Electromagnetic leakage imaging system in a faradized structure and associated method

Abstract

------ Electromagnetic leakage imaging system in a faradized structure and associated method The invention relates to an electromagnetic leakage imaging system (1) in a faradized structure (2), the system (1) comprising an electromagnetic wave receiver (7) and an electromagnetic wave source (3) comprising at least one transmitting antenna (4), a pseudo-random noise generator (51), a frequency modulator and a transmitting signal processing circuit configured so that at least one transmitting antenna (4) emits a pseudo-random noise sequence transposed onto a set of predetermined carrier frequencies, said pseudo-random noise sequence being known to the electromagnetic wave receiver (7),The electromagnetic wave receiver (7) comprises a wideband receiving array of N antennas (8) and a receiving signal processing circuit (10) connected to the receiving antenna array (8) and configured to construct an image from the correlated demodulated signals identifying the position of electromagnetic leakage in the faradized structure (2). Figure to be published with the abbreviation: Figure 1,

Description

Title of the invention: Electromagnetic leakage imaging system in a faradized structure and associated method

[0001] The present invention relates to the field of security, and more particularly to an electromagnetic leakage imaging system in a faradized structure and to an associated method.

[0002] In the context of cybersecurity, electromagnetic shielding is used to protect sensitive sites (e.g., government buildings, data centers, etc.) in order to prevent compromising emissions, or leaks of confidential information, which could cause serious security incidents (espionage, hacking), see for example the publication "Characterization to a TEMPEST testing laboratory and methodology for control to compromising emanation", S. Sebastiani, 1998 IEEE EMC Symposium, International Symposium on Electromagnetic Compatibility, Denver, CO, USA, 1998, pp. 165-170 vol.l.

[0003] Based in particular on the TEMPEST standard (acronym for Telecommunications Electronics Materials Protected from Emanating Spurious Transmissions) of the US National Security Agency (NSA), studies have been carried out on the implementation of faradization techniques to ensure the security of sensitive buildings, but also on the measurement of the quality of the shielding, see for example the publications "Effectiveness of shielded window films used for electromagnetic leakages in communication security", HS Efendioglu, M. Yabuloglu and H. Ozer, 2017 IV International Electromagnetic Compatibility Conference (EMC Turkiye), Ankara, 2017, pp.1-4, and "Estimation of minimum attenuation level for a TEMPEST shielded enclosure", M. Popescu, V. Bîndar, R. Craciunescu and O. Fratu, 2016 International Conference on Communications (COMM), Bucharest, 2016, pp. 513-518.

[0004] Estimating the effectiveness of armoring is indeed essential to guarantee the safety of sensitive structures or buildings. This is based on specific standards in the military or civilian fields (GAMT20, MIL-STD-285, AECTP 510, IEEE-299, EN50147-1, IEC 61587-3:2006, IEC 60917 and IEC 60297).

[0005] However, diagnosing shielding effectiveness at the building scale remains difficult due to the range of known shielding measurement systems that only provide a scalar value of attenuation (internal generator, coupled with an external receiver for intensity measurement), and generally require spatial zoning, as described in the publication "Considerations regarding shielding effectiveness and testing of electromagnetic protected enclosures used in communications security," V. Bîndar, M. Popescu and A. Vulpe, 2014, 1st International Conference on Communications (COMM), Bucharest, 2014, pp. 1-6. To limit the number of measurements required, estimating shielding effectiveness and visualizing the location of defects using imaging are therefore essential..

[0006] The objective of the invention is to propose a system and a method of measurement complementary to existing standardized measurements, capable of detecting, identifying and locating the origin of possible leaks in structures, in particular faradized structures.

[0007] The approach developed in the present invention is based on the principle of interferometric imaging derived from the field of synthetic aperture interferometric radiometry (SAIR, acronym for Synthetic Aperture Interferometric Radiometer).

[0008] The SAIR technique was initially used in various applications such as Earth observation and security scanners, to build thermal cameras operating at millimeter wave frequencies. This approach aims to image the distribution of incoherent electromagnetic sources of thermal origin. The SAIR technique makes it possible to obtain high-resolution imaging with a reduced number of antennas over a wide field of view, because it can synthesize a large virtual aperture from sparsely distributed physical antennas.

[0009] The invention is based on the application of the SAIR imaging principle for the detection and localization of electromagnetic leaks by imaging. Indeed, a faradized structure excited by a broadband source behaves like an oversized cavity within which the amplitude of the electric field temporally follows a statistical Rice distribution, with an exponential decay related to the cavity quality factor (see the publication "Electromagnetic fields in cavities: Deterministic and statistical theories," D.A. Hill, in IEEE Antennas and Propagation Magazine, vol. 56, no. 1, pp. 306-306, Feb. 2014). The potential openings (electromagnetic leaks) will therefore radiate fields spatially uncorrelated. Their detection is therefore similar to a passive radiometric imaging problem.

[0010] The publication “Synthetic aperture interferometric imaging using a passive microwave coding device”, Kpré et al., 2016 IEEE Conference on Antenna Measurements & Applications (CAMA), DOI: 10.1109 / CAMA.2016.7815738, discloses the use of an interferometric radiometer equipped with an N-to-1 switch, with a classic SAIR configuration in which the noise source is an arbitrary noise source requiring the use of two separate receivers respectively associated with two separate switches to achieve simultaneous correlation between each pair of antennas, so that, for an N-antenna array, N2 switching is required, which represents a significant acquisition time.

[0011] The present invention aims to overcome the drawbacks of the prior art and therefore relates to an electromagnetic leakage imaging system in a faradized structure, the system comprising an electromagnetic wave source and an electromagnetic wave receiver, characterized in that the electromagnetic wave source is configured to be disposed within the faradized structure, the electromagnetic wave receiver being configured to be disposed outside and facing the faradized structure, and the electromagnetic wave source comprises a pseudo-random noise generator, an emission signal processing circuit, a frequency modulator and at least one transmission antenna, configured so that the at least one transmission antenna emits a pseudo-random noise sequence transposed onto a set of predetermined carrier frequencies,Given that the pseudo-random noise sequence is known to the electromagnetic wave receiver, the electromagnetic wave receiver comprises an array of N broadband receiving antennas and a receiving signal processing circuit connected to the receiving antenna array and configured to sequentially switch the N receiving antennas, to demodulate the signals received from the N receiving antennas at the frequencies of the predetermined carrier frequency set, to correlate the demodulated signals with the pseudo-random noise sequence, and to construct an image from the cross-correlation, for each pair of receiving antennas, of the correlated demodulated signals, the resulting image identifying the position of electromagnetic leakage in the faradized structure.

[0012] The faradized structure can be any structure, including an enclosure, a building or part of a building.

[0013] The use by the system of a pseudo-random noise sequence allows for a non-cooperative architecture, i.e. without the need for synchronization with the transmission system between the source of electromagnetic waves and the receiver of electromagnetic waves.

[0014] In the invention, the modulating signal is baseband and modulates in amplitude and phase a carrier frequency fp in the set of predetermined carrier frequencies.

[0015] The electromagnetic leakage imaging system of the invention makes it possible to complete the existing product range by combining shielding attenuation measurement with localization of electromagnetic leakage sources by imaging.

[0016] The reconstruction of maps of potential leaks (faradization defects) is ensured by a multi-antenna reception associated with passive imaging algorithms.

[0017] The use of a non-cooperative noise source (i.e., the emitted noise signal has the correlation properties of white noise and is known to the receiver) allows the use of only one N-to-1 switch, unlike a conventional SAIR configuration. With the use of the non-cooperative noise source, the capture of signals from each pair of antennas does not need to be performed simultaneously, so that, for an N-antenna array, N switches are required instead of N2 in the conventional SAIR configuration, thus minimizing acquisition time and cost.

[0018] At the electromagnetic wave receiver level, the first correlation of each receiving antenna signal by the known pseudo-random noise sequence allows the complex information associated with each receiving antenna to be extracted, while the intercorrelation of signals between each pair of receiving antennas in the imaging process allows the visibility matrix (synthetic aperture) to be constructed.

[0019] Since SAIR imaging is based on a passive initial approach, it offers several major advantages in the context of the invention:

[0020] - it is not necessary to synchronize the electromagnetic wave receiver with The source of electromagnetic waves: the electromagnetic wave receiver may or may not be synchronized with the source of electromagnetic waves. The idea is to place one (or more) independent electromagnetic source(s) inside the faradized structure;

[0021] - noise correlation will allow good sensitivities to be achieved as a function of the integration time (compression gain).

[0022] In practice, a faradized structure is preferably used, which corresponds to a cavity oversized relative to the wavelength (see the aforementioned publication). "Electromagnetic fields in cavities: Deterministic and statistical theories", DA Hill, in IEEE Antennas and Propagation Magazine, vol. 56, no. 1, pp. 306-306, Feb. 2014).

[0023] In practice for a building-type structure, the wavelength of the electromagnetic waves emitted by the electromagnetic wave source is in the decimeter or centimeter wavelength band, preferably in the 100MHz to 10GHz wavelength band, relative to the dimensions of a building or a room in a building, on the order of a meter.

[0024] Depending on the number of electromagnetic leakage sources, certain wavelengths will be preferred according to the dimensions of the faradized structure whose leakage we seek to detect.

[0025] Thus, for several sources of electromagnetic leakage, the wavelength of the electromagnetic waves emitted by the sources of electromagnetic waves will preferably be at least three times smaller, preferably at least ten times smaller, than the dimensions of the faradized structure whose leakage we seek to detect.

[0026] For a single source of electromagnetic leakage, the system and method of the invention can function to image an electromagnetic leak, whether the wavelength of the electromagnetic waves is small or large compared to the dimensions of the Faraday cage whose leakage is to be detected. Even if the Faraday cage is small compared to the wavelength, the system of the invention will be able to detect it.

[0027] Potential leaks, distributed throughout this cavity, can therefore be considered as temporally and spatially incoherent sources. The principle of SAIR imaging is thus perfectly compatible for the detection and imaging of these sources.

[0028] In this type of configuration, it is well known that the impulse response between two distant points (in the case of the invention, between an electromagnetic wave source placed inside the Faraday cage and a leakage source) has a Rice distribution in line-of-sight (LOS) and a Rayleigh distribution in non-line-of-sight (NLOS). In an NLOS case, for example, this response is therefore equivalent to Gaussian white noise with exponential decay over time, with a decay time (Te) proportional to the composite quality factor of the cavity (Q = 2ir.f.Tc). The degree of decorrelation between sources is proportional to the product Band x Te, where Band represents the bandwidth of the signal emitted by the electromagnetic wave source.

[0029] If the quality factor decreases (possible losses), Te decreases, resulting in a greater correlation between sources of leakage.

[0030] To compensate for the decrease in Te, it is necessary to increase the bandwidth.

[0031] Since an instantaneous broadband system is complex and expensive, it is therefore preferable to use a narrowband frequency diversity system.

[0032] This is possible by considering that the frequency sum of the visibility matrices (correlations between antennas) is equivalent to the broadband visibility matrix (this is valid for Gaussian noise, but this approximation is made here).

[0033] The invention provides a narrowband reception architecture system with frequency diversity, different from those used in the field of radiometry, making it possible to do without the implementation of a more complex and more expensive broadband system.

[0034] The system of the invention thus makes it possible to evaluate the quality of faradization and to locate electromagnetic leaks of electronic boxes or buildings in the field of EMC (Electromagnetic Compatibility), in centimeter band (500MHz-10 GHz).

[0035] According to one embodiment, the transmit signal processing circuit includes computing and memory circuits and, in series after the pseudo-random noise sequence generator, a digital filter configured to filter the pseudo-random noise sequence, an IQ modulator configured to successively modulate in phase and amplitude the set of predetermined carrier frequencies using the filtered pseudo-random noise sequence, and optionally a signal amplifier before at least one transmit antenna.

[0036] According to one embodiment, the receiving signal processing circuit comprises computing and memory circuits and an N-way to 1-way RF switch, the N channels of the RF switch being respectively connected to the N receiving antennas, and, in series after the single output channel of the RF switch, a processing chain comprising an IQ demodulator configured to demodulate in phase and amplitude the set of predetermined carrier frequencies using the known pseudo-random noise sequence, at least one low-pass filter, at least one analog-to-digital converter, with optionally at the output of the RF switch upstream of the processing chain, at least one of a signal amplifier and a band-pass filter.

[0037] According to one embodiment, the receiving antenna array comprises N antennas arranged in one of the following configurations: a cross configuration, a Y configuration, a spiral configuration, and a random configuration. The antennas are preferably located in the same plane, configured for vertical operation and facing the Faraday cage whose electromagnetic leakage is to be detected.

[0038] According to one embodiment, the receiving antenna array comprises N antennas arranged on a plane configured to be vertical in use.

[0039] The distance between antennas affects the field of view (FOV).

[0040] The relationship between the separation distance of the antennas (Dant) and the FOV is generally defined as follows: Dant=R*lambda / FOV, with R representing the depth distance between the antenna plane and the observation plane, lambda representing the wavelength of the maximum frequency fp, FOV representing the visibility window (maximum observable area without aliasing).

[0041] Once Dant is fixed (to reach a given visibility area), the number of antennas to adjust the resolution can be set approximately as follows:

[0042] Average_resolution = R*average_lambda / Dr, where average_lambda represents the wavelength of the center frequency, and Dr represents the total size of the virtual antenna array (synthetic aperture obtained after correlation between each pair of antennas). In the case of the cross array, the size of the synthetic aperture is equivalent to that of the physical array.

[0043] According to one embodiment, at least one transmitting antenna transmits in the centimeter band.

[0044] According to one embodiment, the N-channel RF switch is controlled by a clock frequency configured for the capture, by the N-channel RF switch, of at least three pseudo-random noise sequence frames per receiving antenna.

[0045] According to one embodiment, the electromagnetic wave source includes a self-contained power supply battery, which eliminates the need to power the electromagnetic wave source from outside the shielded structure. The electromagnetic wave source can also be powered by a power cord connected to the mains.

[0046] According to one embodiment, the set of predetermined carrier frequencies comprises discrete frequencies between 500MHz and 10 GHz.

[0047] In a non-limiting example, the set of predetermined carrier frequencies comprises discrete frequencies between 2.5 and 3.5 GHz, spaced 50 MHz apart.

[0048] The spacing may be smaller or larger, for example 25 MHz or 100 MHz, without departing from the scope of the present invention.

[0049] Preferably, a bandwidth of the transmitted signal lower than the coherence band of the channel is chosen to avoid interference between symbols (in time: symbol time > decay time of the enclosure Te).

[0050] Once this bandwidth is fixed, the total bandwidth is preferably swept with contiguous sub-bands (equal to the signal bandwidth), to optimize frequency diversity.

[0051] According to one embodiment, the computing circuits are one or more of a microcontroller, a processor, a microprocessor, a digital signal processor, DSP, a programmable pre-diffused matrix, FPGA, an application-specific integrated component, ASIC, a computer.

[0052] The invention also relates to a method for imaging electromagnetic leakage in a faradized structure using an electromagnetic leakage imaging system as described above, characterized in that it comprises:

[0053] - to place the electromagnetic wave source in the faradized structure;

[0054] - position the electromagnetic wave receiver opposite the structure faradized;

[0055] - emit, through the electromagnetic wave source, a pseudo-noise sequence random transposed onto a set of predetermined carrier frequencies;

[0056] - to receive, via the electromagnetic wave receiver, the electromagnetic waves emitted by the source of electromagnetic waves;

[0057] - sequentially switch the N receiving antennas of the wave receiver electromagnetic, demodulate at the frequencies of the set of predetermined carrier frequencies the signals received from the N receiving antennas, correlate the demodulated signals with said pseudo-random noise sequence and construct an image from the cross-correlation, for each pair of receiving antennas, of the correlated demodulated signals, the resulting image identifying the position of electromagnetic leaks in the faradized structure.

[0058] To better illustrate the object of the present invention, preferred embodiments will now be described, in conjunction with the accompanying drawings.

[0059] On these drawings:

[0060] [Fig.1] is a schematic view of an electromagnetic leakage imaging system according to the invention.

[0061] [Fig.2] is a schematic view of an electromagnetic wave receiver of the system of [Fig.1].

[0062] [Fig.3] is a general diagram of the principle of interferometric aperture synthesis.

[0063] [Fig.4] is a circuit diagram of the signal processing circuit for the transmission of the source of electromagnetic waves of the system of the [Fig.l].

[0064] [Fig.5] is a representation of a time frame containing a pseudo-random noise sequence.

[0065] [Fig.6] is a representation of the time frame containing the pseudo-random noise sequence emitted by the electromagnetic wave source of the system of [Fig.1].

[0066] [Fig.7] is a circuit diagram of the receiver signal processing circuit of the electromagnetic wave receiver of the system of the [Fig.l].

[0067] [Fig.8] represents an organizational chart of the operation of the system of the invention.

[0068] [Fig.9a] represents an implementation of the system of the invention.

[0069] [Fig.9b] represents the electromagnetic leakage detected on the implementation of the [Fig.9a]

[0070] Referring to [Fig. 1], one can see that an electromagnetic leakage imaging system 1 according to the invention is shown, implemented on a faradized structure 2, symbolized by a cube. The faradized structure 2 can be any structure, for example a building or part of a building, or even an enclosure or a shielded box, without the invention being limited in this respect.

[0071] The system 1 comprises an electromagnetic wave source 3, described in more detail below, including an antenna 4 connected to a transmitting signal processing circuit 5 by a wire 6. Although the transmitting signal processing circuit 5 has been shown outside the Faraday cage 2, it is understood that the invention is not limited in this respect and that the transmitting signal processing circuit 5 could be located in the Faraday cage 2 with the antenna 4. Also, although only one antenna 4 has been shown in this non-limiting embodiment, it is understood that the electromagnetic wave source 3 could include several antennas in the Faraday cage 2, which antennas may then be synchronized or not with each other.

[0072] The faradized structure 2 comprises two holes 2a, 2b, from which electromagnetic leakage signals, respectively S2a, S2b, escape, which are to be detected with the system 1 according to the invention.

[0073] The system 1 also includes an electromagnetic wave receiver 7, described in more detail below, comprising a plurality of receiving antennas 8 (visible in [Fig. 2]) mounted on a panel 9a and connected to a receiving signal processing circuit 10 by a wire 11. In the non-limiting schematic embodiment shown, the electromagnetic wave receiver 7 has the form of a vertical panel 9a mounted on a base 9b, and carrying a plurality of antennas 8, as shown in [Fig. 2], on the face of the panel 9a facing the Faraday cage 2 in use. In [Fig. 2], the receiving signal processing circuit 10 has not been shown to facilitate reading the drawing.

[0074] Although the plurality of antennas 8 have been represented with a Greek cross-type cross configuration on [Fig.2], it is understood that this configuration is not limiting, and that the plurality of antennas 8 can be arranged on the panel 9a in any configuration, for example in a St. Andrew's cross, in a Y, in a spiral, or even in any random shape.

[0075] Preferably, the antennas 8 carried by the panel 9a are identical and adapted to the frequency emitted by the antenna 4 (where applicable by the plurality of antennas) of the electromagnetic wave emission source 3.

[0076] The resolution of the receiver 7 then depends on the total size of the virtual network (visibility function) obtained after pair-to-pair correlation of all the physical antennas 8.

[0077] The field of view (FOV) of the receiver 7 depends on the spacing between the antennas 8 and the distance from the receiver 7 to the Faraday cage 2. The greater the spacing between the antennas 8, the smaller the FOV. The greater the distance between the receiver 7 and the Faraday cage 2, the greater the FOV. A person skilled in the art will know how to adjust the distance between the antennas 8 and the distance between the receiver 7 and the Faraday cage 2 to obtain the appropriate FOV.

[0078] In the non-limiting embodiment schematically represented in [Fig. 1], the antenna 4 of the electromagnetic wave source 3 is placed in the faradized structure 2 and the electromagnetic wave receiver 7 is placed outside the faradized structure 2, opposite it.

[0079] As can be seen in [Fig.1], the electromagnetic wave source 3, in particular its transmit signal processing circuit 5, and the electromagnetic wave receiver 7, in particular its receive signal processing circuit 10, are not connected, which means that the antennas 8 are not synchronized with the antenna 4.

[0080] The physical system can be modeled by the simplified diagram in [Fig.3]. We consider a gap network of antennas (for example in Y in this case, represented by the black circles on the left part of [Fig.3]) positioned parallel to the face of the structure to be imaged.

[0081] The scene can be considered to be discretized into noise sources illuminating the antenna array (the size of a pixel corresponds globally to the achievable resolution depending on the size of the synthesized array).

[0082] The technique consists of constructing a virtual network (synthetic aperture in white circles) from the physical network (in black circles), corresponding to a sampling grid of the "visibility function" V(u,v), in the space of spatial frequencies (u,v), where u and v are the relative spacings of the antennas normalized by wavelength.

[0083] More specifically, for each pair (k,l) of antennas, the cross correlation between the 'S' signals received by each antenna is calculated (the result constitutes a sample of the visibility function).

[0084] [Math.l]

[0085] With:

[0086] [Math.2]

[0087] With Xk, Yk, the coordinates of the antenna k in the plane of the source to be imaged.

[0088] Considering an ideal system (isotropic antennas, far-field conditions, narrow bandwidth), the visibility function corresponds to the spatial Fourier transform of the noise source distribution:

[0089] [Math.3]

[0090] with R the separation distance between the network and the structure, Xq, Yq, the coordinates of the source in the plane of the scene, and Tq the total power of the source q-

[0091] We then obtain classically the mapping of the sources (therefore of the electromagnetic leaks EM) by an inverse Fourier transform.

[0092] In the context of the invention, imaging is performed in the near field. This requires correcting the spherical wavefront by a quadratic phase correction.

[0093] The distribution map of the sources can therefore be found in the following way:

[0094] [Math.4] Tq = TF 1 [Ve^i' v ki)]

[0095] With,

[0096] [Math.5] * = TÊ (K + YD-fxl+Yl))

[0097] representing quadratic phase correction under near-field conditions.

[0098] Such an approach is described in the publication "An accurate imaging algorithm for millimeter wave synthetic aperture imaging radiometer in near-field", Jianfei, Chen & Li, Yuehua & Wang, Jianqiao & Li, Yuanjiang & Zhang, Yilong, (2013), Progress In Electromagnetics Research, 141, 517-535, 10.2528 / PIER13060702.

[0099] In addition, broadband processing is performed to provide sufficient frequency diversity to decorrelate the sources.

[0100] To avoid creating a complex system, we will consider here that the calculation of the broadband visibility function is equivalent to the sum of the visibility functions calculated by sub-bands (correlation properties approximated to those of white noise). The architecture of the invention is therefore based on narrowband, frequency-scanning (or frequency-hopping) receiver chains.

[0101] The final frequency diversity image is therefore calculated from the inverse Fourier transform of the sum of the phase-corrected visibility functions.

[0102] To calculate the source mapping, it is also possible to model the system by a transition matrix (matrix [G]) between the noise power space and visibility functions (see the publication "Regularization of an inverse problem in remote sensing imaging by aperture synthesis", E. Anterrieu, Proc. IEEE Int. Conf. Acoust., Speech Signal Process. (ICASSP), vol. 2. May 2006, p. 2):

[0103] [Math.6]

[0104] With M: Number of antennas

[0105] P = M x M: Number of visibility samples

[0106] Q: Number of pixels (sources)

[0107] [ipq(f)] : matrix of Green's functions between the source q and the antennas

[0108] The preceding equation can be put into the following matrix form (by vectorizing V):

[0109] [Math.7] FL rGilV G / 1.1 / ... Grll: I I_I Gy 12 / Gy 1.2 / ... .... ||7;| $ ' : t ' '■ v 1 v"' j [ | GJMMi ...

[0110] The distribution of sources can then be calculated by a classical pseudo-inversion method of G: [YES] [Math. 8]

[0112] This method allows us to directly consider the near-field Green's function.

[0113] Referring to [Fig.4], one can see that a signal processing circuit for the transmission 5 of an electromagnetic wave source 3 according to the invention is shown, comprising, in order, a pseudorandom noise sequence generator 51, a digital filter 52 configured to filter the pseudorandom noise sequence, an IQ modulator 53 configured to successively modulate in phase and amplitude a set of predetermined carrier frequencies fp using the filtered pseudorandom noise sequence, and optionally a signal amplifier 54 before at least one transmitting antenna 4, and a computing circuit 55, of the type microcontroller, processor, microprocessor, digital signal processor, DSP, programmable pre-diffused matrix, FPGA, application-specific integrated component, ASIC, or computer which controls the various elements to transmit a signal to the antenna 4.

[0114] The pseudo-random noise sequence generated by the pseudo-random noise sequence generator 41 is represented in [Fig.5] and comprises a frame of duration Tframe consisting of several identical subframes, each subframe comprising for example a pseudo-noise sequence PN followed by a sequence of zeros, the duration of a subframe being TPN.

[0115] At the output to the antenna 4, we obtain a signal such as represented in [Fig.6], namely a baseband modulated signal in amplitude and phase by the pseudo-random noise sequence modulating and transposed sequentially, from a start frequency fstart to an end frequency fstop, by frequency jumps Af, to a set of predetermined carrier frequencies fp.

[0116] By way of example, the 2.5–3.5 GHz band can be covered with the predetermined set of carrier frequencies, in which case the set of predetermined carrier frequencies will comprise discrete frequencies between 2.5 and 3.5 GHz, for example spaced 50 MHz apart. A person skilled in the art will be able to adapt the frequency band covered and the number of discrete frequencies desired to a specific application, bearing in mind that the wavelength emitted by the antenna 4 must be less than the dimensions of the Faraday cage, preferably at least three times less.

[0117] Referring now to [Fig. 7], we can see that the processing circuit 10 of the electromagnetic wave receiver 7 is shown. The processing circuit 10 comprises an N-way to 1-way RF switch 101 controlled by a driver 108, the N channels of the RF switch 101 being respectively connected to the N receiving antennas 8, and, in series after the single output channel of the RF switch 101, a processing chain comprising an IQ demodulator 104 configured to demodulate in phase and amplitude the set of predetermined carrier frequencies using the pseudo-random noise sequence known from the processing circuit 10, and filters low-pass 105, analog-to-digital converters 106, with optionally at the output of the RF switch 101 upstream of the processing chain, at least one of a signal amplifier 102 and a band-pass filter 103, and a signal processing circuit 107, controlling the local oscillator 109 and the analog-to-digital converters 106, to generate an image of electromagnetic leakage from the received signals.

[0118] The role of the processing circuit 10 is to control the capture time between each antenna 8 switching to maintain phase continuity. The control of switch 101 is coded on q bits (N=2Aq, with N being the number of antennas 8). A suitable clock frequency allows three frames to be captured per antenna 8, over a total acquisition time of T^g = 3 x N x 7^^.

[0119] Unlike a conventional interferometric aperture synthesis architecture, the approach proposed by system 1 of the invention requires only a single receiver chain. From a practical point of view, this solution offers numerous advantages:

[0120] - the architecture of system 1 will be greatly simplified, with a single N switch antennas towards 1 channel;

[0121] - for a network with N antennas, N switching operations are performed instead of N2 in the case classic, which will consequently minimize acquisition time;

[0122] - the acquisition will be carried out sequentially for each antenna 8, and not for a pair of antennas simultaneously, effectively eliminating interference related to two receiving chains. It also makes the system more robust to external signals (it is unlikely to correlate parasitic electromagnetic disturbances at different times);

[0123] - the correlation performed between the source and the receptor intrinsically gives Additional information about the depth would be needed. It would therefore be possible to reconstruct a 3D image to detect electromagnetic leaks.

[0124] The captured sequence (after analog-to-digital conversion) therefore corresponds to the signals received by the N antennas in series. It is then digitally processed according to the following steps:

[0125] 1 - Frequency retrieval algorithm:

[0126] - Control of the receiving PLL (phase-locked loop) to perform demodulation at the first carrier frequency.

[0127] - For each frequency:

[0128] - Optimal subsampling: calculation of the symbol energy for estimation of the optimal index, corresponding to the most energetic sample for each symbol.

[0129] - Frame start identification: correlation of the sequence received by the frame The PN is known at the receiver. It should be noted here that this processing amounts to calculating the channel's impulse response with low temporal resolution. It is important to verify at this stage that the channel is non-frequency-selective (frequency-flat channel: the channel must be "contained" between two peaks to avoid any inter-symbol interference).

[0130] - Calculation of the signal-to-noise ratio (SNR): estimation of the signal-to-noise ratio average (average peak power relative to the noise floor). A threshold SNR must be reached (estimated based on receiver sensitivity) for the captured sequence to be considered valid. If this SNR is not reached, the capture is repeated until a timeout occurs (in which case the current carrier frequency is not considered).

[0131] 2 - Imaging:

[0132] - After scanning all frequencies, a cube matrix (antennas / time / frequency) is thus constituted. For each antenna and each frequency, the antenna signals are then constituted by complex coefficients, resulting from the N previous correlations;

[0133] - constitution of the 2D visibility matrix (antennas / frequencies) by correlation of all pairs of antenna signals.

[0134] - calculation of the image at each frequency by inverse Fourier transform of the visibility function with near-field quadratic phase correction (see publication “An accurate imaging algorithm for millimeter wave synthetic aperture imaging radiometer in near-field” [An accurate imaging algorithm for a near-field millimeter wave synthetic aperture imaging radiometer], Jianfei, Chen & Li, Yuehua & Wang, Jianqiao & Li, Yuanjiang & Zhang, Yilong. (2013)., Progress In Electromagnetics Research. 141.517-535.10.2528 / PIER13060702)

[0135] - calculation of the total image by frequency recombination (frequency diversity).

[0136] The image is then analyzed to detect electromagnetic leakage on the faradized structure.

[0137] The flowchart in [Fig.8] details the algorithm implemented.

[0138] The local oscillator is driven at the frequency fp.

[0139] A received frame is captured to obtain IQ signals.

[0140] The antenna switch driver is triggered automatically, the duration of the received frame being greater than or equal to Ttrig = 3*Ttrame*N.

[0141] A Nyquist filtering is then performed, before the received frame is cut into N subframes, each subframe corresponding to one of the antennas.

[0142] Next, for each of the N subframes, a loop is performed, comprising:

[0143] - a synchronization of the symbols in the subframe under consideration,

[0144] - an optimal subsampling by symbol energy calculation,

[0145] - a correlation between the received frame and the PN sequence,

[0146] - a search for a second correlation peak that corresponds to the sn antenna signal,

[0147] - a phase correction related to a possible sample shift,

[0148] - a calculation of received power by autocorrelation,

[0149] - an estimate of the signal-to-noise ratio (SNR).

[0150] The average power and the average SNR on the N antennas are then calculated.

[0151] If the average SNR is below a threshold, a new frame is captured, otherwise, a visibility matrix is ​​calculated.

[0152] This is then normalized by a Vcal(fp) calibration matrix obtained by a prior calibration step of the system in wired mode.

[0153] A near-field quadratic phase correction is then performed, before the image at the frequency fp is calculated.

[0154] The final frequency diversity image is obtained once all the frequencies fp have been scanned.

[0155] A first demonstrator for the simultaneous measurement of shielding attenuation and imaging of electromagnetic leakage has been produced with the following characteristics:

[0156] - Total bandwidth: 2 - 4 GHz

[0157] - Instantaneous bandwidth: 5 MHz

[0158] - 8 x 8 cross antennas (16-channel receiver)

[0159] - Resolution: 20 cm @ 2m, 50 cm @ 5m

[0160] - Field Of View: 1.5m @ 2m, 3.6m @ 5m

[0161] - Source output power: 20 dBm

[0162] - Maximum receiver input power: -2 dBm

[0163] - Sensitivity: -145dBm / Hz.

[0164] This first demonstrator D was used to detect electromagnetic leakage on a faradized structure SF, as shown in [Fig.9a].

[0165] For the demonstrator, the size of the virtual network corresponds to that of the physical network.

[0166] Let D be the total size, R the observation distance, lambda the wavelength, the resolution can be approximated by: R x lambda / D.

[0167] As can be seen in [Fig.9b], representing the image obtained by the System D on the SF structure, we visualize in SF1 and SF2 two electromagnetic leakage points on the bar of the SF structure's gate

Claims

Demands

1. - System (1) for imaging electromagnetic leakage in a faradized structure (2), the system (1) comprising an electromagnetic wave source (3) and an electromagnetic wave receiver (7), characterized in that the electromagnetic wave source (3) is configured to be disposed in the faradized structure (2), the electromagnetic wave receiver (7) being configured to be disposed outside and opposite the faradized structure (2), and the electromagnetic wave source (3) comprising a pseudo-random noise generator (51), an emitting signal processing circuit (5), a frequency modulator (53) and at least one emitting antenna (4), configured such that the at least one emitting antenna (4) emits a pseudo-random noise sequence transposed onto a set of predetermined carrier frequencies, said pseudo-random noise sequence being known to the electromagnetic wave receiver (7),The electromagnetic wave receiver (7) comprises a broadband receiving array of N antennas (8) and a receiving signal processing circuit (10) connected to the receiving antenna array (8) and configured to sequentially switch the N receiving antennas (8), to demodulate at the frequencies of the predetermined carrier frequency set the signals received from the N receiving antennas (8), to correlate the demodulated signals with said pseudo-random noise sequence and to construct an image from the cross-correlation, for each pair of receiving antennas, of the correlated demodulated signals, the resulting image identifying the position of electromagnetic leakage in the faradized structure (2).

2. - System (1) according to claim 1, characterized in that the transmit signal processing circuit (5) comprises computing circuits (55) and memory and, in series after the pseudo-random noise sequence generator (51), a digital filter (52) configured to filter the pseudo-random noise sequence, an IQ modulator (53) configured to successively modulate in phase and amplitude the set of predetermined carrier frequencies using the pseudo-random noise sequence filtered, and optionally a signal amplifier (54) before at least one transmitting antenna.

3. - A system (1) according to claim 1 or claim 2, characterized in that the receive signal processing circuit (10) comprises computing circuits (107) and memory and an N-way to 1-way RF switch (101), the N channels of the RF switch (101) being respectively connected to the N receive antennas (8), and, in series after the single output channel of the RF switch (101), a processing chain comprising an IQ demodulator (104) configured to phase- and amplitude-demodulate all predetermined carrier frequencies using the known pseudo-random noise sequence, at least one low-pass filter (105), at least one analog-to-digital converter (106), at least one local oscillator (109), with optionally at the output of the RF switch (101) upstream of the processing chain, at least one signal amplifier (102) and of a bandpass filter (103),the computing circuits (107) controlling the local oscillator (109) and at least one analog-to-digital converter (106).

4. - System (1) according to any one of claims 1 to 3, characterized in that the receiving antenna array comprises N antennas (8) arranged in a configuration among a cross configuration, a Y configuration, a spiral configuration and a random configuration.

5. - System (1) according to any one of claims 1 to 4, characterized in that the receiving antenna array comprises N antennas (8) arranged on a plane (9a) configured to be vertical in use.

6. - System (1) according to any one of claims 1 to 5, characterized in that at least one transmitting antenna (4) transmits in the centimeter band.

7. - System (1) according to claim 3 or according to any one of claims 4 to 6 taken in combination with claim 3, characterized in that the N-way to 1 RF switch (101) is controlled by a clock frequency configured for the capture, by the N-way to 1 RF switch (101), of at least three pseudo-random noise sequence frames per receiving antenna (8).

8. - System (1) according to any one of claims 1 to 6, characterized in that the electromagnetic wave source (3) includes a self-contained power supply battery.

9. - System (1) according to any one of claims 1 to 8, characterized in that the set of predetermined carrier frequencies comprises discrete frequencies between 500MHz and 10 GHz, preferably between 2.5 and 3.5 GHz, preferably spaced 50 MHz apart.

10. - System (1) according to any one of claims 2 and 3, or according to any one of claims 4 to 9 taken in combination with any one of claims 2 and 3, characterized in that the computing circuits (55, 110) are one or more of a microcontroller, a processor, a microprocessor, a digital signal processor, DSP, a programmable pre-diffused matrix, FPGA, an application-specific integrated component, ASIC, a computer.

11. - A method for imaging electromagnetic leakage in a faradized structure using a system (1) according to any one of claims 1 to 10, characterized in that it comprises: - arranging the electromagnetic wave source in the faradized structure; - arranging the electromagnetic wave receiver opposite the faradized structure; - emitting, by the electromagnetic wave source, a pseudo-random noise sequence transposed onto a set of predetermined carrier frequencies; - receiving, by the electromagnetic wave receiver, the electromagnetic waves emitted by the electromagnetic wave source;- sequentially switch the N receiving antennas of the electromagnetic wave receiver, demodulate the signals received from the N receiving antennas at the frequencies of the predetermined carrier frequency set, correlate the demodulated signals with said pseudo-random noise sequence and construct an image from the cross-correlation, for each pair of receiving antennas, of the correlated demodulated signals, the resulting image identifying the position of electromagnetic leaks in the faradized structure.;

Images