, the results are displayed on the operator interface 600 shown in [Fig.5].

[0111] The operator interface 600 mainly comprises a name 60 of the monitored area 300, a box 64 comprising a map 641 of said area and on which markers 643 are positioned in order to locate in particular the surveillance cameras 12 as well as the threats 200 which have been detected. In a particular embodiment, when a threat is detected, the box 64 comprises a frame 642 (represented in dotted lines in [Fig. 5]) which can for example flash and become red in order to attract the attention of the operator. Similarly, the operator interface 600 comprises a pictogram 61 which can flash and become red when a threat is present in the monitored area 300.

[0112] The frame 642 and the pictogram 61 are for example green in color when no threat is detected in the monitored zone 300.

[0113] The operator interface 600 also includes a drop-down menu 63 for selecting the wideband IV frequency band for which the operator wishes to have a situation analysis.

[0114] Finally, the markers 643 are clickable to display specific information.

[0115] As examples, when the marker 643m corresponds to that of a threat (icon with the exclamation point in a triangle), a pop-up window 62a-b is displayed.

[0116] The pop-up window 62a-b includes a pictogram 621a-b to identify the type of threat (drone, remote pilot, etc.) as well as geolocation information 622a-b.

[0117] The same applies to the markers 643c, in the form of a cross, corresponding to the surveillance cameras 12 present on the map 64 of the monitored area 300. A pop-up window 65 is displayed by clicking on one of the markers 643c, said window comprising a pictogram 651 and a numbering 652 facilitating the identification of the type of element which is analyzed, as well as a video stream 653.

[0118] In another particular embodiment, the position of potential threats is represented by means of a color map, superimposed on the plan of the monitored area.

[0119] In one embodiment, the operator interface 600 is an integral part of the control unit 10.

[0120] [Fig.6] represents a functional diagram of the control unit 10 allowing the implementation of the method 500 for detecting and locating threats, said unit comprising in the preferred embodiment: - an electrical energy storage means 108; - a charge and discharge controller 106; - a transmitter / receiver module 103 allowing data to be communicated with third-party entities; - a data acquisition module 104 receiving the analog signals y, coming from the receivers 111 and converting said analog signals into digital signals yn; - a memory 101 for storing program instructions as well as data; - a data processor 102 capable of executing at least one step of the method 500 for detecting and locating a threat; - a display module 105 for digital content (touch screen, non-touch screen, etc.) on which, among other things, the operator interface 600 is displayed; and - a means of interaction 107 with the displayed digital content (button, keyboard, pointing device, etc.).

[0121] In a particular embodiment, the electrical energy storage means 107 can easily be changed by another equivalent storage means, in order to have greater energy autonomy when using the control unit.

[0122] In another embodiment of the invention, the electrical energy storage means is rechargeable by means of solar panels.

[0123] These particular embodiments are given as examples, and do not represent limits to the present invention. Indeed, to implement the method 500 for detecting and locating threats, those skilled in the art have the ability to use control units 10 having functionalities that are all or partly identical to those described previously.

[0124] For example, in a particular embodiment, the control unit 10 does not include a wireless transmission module so as not to be itself detectable by means of passive methods.

[0125] Preferably, the control unit 10 is a system including both the hardware part and the software part for detecting and locating the drones 21 and the remote pilots 22.

Claims

Claims

1. Method for detecting a drone (21) and / or a remote pilot (22) in a monitored area (300), said method comprising a step of acquiring (510) on at least one receiver (111) a plurality of frames of a digital signal (ym y'n) to be analyzed, said method being characterized in that it comprises the following steps: - (513) of digital frequency filtering of the digital signal (y 'n) in narrow frequency bands (wO producing a number K of filtered digital signals (zn,t); - (521) of estimating the energy levels (¾) of the filtered digital signals (zn,ù obtained in the digital filtering step (513); - (523) of constructing a time-frequency state matrix (ME) of the energy levels (Ek); - (524) of calculating a transition matrix (MT), based on a first-order derivative of the state matrix (ME); - (525) of calculating an average rate of variation (a) of the transition matrix (MT);and - (526) detecting the presence of the drone (21) and / or the remote pilot (22) as a function of the value of the average variation rate (a).;

2. Method according to claim 1, further comprising, and before the step (523) of constructing a time-frequency state matrix (ME) of the energy levels (Ek), a step (522) of comparing the energy levels (E^ with the value of an energy threshold (Xenergy).

3. Method according to claim 2, in which the value of the energy threshold (Xenergy) is evaluated beforehand and according to a Neyman-Pearson decision principle.

4. Method according to any one of the preceding claims, in which the step (526) of detecting the presence of the drone (21) and / or the remote pilot (22) is carried out by comparing the average variation rate (a) with respect to a detection threshold (adetection).

5. Method according to claim 4, in which the value of the detection threshold (adetectim) is evaluated beforehand and according to the Neyman-Pearson decision principle.

6. Method according to any one of the preceding claims, further comprising, and after the step (526) of detecting the presence of the drone (21) and / or the remote pilot (22), a step (530) of triggering an alert.

7. Method according to any one of the preceding claims, further comprising, and after the step (530) of triggering an alert, a step (540) of locating the drone (21) and / or the remote pilot (22) which have been detected.

8. Method according to claim 7, in which the step (540) of locating the drone (21) and / or the remote pilot (22) is based on a multiple signal classification algorithm (MUSIC).

9. Method according to claim 8, in which the step (540) of locating the drone (21) and / or the remote pilot (22) comprises the following consecutive steps: - (541) of estimating a spectrogram of the yn signals; - (542) of estimating the covariance matrix of the spectrogram of the yn signals; - (543) of calculating eigenvalues ​​and eigenvectors of the spectrogram of the yn signals; - (544) of constructing a vector subspace of the noise; - (545) of searching for vectors orthogonal to the vector subspace of the noise.

10. A method according to any preceding claim, further comprising a step (550) of updating an operator interface (600) of a system (100).

11. System (100) implementing a method (500) for detecting a drone (21) and / or a remote pilot (22), according to any one of claims 1 to 10, characterized in that said system comprises a network (11) of receivers (111), a control unit (10).

12. The system (100) of claim 11, further comprising at least one surveillance camera (12).

13. System (100) according to claim 11 or claim 12, in which the control unit (10) comprises, wholly or in part, an electrical energy storage means (108), a charge and discharge controller (106), a transmitter / receiver module (103), a data acquisition module (104), a storage memory (101), a data processor (102), a digital content display module (105) and a means of interaction (107) with the displayed digital content.

14. Computer program product characterized in that it comprises a set of program code instructions which, when executed by a processor (102) of a control unit (10) of a system (100), implement a method (500) for detecting a drone (21) and / or a remote pilot (22) according to one of claims 1 to 10.

Images