METHOD FOR RECEIVING DATA IN A RADIO FREQUENCY TRANSMISSION

Description

[0001] Embodiments and implementations relate to radio frequency transmission, and in particular software radios.

[0002] A software-defined radio (also referred to by the acronym "SDR") is a radio frequency receiver or transmitter device made primarily by software and to a lesser extent by hardware.

[0003] In particular, a software radio receiver device generally comprises an antenna configured to receive a modulated radio frequency signal and a radio frequency-to-digital converter for converting the received radio frequency signal into a raw digital signal. The software radio receiver device may also comprise an amplitude demodulator. The demodulated digital signal thus obtained can then be processed by software.

[0004] Such processing operations can be carried out using a microprocessor dedicated to signal processing such as a DSP (Digital Signal Processor), a dedicated integrated circuit, such as an ASIC (Application Specific Integrated Circuit), a programmable electronic circuit, such as an FPGA (Field Programmable Gate Array), or using a computer processor.

[0005] Software-defined radios have the advantage that the software can be changed or adapted to work with a different radio system. For example, software-defined radios can be used to receive Wi-Fi signals, LTE (Long Term Evolution) signals, and Bluetooth signals.

[0006] However, known software radios consume a lot of power.

[0007] In particular, radio frequency-to-digital converters in software radios require generating sampling clocks that consume a large amount of power.

[0008] In particular, it is often necessary to use high-frequency sampling clocks to obtain a digital signal representative of the original radio frequency signal.

[0009] Additionally, sampling requires limiting the signal bandwidth.

[0010] Furthermore, it is possible that some radio frequency signals are sampled to obtain digital signals which are then demodulated while the receiving device is not the recipient of this signal. The energy supplied to sample the radio frequency signal and then demodulate the digital signal is therefore consumed unnecessarily.

[0011] However, in certain applications it is preferable to limit energy consumption as much as possible. In particular, software radios can be used in the field of the Internet of Things so as to enable communication between an object and the Internet network. US 2019 / 319658 discloses a receiver device configured to process a radio frequency signal, the device comprising an analog-to-digital converter and a neural network configured to perform a classification of a modulation scheme of the digital signal.

[0012] Thus, there is a need to propose a method for receiving a radiofrequency signal making it possible to reduce energy consumption compared to known reception methods.

[0013] According to one aspect, there is provided a method of receiving data in a radio frequency transmission comprising: a reception of a radiofrequency signal of a given nature and a conversion of the radiofrequency signal into an electrical signal, at least one detection of at least one voltage level in the electrical signal, a development of at least one train of pulses representative of each detection, a determination of the nature of the radiofrequency signal from said at least one train of pulses.

[0014] The nature of the radio frequency signal can be either compliance with a standard or a protocol. For example, the radio frequency signal can be a Wi-Fi or Bluetooth signal or even LTE (Long Term Evolution).

[0015] The pulse train is associated with the radio frequency signal.

[0016] Determining the nature of the radio frequency signal allows us to know whether the radio frequency signal needs to be processed or not.

[0017] Thus, determining the nature of the radiofrequency signal from the development of a pulse train makes it possible to avoid, for any radiofrequency signal received, sampling the radiofrequency signal in order to obtain a digital signal, then demodulating the digital signal.

[0018] Thus, the reception method makes it possible to avoid generating a sampling clock for determining the nature of the radiofrequency signal.

[0019] Such a method therefore makes it possible to reduce the energy consumption to determine the nature of a radiofrequency signal. Such a method also makes it possible to reduce the number of samples to be analyzed by analyzing only a reduced number of pulses.

[0020] Such a method also makes it possible to avoid having limited conditions concerning the bandwidth of the received radio frequency signal.

[0021] Preferably, such a method can make it possible to detect several types of radiofrequency signal. However, it is possible to provide a reception method in which only one type of signal can be determined.

[0022] In an advantageous embodiment, the radiofrequency signal comprises a header representative of the nature of the radiofrequency signal, said development comprising a development of at least one pulse train associated with the header of the received radiofrequency signal.

[0023] Determining the nature of the radiofrequency signal involves the implementation of an artificial neural network.

[0024] Preferably, the neural network is pre-trained to recognize at least one nature of radio frequency signal from already classified pulse trains.

[0025] The artificial neural network comprises a succession of layers including an input layer for recovering the generated pulse trains, at least one hidden convolution and / or propagation layer for generating corrected pulse trains comprising pulses representative of the nature of the radiofrequency signal and an output layer for transmitting the corrected pulse trains.

[0026] The nature of the radio frequency signal is identified from the corrected pulse trains.

[0027] Preferably, detecting at least one voltage level of the electrical signal comprises comparing the electrical signal to at least one threshold.

[0028] In an advantageous embodiment, the method comprises, once the nature of the signal has been determined, sampling and then processing of the electronic signal in accordance with the nature of the radiofrequency signal determined.

[0029] Alternatively, the method comprises decoding the radiofrequency signal from said at least one corrected pulse train, depending on the determined nature of the radiofrequency signal.

[0030] In one implementation, the radio frequency signal is selected from a Wi-Fi signal, a Bluetooth signal, and an LTE signal.

[0031] According to another aspect, there is provided a radio frequency receiver device comprising: an input interface configured to receive a radiofrequency signal of a given nature and convert it into an electrical signal, detection means configured to detect at least one voltage level in the electrical signal, a pulse generator configured to produce at least one train of pulses representative of the detected voltage levels, ∘ a processing unit configured to determine the nature of the radiofrequency signal from said at least one pulse train.

[0032] The input interface may include an antenna configured to receive the radio frequency signal and convert it into an electrical signal. The input interface may further include an amplifier, including a low noise amplifier, configured to amplify the electrical signal output by the antenna.

[0033] In an advantageous embodiment, the input interface is configured to receive a radiofrequency signal comprising a header representative of the nature of the radiofrequency signal. The pulse generator is then configured to develop at least one pulse train associated with the header of the received radiofrequency signal.

[0034] The processing unit is configured to implement an artificial neural network to determine the nature of the radiofrequency signal. The artificial neural network comprises a succession of layers including an input layer for recovering the generated pulse trains, at least one hidden convolution and / or propagation layer for generating corrected pulse trains comprising pulses representative of the nature of the radiofrequency signal and an output layer for transmitting the corrected pulse trains.

[0035] The receiving device comprises a decoder configured to identify the nature of the radio frequency signal from the pulses of the corrected pulse trains.

[0036] The processing unit can be configured to implement this decoder.

[0037] In an advantageous embodiment, the detection means comprise at least one comparator configured to detect at least one voltage level of the electrical signal by comparing the electrical signal to at least one threshold.

[0038] The pulse generator then takes the comparator's output signals as input and generates a pulse when the comparator's output signals indicate that the voltage level of the radio frequency signal is greater than at least one threshold.

[0039] In an advantageous embodiment, the receiving device further comprises signal processing means, configured to, once the nature of the radiofrequency signal has been determined, sample and then process the electrical signal in accordance with the nature of the determined signal, in particular if the determined nature of the radiofrequency signal is a nature that can be processed by the signal processing means.

[0040] Sampling may be performed using a sampling clock generated by a clock generator of the receiving device. The clock generator may be powered down until the nature of the signal is determined, thereby reducing the power consumption of the receiving device.

[0041] The comparators, the pulse generator and the processing unit are then used as a radio wake-up system (better known to those skilled in the art as a “wake-up radio”) to indicate that a radio frequency signal of an expected nature (i.e. a nature that can be processed by the signal processing means) has been received.

[0042] The receiving device is then configured to turn off the clock generator until a radio frequency signal of an expected nature is received.

[0043] Alternatively, the receiving device may comprise a decoder configured to decode the radio frequency signal from said at least one corrected pulse train, depending on the determined nature of the radio frequency signal.

[0044] The receiving device is then configured to extract all of the useful information from the radiofrequency signal from said at least one elaborated pulse train.

[0045] In one embodiment, the radio frequency signal is selected from a Wi-Fi signal, a Bluetooth signal, and an LTE signal.

[0046] According to another aspect, there is provided an object comprising a receiving device as described above.

[0047] Such an object can be used in particular in the field of the Internet of Things. In particular, such an object is then configured to receive data from a remote device or server, via the Internet network for example. Such an object makes it possible in particular to receive data via a radiofrequency signal while consuming little energy.

[0048] Other advantages and characteristics of the invention will appear on examining the detailed description of modes of implementation and embodiment, which are in no way limiting, and the appended drawings in which: [ Fig 1 ] [ Fig 2 ] [ Fig 3 ] [ Fig 4 ] [ Fig 5 ] [ Fig 6 ] [ Fig 7 ] schematically illustrate embodiments and implementations of the invention.

[0049] There figure 1 illustrates a DIS1 receiver device according to an embodiment of the invention which can serve as software radio.

[0050] The DIS1 receiver device comprises an antenna ANT configured to receive a radio frequency signal and convert this signal into an electrical signal SE.

[0051] The radio frequency signal is of a given nature. The nature of the radio frequency signal may be in compliance with a standard, a protocol. For example, the radio frequency signal may be a Wi-Fi or Bluetooth signal or even LTE (Long Term Evolution).

[0052] The radio frequency signal is of a given nature and carries a header representative of the nature of the radio frequency signal. The radio frequency signal also carries useful information, that is, data that can be used by the receiving device.

[0053] The nature of the radio frequency signal and the useful information of the radio frequency signal are represented in the form of symbols.

[0054] The received radio frequency signal may include noise and errors introduced in the transmission of the radio frequency signal.

[0055] The receiving device also includes a wideband low-noise amplifier AMP connected to the antenna. The amplifier AMP is configured to receive as input the electrical signal SE generated by the antenna ANT and to amplify this electrical signal SE. The amplified signal SA obtained is delivered as output from the amplifier AMP.

[0056] The receiving device comprises an IGD pulse generating device shown in more detail in figure 2 .

[0057] The pulse generation device IGD comprises a plurality of comparators CP0, CP1, CP2, CP3 connected to the output of the amplifier AMP. Each comparator CP0, CP1, CP2, CP3 receives respectively as input the amplified signal SA and a comparison signal SC0, SC1, SC2, SC3 each defining a threshold voltage. Each comparator CP0, CP1, CP2, CP3 is thus associated with a given threshold voltage.

[0058] Comparators CP0, CP1, CP2, CP3 are thus configured to detect overshoots with respect to the threshold voltages. These overshoots are signaled respectively by signals SD0, SD1, SD2 and SD3 at the output of comparators CP0, CP1, CP2 and CP3.

[0059] More particularly, preferably, the IGD pulse generating device comprises a first comparator CP0 used to detect an exceedance of a threshold voltage of 0 volts.

[0060] The IGD pulse generating device further comprises a series of comparators CP1, CP2, CP3 for detecting overshoots with respect to other threshold voltages. In the embodiment shown, the IGD pulse generating device comprises three comparators. However, it is possible to provide a different number of comparators depending on a desired detection resolution.

[0061] The IGD pulse generation device further comprises a time-to-digital converter TDC (in English " Time to digital converter ”) . Preferably, the time-to-digital converter TDC receives as input the output of the comparator CP0. The time-to-digital converter TDC is then configured to measure a duration between two detection sequences, a detection sequence ending when no overshoot has been detected over a duration greater than the transmission time of a symbol.

[0062] The IGD pulse generating device also includes an IG pulse generator. The IG pulse generator includes several inputs configured to receive the detection signals from the comparators CP1, CP2, CP3.

[0063] The pulse generator IG is also configured to generate a pulse as soon as an exceedance of a voltage threshold is detected by a comparator CP1, CP2, CP3. Thus, the pulse generator comprises several outputs O1, O2, O3, each output O1, O2, O3 being associated with a detection linked respectively to the comparator CP1, CP2, CP3.

[0064] The IG pulse generator therefore makes it possible to generate several pulse trains TI1, TI2, TI3 when several voltage threshold overshoots are detected by several comparators CP1, CP2 and CP3 during reception of the radiofrequency signal.

[0065] Some pulses of these pulse trains TI1, TI2, TI3 may be representative of symbols included in the received radiofrequency signal.

[0066] Other pulses may result from noise in the received radio frequency signal or from errors introduced into the radio frequency signal during transmission.

[0067] The IG pulse generator can be formed in particular by an AND type logic gate with a delay on one of its inputs.

[0068] The receiving device DIS1 further comprises a processing unit UT configured to implement an artificial neural network NN.

[0069] The processing unit UT can be a microprocessor for example.

[0070] This artificial neural network NN is stored as a computer program in a memory of the processing unit.

[0071] The neural network NN receives as input the pulse trains TI1, TI2 and TI3 generated by the pulse generator IG.

[0072] In particular, the artificial neural network is previously trained to recognize pulses representative of symbols in the pulse trains it receives so as to generate corrected pulse trains in which the pulses resulting from noise or errors from the transmission are eliminated.

[0073] The pulses of the corrected pulse trains are therefore representative of the symbols contained in the radiofrequency signal.

[0074] In particular, the neural network represented makes it possible to generate the corrected impulse trains TIC1, TIC2 and TIC3.

[0075] The neural network thus generates corrected pulse trains TIC1, TIC2 and TIC3 as output.

[0076] In particular, the NN neural network comprises a succession of layers of neurons.

[0077] Each layer takes data as input to which weights are applied and outputs output data after processing by activation functions of the neurons in that layer. This output data is passed to the next layer in the neural network.

[0078] Weights are data, more specifically parameters, of neurons that can be configured to obtain good output data.

[0079] The weights are adjusted during a generally supervised learning phase, notably by running the neural network with already classified pulse trains from a reference database as input data.

[0080] For example, as shown in the figure 3 for illustration, the first layer CI is an input layer comprising a neuron I1 receiving the pulse train TI1, a neuron I2 receiving the pulse train TI2 and a neuron I3 receiving the pulse train TI3.

[0081] The last CO layer is an output layer comprising an O1 neuron delivering the corrected TIC1 impulse train, an I2 neuron delivering the corrected TIC2 impulse train and an I3 neuron delivering the corrected TIC3 impulse train.

[0082] The neural network NN also includes one or more hidden layers CH between the input layer CI and the output layer CO. At the figure 3 , only one hidden layer CH is represented. This layer CH includes neurons H1, H2 and H3 taking as input the outputs of neurons I1, I2 and I3 of the input layer CI. Neurons O1, O2, O3 of the output layer CO then take as input the outputs of these neurons H1, H2 and H3.

[0083] The hidden layers can be convolution and / or propagation layers. These layers are configured to identify the pulses representative of symbols in the generated pulse trains.

[0084] The durations measured by the time-to-digital converter between two sequences of detections of the comparator CP0 are used to define time windows during which the pulse trains are to be analyzed by the NN neural network.

[0085] The receiving device DIS1 comprises a decoder DEC which can be implemented by the processing unit UT.

[0086] The DEC decoder is configured to identify symbols in the corrected pulse trains it receives.

[0087] In particular, the DEC decoder is configured to recognize the nature of the radio frequency signal included in the signal header from the identified symbols.

[0088] The decoder DEC is also configured to translate the identified symbols of the corrected pulse trains TIC1, TIC2, TIC3 delivered by the neural network NN into a binary data stream DAT representative of the information contained in the radiofrequency signal, depending on the identified nature of the radiofrequency signal.

[0089] The receiving device DIS1 is thus configured to implement a reception method represented in figure 4 .

[0090] The receiving method comprises a receiving step 40 in which the antenna ANT receives a radio frequency signal and converts it into an electrical signal SE.

[0091] The reception method then comprises an amplification step 41 in which the amplifier AMP amplifies the electrical signal SE produced by the antenna ANT. The amplified signal SA obtained is then transmitted to the comparators CP0, CP1, CP2 and CP3.

[0092] Thus, the method comprises a step 42 of detecting overshoots in which the comparators CP0, CP1, CP2 and CP3 compare the amplified signal with the different threshold voltages. When the amplified signal becomes greater than a threshold voltage, the comparator CP0, CP1, CP2 and CP3 associated with this threshold voltage emits a detection signal.

[0093] The method further comprises a step 43 of generating a pulse train by the pulse generator IG. Each pulse represents a detection of an exceeding of a threshold voltage.

[0094] The IG pulse generator therefore makes it possible to translate each detection of an excess of a threshold voltage into a pulse.

[0095] The method then comprises a step 44 of developing the corrected pulse trains TIC1, TIC2, TIC3.

[0096] In this step, the processing unit executes the neural network NN in order to develop the corrected pulse trains TIC1, TIC2, TIC3 from the generated pulse trains TI1, TI2, TI3.

[0097] The corrected pulse trains TIC1, TIC2 and TIC3 are then transmitted to the decoder.

[0098] The method then comprises a decoding step 45 implemented by the decoder DEC. In this decoding step, the decoder DEC first identifies the symbols in the corrected pulse trains TIC1, TIC2 and TIC3.

[0099] The DEC decoder then determines the nature of the radio frequency signal from an identified symbol from the radio frequency signal header.

[0100] The DEC decoder then translates the identified symbols associated with the useful information into a DAT binary data stream representative of the useful information of the radio frequency signal, depending on the determined nature of the radio frequency signal.

[0101] Determining the nature of the radiofrequency signal makes it possible to know, in particular, whether the radiofrequency signal must be processed or not.

[0102] Thus, the development of a pulse train from a detection of voltage level overshoots makes it possible to extract useful information from the radiofrequency signal without sampling the radiofrequency signal.

[0103] Such a process therefore makes it possible to reduce energy consumption to extract useful information from a radiofrequency signal.

[0104] Such a method also makes it possible to reduce the number of samples to be analyzed by analyzing only a reduced number of pulses. Such a method also makes it possible to avoid having limited conditions concerning the bandwidth of the received radiofrequency signal.

[0105] There figure 5 illustrates a DIS2 receiver device according to another embodiment of the invention which can also serve as a software radio.

[0106] Here, the receiver device DIS2 also comprises an antenna ANT, a low noise amplifier AMP, a pulse generation device IGD as well as a processing unit UT as described previously.

[0107] The receiving device of the figure 5 differs from that of the figure 1 by the fact that the processing unit UT is not used to decode the useful information of the radio frequency signal from the corrected pulse trains.

[0108] In particular, useful information is extracted from the radio frequency signal using MT signal processing means.

[0109] The processing means comprise a high-frequency analog-to-digital converter CAD configured to sample the amplified electrical signal SA delivered by the amplifier AMP. In particular, the processing means comprise a clock generator CG configured to be able to generate a sampling clock. This sampling clock is used to clock the sampling performed by the analog-to-digital converter.

[0110] The MT processing means also comprise an IQD amplitude demodulator, in particular a quadrature amplitude demodulator well known to those skilled in the art, in particular its use in software radios.

[0111] The MT processing means are configured to extract useful information from a radiofrequency signal of a given nature.

[0112] The IGD pulse generation device and the UT processing unit are used to determine the nature of the radiofrequency signal and then trigger the MT processing means if the nature of the signal is that for which these MT processing means are configured.

[0113] In particular, the processing unit UT is configured to implement the neural network NN in order to develop corrected pulse trains from pulse trains generated by the pulse generation device, as described for the receiving device of the figure 1 .

[0114] Furthermore, the decoder DEC is configured to identify a symbol, from the signal header, representative of an expected nature of the radiofrequency signal from the corrected pulse trains, the expected nature being a nature that can be processed by the processing means MT.

[0115] The DEC decoder is then configured to trigger the MT processing means when it has identified a symbol representative of the said expected nature.

[0116] The IGD pulse generating device and the UT processing unit can then be used as a radio alarm system to indicate that a radio frequency signal of the expected nature has been received.

[0117] Thus, the DIS2 receiver device of 1a figure 5 is configured to implement a reception method illustrated in figure 6 .

[0118] At the beginning of such a process, the clock generator CG is deactivated and therefore does not consume energy.

[0119] The receiving method comprises a receiving step 60 in which the antenna ANT receives a radio frequency signal and converts it into an electrical signal.

[0120] The reception method then comprises an amplification step 61 in which the amplifier AMP amplifies the electrical signal SE produced by the antenna. The amplified signal SA obtained is then transmitted to the comparators of the pulse generation device IGD.

[0121] Thus, the method comprises a step 62 of detecting overshoots in which the comparators compare the amplified signal with the different threshold voltages. When the amplified signal becomes greater than a threshold voltage, the comparator emits a detection signal.

[0122] The method further comprises a step 63 of generating a pulse train by the pulse generator of the IGD pulse generation device. Each pulse represents a detection of an exceeding of a threshold voltage.

[0123] The pulse generator therefore makes it possible to translate each detection of an excess of a threshold voltage into a pulse.

[0124] The method then comprises a step 64 of determining the nature of the radiofrequency signal from each train of pulses generated by the pulse generator.

[0125] The method then comprises a step 44 of developing the corrected pulse trains TIC1, TIC2, TIC3.

[0126] In this step, the processing unit executes the neural network NN in order to develop the corrected pulse trains TIC1, TIC2, TIC3 from the generated pulse trains TI1, TI2, TI3.

[0127] The corrected pulse trains TIC1, TIC2 and TIC3 are then transmitted to the decoder.

[0128] The method then comprises a decoding step 45 implemented by the decoder DEC. In this decoding step, the decoder DEC identifies whether pulses of the corrected pulse trains correspond to a symbol representative of the expected nature of the radiofrequency signal.

[0129] If the decoder DEC identifies a symbol representative of the expected nature from the corrected pulse trains, then the processing unit UT sends an activation signal SAC to the processing means MT. In particular, the activation signal enables the clock generator CG to be activated.

[0130] Once the clock generator is activated, the method comprises a step 65 of sampling the amplified electrical signal SA from the radiofrequency signal received by the analog-to-digital converter ADC. This sampling makes it possible to obtain a digital signal SB representative of the radiofrequency signal.

[0131] Then, the method comprises a step 66 of demodulation of the digital signal SB by the amplitude demodulator IQD in order to be able to extract the useful information carried by the radiofrequency signal.

[0132] Once the useful information has been extracted, the clock generator CG is disabled to reduce the power consumption of the receiving device.

[0133] There figure 7 represents an object OBJ comprising a receiving device DIS which can be chosen from the receiving device DIS1 shown in figure 1 and the DIS2 receiving device shown in figure 5 .

[0134] Such an OBJ object can be used in particular in the field of the Internet of Things. In particular, such an object is then configured to receive data from a remote device or server, via the Internet network for example. Such an OBJ object makes it possible to receive data via a radiofrequency signal while consuming little energy.

Claims

1. A method for receiving data in a radio frequency transmission comprising: - receiving (40, 60) a radio frequency signal of a given nature and converting the radio frequency signal into an electrical signal, - at least one detection (42, 62) of different voltage levels in the electrical signal, - generating (43, 63) a pulse train for each voltage level detected in the electrical signal, the pulses of each pulse train being representative of each detection of a same voltage level in the electrical signal, - determining (44, 64) the nature of the radio frequency signal from the pulse trains, determining (44, 64) the nature of the radio frequency signal comprising implementing an artificial neural network (NN), the artificial neural network comprising a succession of layers (CI, CH, CO) including an input layer (CI) to recover the generated pulse trains, at least one hidden convolution and / or propagation layer (CH) to generate corrected pulse trains (TIC1, TIC2, TIC3) comprising pulses representative of the nature of the radio frequency signal and an output layer (CO) to transmit the corrected pulse trains, the nature of the radio frequency signal being identified from the corrected pulse trains (TIC1, TIC2, TIC3).

2. The method according to claim 1, wherein the radio frequency signal includes a header representative of the nature of the radio frequency signal, said generation (43, 63) comprising a generation of pulse trains associated with the header of the received radio frequency signal.

3. The method according to one of claims 1 to 2, wherein the detection (42, 62) of at least one voltage level of the electrical signal comprises a comparison of the electrical signal with at least one threshold.

4. The method according to one of claims 1 to 3, comprising, once the nature of the signal has been determined, sampling (65) then processing (66) the electronic signal in accordance with the nature of the determined radio frequency signal.

5. The method according to one of claims 1 to 3, comprising decoding (45) the radio frequency signal from the corrected pulse trains, depending on the determined nature of the radio frequency signal.

6. The method according to one of claims 1 to 5, wherein the radio frequency signal is selected from a Wi-Fi signal, a Bluetooth signal, and an LTE signal.

7. A radio frequency receiver device comprising: - an input interface (ANT, AMP) configured to receive a radio frequency signal of a given nature and convert it into an electrical signal, - detection means (CP1, CP2, CP3) configured to detect voltage levels in the electrical signal, - a pulse generator (IG) configured to generate a pulse train for each voltage level detected in the electrical signal, the pulses of each pulse train being representative of each detection of a same voltage level in the electrical signal, - a processing unit (UT) configured to determine the nature of the radio frequency signal from the pulse trains (TI1, TI2, TI3) by implementing an artificial neural network (NN), the artificial neural network comprising a succession of layers (CI, CH, CO) including an input layer (CI) for recovering the generated pulse trains, at least one convolution and / or propagation hidden layer (CH) for generating corrected pulse trains (TIC1, TIC2, TIC3) comprising pulses representative of the nature of the radio frequency signal and an output layer (CO) to transmit the corrected pulse trains, - a decoder (DEC) configured to identify the nature of the radio frequency signal from the pulses of the corrected pulse trains (TIC1, TIC2, TIC3).

8. The receiver device according to claim 7, wherein the input interface (ANT, AMP) is configured to receive a radio frequency signal including a header representative of the nature of the radio frequency signal, the pulse generator (IG) being configured to generate pulse trains associated with the header of the received radio frequency signal.

9. The receiver device according to one of claims 7 to 8, wherein the detection means comprise at least one comparator (CP1, CP2, CP3) configured to detect at least one voltage level of the electrical signal by comparing the electrical signal with at least one threshold.

10. The receiver device according to any one of claims 7 to 9, further comprising signal processing means (MT), configured to, once the nature of the radio frequency signal has been determined, sample then process the electrical signal in accordance with the nature of the determined signal.

11. The receiver device according to any one of claims 7 to 9, wherein the decoder (DEC) is configured to decode the radio frequency signal from the corrected pulse trains, according to the determined nature of the radio frequency signal.

12. The receiver device according to any one of claims 7 to 11, wherein the radio frequency signal is selected from a Wi-Fi signal, a Bluetooth signal, and an LTE signal13. An object comprising a receiver device according to one of claims 7 to 12.