A system and method for monitoring radio frequency fingerprints of rogue drones

Abstract

The application discloses a system and method for monitoring the radio frequency fingerprint of a UAV in violation of regulations. The system includes a radio frequency receiving module, a radio frequency fingerprint identification module, a remote identification detection channel, a trigger control unit, and a camera evidence taking module. The radio frequency receiving module receives the radio frequency signal of the UAV and outputs in-phase and quadrature sampling data. The radio frequency fingerprint identification module extracts the radio frequency fingerprint features and makes a classification decision. The remote identification detection channel simultaneously listens to and analyzes the remote identification broadcast signal. The trigger control unit marks the signal source UAV as a cooperative target when the remote identification broadcast signal is analyzed. If the remote identification broadcast signal is not analyzed and the radio frequency fingerprint classification result is a non-cooperative UAV, the signal source UAV is determined to be a target in violation of regulations and the camera evidence taking module takes evidence images. The system provided by the application can monitor and take evidence of UAVs in violation of regulations.

Description

Technical Field

[0001] This application relates to the fields of low-altitude security and drone monitoring technology, and more specifically, to a radio frequency fingerprint monitoring system and method for monitoring drones flying illegally. Background Technology

[0002] With the continuous growth in the number of civilian drones, they are increasingly being used in fields such as aerial photography, inspection, surveying, logistics, and entertainment. At the same time, drones flying without permission, failing to broadcast their identity information as required, or unable to verify their identity may pose a security threat to areas such as power transmission lines, nuclear power plants, and large event venues. Therefore, how to accurately identify and promptly document drones within the monitored area has become a problem that needs to be solved in the field of low-altitude security.

[0003] Existing solutions mainly include detection schemes based on remote identification broadcast signals, spectrum scanning schemes based on radio frequency energy detection, classification and identification schemes based on radio frequency fingerprints, and visual recognition schemes based on optical images. Specifically, the remote identification and detection scheme identifies targets by analyzing information such as the identity, location, and flight status broadcast by the UAV; the radio frequency energy detection scheme determines the presence of a target by receiving radio frequency signals within the UAV's communication frequency band; the radio frequency fingerprinting scheme classifies targets by extracting physical layer features generated by the radio frequency front-end; and the visual recognition scheme acquires images using camera equipment and identifies UAV targets.

[0004] Among these challenges, remote identification and detection schemes struggle to detect drones that are not broadcasting remote identification signals; radio frequency energy detection schemes have difficulty distinguishing drone signals from those of other wireless devices on the same frequency band; existing radio frequency fingerprinting schemes typically lack collaborative decision-making with remote identification and detection, potentially leading to repetitive classification processing of cooperative targets, and they lack automatic linkage from target identification to image forensics; visual recognition schemes are also susceptible to the effects of distance, lighting conditions, and weather conditions. In summary, existing solutions struggle to simultaneously address remote identification blind spot monitoring and automatic forensics.

[0005] Therefore, this application provides a radio frequency fingerprint monitoring system and method for unauthorized unmanned aerial vehicles (UAVs) to solve one of the aforementioned technical problems. Summary of the Invention

[0006] The purpose of this application is to provide a radio frequency fingerprint monitoring system and method for monitoring illegally flying drones, which can solve at least one of the technical problems mentioned above. The specific solution is as follows: According to a specific embodiment of this application, in a first aspect, this application provides a radio frequency fingerprint monitoring system for unauthorized unmanned aerial vehicles (UAVs), comprising: The system includes: an RF receiving module for receiving RF signals from a UAV and outputting in-phase quadrature sampling data; an RF fingerprinting module connected to the RF receiving module for extracting RF fingerprint features from the in-phase quadrature sampling data, classifying and determining the data using an RF fingerprint classification model, and outputting the RF fingerprint classification result to a trigger control unit; a remote identification detection channel operating in parallel with the RF fingerprinting module for listening to and parsing remote identification broadcast signals conforming to the remote identification data format, and outputting the remote identification parsing result to the trigger control unit; the trigger control unit for marking the UAV source as a cooperative target and not outputting an evidence collection trigger signal when the remote identification parsing result indicates that the remote identification broadcast signal has been parsed, or for determining the UAV source as an illegal flight target and outputting the evidence collection trigger signal when the remote identification parsing result indicates that the remote identification broadcast signal has not been parsed and the RF fingerprint classification result indicates that the UAV source is a non-cooperative UAV; and a video evidence collection module for capturing evidence images of the airspace where the illegal flight target is located after receiving the evidence collection trigger signal.

[0007] According to a specific embodiment of this application, in a second aspect, this application provides a method for monitoring the radio frequency fingerprint of illegally flying drones, applied to an radio frequency fingerprint monitoring system for illegally flying drones including a radio frequency receiving module, a radio frequency fingerprint recognition module, a remote identification and detection channel, a trigger control unit, and a camera evidence collection module. The method includes: The system controls the radio frequency receiving module to receive the UAV radio frequency signal and obtain in-phase quadrature sampling data; it controls the remote identification detection channel and the radio frequency fingerprint recognition module to perform remote identification detection and radio frequency fingerprint recognition in parallel; wherein, the remote identification detection includes listening to and parsing the remote identification broadcast signal that conforms to the remote identification data format to obtain the remote identification parsing result, and the radio frequency fingerprint recognition includes extracting radio frequency fingerprint features from the in-phase quadrature sampling data and using the radio frequency fingerprint classification model to classify and decide to obtain the radio frequency fingerprint classification result; the system controls the trigger control unit to mark the UAV from which the signal originates as a cooperative target and not output the evidence collection trigger signal when the remote identification parsing result is that the remote identification broadcast signal has been parsed, or to determine the UAV from which the signal originates as an illegal flight target and output the evidence collection trigger signal when the remote identification parsing result is that the remote identification broadcast signal has not been parsed and the radio frequency fingerprint classification result is that the UAV from which the signal originates is a non-cooperative UAV; and the system controls the camera evidence collection module to take evidence images of the airspace where the illegal flight target is located after receiving the evidence collection trigger signal.

[0008] Compared with the prior art, the above-described solution of this application has at least the following beneficial effects: This application provides a radio frequency fingerprint monitoring system for unauthorized drones. By having a radio frequency receiving module, a radio frequency fingerprint recognition module, and a remote identification detection channel work in parallel, the system can first identify and filter cooperative targets based on remote identification broadcast signals. Then, when no remote identification broadcast signals are parsed, it can identify non-cooperative drones by combining the radio frequency fingerprint classification results. The trigger control unit then links the camera evidence collection module to capture evidence images. This system provides a way to balance the filtering of compliant targets, compensation for remote identification blind spots, and automatic evidence collection. Attached Figure Description

[0009] Figure 1 A flowchart for a method of monitoring the radio frequency fingerprint of drones flying illegally.

[0010] Figure 2 This is a block diagram of a radio frequency fingerprint monitoring system for unauthorized drone flights. Figure 3 A schematic diagram illustrating the deployment of a radio frequency fingerprint monitoring system for illegally flying drones on a high-voltage power transmission tower; Figure 4 The structure diagram of a lightweight one-dimensional convolutional neural network; Figure 5 This is a schematic diagram illustrating the principle of angle-of-arrival direction finding using three antennas. Figure Labels

[0011] 201. Radio frequency receiving module; 202. Radio frequency fingerprint recognition module; 203. Remote identification and detection channel; 204. Trigger control unit; 205. Camera evidence collection module; 206. Angle of arrival direction finding unit; 207. Power supply module; 208. Data transmission interface; 209. First antenna channel; 210. Second antenna channel; 211. Third antenna channel; 212. Edge processor. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the application. The singular forms “a,” “said,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0014] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0015] It should be understood that although the terms first, second, third, etc., may be used in the embodiments of this application, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, first may also be referred to as second without departing from the scope of the embodiments of this application, and similarly, second may also be referred to as first.

[0016] Depending on the context, the words “if” or “suppose” as used here can be interpreted as “when” or “in response to determination” or “in response to detection.” Similarly, depending on the context, the phrases “if determination” or “if detection (of the stated condition or event)” can be interpreted as “when determination” or “in response to determination” or “when detection (of the stated condition or event)” or “in response to detection (of the stated condition or event).”

[0017] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a product or system comprising a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a product or system. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the product or system that includes said element.

[0018] It should be noted that any symbols and / or numbers present in the specification that are not marked in the accompanying drawings are not reference numerals.

[0019] The optional embodiments of this application are described in detail below with reference to the accompanying drawings.

[0020] The embodiments provided in this application are embodiments of a method for monitoring the radio frequency fingerprint of unauthorized flying drones.

[0021] The following is combined Figure 1 The embodiments of this application will be described in detail.

[0022] Figure 1 The flowchart for the radio frequency fingerprint monitoring method for illegally flying drones is shown below. Figure 1This method is applied to an RF fingerprint monitoring system for illegally flying drones, which includes an RF receiving module, an RF fingerprint recognition module, a remote identification and detection channel, a trigger control unit, and a camera evidence collection module. The method includes the following steps: S101, control the radio frequency receiving module to receive the radio frequency signal of the UAV and obtain in-phase quadrature sampling data.

[0023] S102, control the remote identification and detection channel and the radio frequency fingerprint recognition module to perform remote identification and detection and radio frequency fingerprint recognition in parallel.

[0024] The remote identification detection includes listening to and parsing remote identification broadcast signals that conform to the remote identification data format to obtain remote identification parsing results. The radio frequency fingerprint identification includes extracting radio frequency fingerprint features from in-phase orthogonal sampling data and using a radio frequency fingerprint classification model to make classification decisions to obtain radio frequency fingerprint classification results.

[0025] S103a When the remote identification analysis result is that a remote identification broadcast signal has been analyzed, the control trigger control unit marks the UAV from which the signal comes as a cooperative target and does not output an evidence collection trigger signal.

[0026] S103b, when the remote identification analysis result is that no remote identification broadcast signal is analyzed and the radio frequency fingerprint classification result is a non-cooperative drone, the control trigger control unit determines the drone from which the signal comes as an illegal flight target and outputs an evidence collection trigger signal.

[0027] S104. After receiving the evidence collection trigger signal, the control camera evidence collection module takes evidence images of the airspace where the illegal flight target is located.

[0028] For example, in step S101, the radio frequency receiving module uses a fully passive receiving method to listen to the radio frequency signals in the monitoring area and does not transmit electromagnetic signals outward.

[0029] For example, the radio frequency receiving module can cyclically scan the frequency bands where the drone remote control link and image transmission link are located. After receiving the drone radio frequency signal, it converts the drone radio frequency signal into an original sampling sequence composed of in-phase and quadrature components.

[0030] For example, in-phase quadrature sampling data retains physical layer information such as instantaneous frequency offset, phase noise, envelope waveform, and in-phase quadrature amplitude-phase imbalance, and is transmitted to the radio frequency fingerprint recognition module for subsequent processing.

[0031] For example, in step S102, the remote identification detection channel and the RF fingerprint recognition module form two parallel signal analysis channels. The remote identification detection channel listens to the remote identification broadcast signal and attempts to parse the remote identification information from the captured broadcast frame; the RF fingerprint recognition module extracts the physical layer features of the RF front end based on the in-phase quadrature sampling data and outputs the classification results corresponding to different signal categories using the RF fingerprint classification model. Further, the processing results of the two channels are sent to the trigger control unit, which performs hierarchical decision-making according to the priority of the remote identification parsing results.

[0032] In a specific example, the RF receiver module cycles through the scanning frequencies between the 2.4 GHz and 5.8 GHz bands. For instance, 2420 MHz, 2450 MHz, 2480 MHz, 5800 MHz, 5825 MHz, and 5850 MHz are used as scanning frequencies, with each scanning frequency staying for 200 ms and each scanning cycle lasting 1.2 s.

[0033] For example, when the detected signal energy exceeds a preset energy condition, the RF receiving module captures at least 500ms of in-phase quadrature sampling data and sequentially performs bandpass filtering, downsampling, and normalization. Based on this, the remote identification and detection channel simultaneously attempts to parse Wi-Fi Beacon frames or Neighbor Awareness Networking (NAN) broadcast frames. The RF fingerprinting module performs feature extraction and classification decisions on the pre-processed in-phase quadrature sampling data.

[0034] For example, in step S103, the trigger control unit receives the remote identification parsing result and the radio frequency fingerprint classification result. When the remote identification detection channel successfully parses a legitimate remote identification broadcast signal, the trigger control unit marks the corresponding signal source drone as a cooperative target, records the parsed remote identification information, and suppresses the evidence collection triggering of the cooperative target. When the remote identification detection channel does not parse a remote identification broadcast signal, and the radio frequency fingerprint classification result indicates that the signal source drone is a non-cooperative drone, the trigger control unit determines the signal source drone as an illegal flight target.

[0035] In one specific judgment method, the radio frequency fingerprint classification model outputs the probability distribution corresponding to each category. The trigger control unit outputs an evidence collection trigger signal when the confidence level for the non-cooperative drone category is greater than 0.85 and the remote identification detection channel does not resolve a legitimate remote identification broadcast signal. However, if neither channel obtains a valid signal, or if the classification result does not meet the criteria for determining an illegal flight target, the radio frequency receiving module continues to perform frequency band scanning.

[0036] For example, in step S104, the camera evidence collection module responds to the evidence collection trigger signal, takes a picture in the direction of the illegal flight target, and saves the captured evidence image in the local storage medium. After completing the shooting, the system can generate an alarm message.

[0037] In one specific embodiment, the alarm message includes a timestamp, system deployment location, RFID fingerprint classification result, classification confidence level, target orientation angle, remote identification status, and evidence-collected image. This information is uploaded to the monitoring center via a data transmission interface. Once the upload is complete, the alarm and image capture process ends, and the system resumes RFID scanning.

[0038] Figure 2 This is a block diagram of a radio frequency fingerprint monitoring system for unauthorized drone flights.

[0039] In some embodiments, refer to Figure 2 The radio frequency fingerprint monitoring system for unauthorized drone flights includes a radio frequency receiving module 201, a radio frequency fingerprint recognition module 202, a remote identification and detection channel 203, a trigger control unit 204, and a camera evidence collection module 205.

[0040] The radio frequency receiver module 201 is used to receive radio frequency signals from the UAV and output in-phase quadrature sampling data.

[0041] The radio frequency fingerprint recognition module 202 is connected to the radio frequency receiving module 201. It is used to extract radio frequency fingerprint features from in-phase quadrature sampling data, classify and decide using the radio frequency fingerprint classification model, and output the radio frequency fingerprint classification result to the trigger control unit 204.

[0042] The remote identification detection channel 203 works in parallel with the radio frequency fingerprint recognition module 202 to listen to and parse remote identification broadcast signals that conform to the remote identification data format, and output the remote identification parsing results to the trigger control unit 204.

[0043] The trigger control unit 204 is used to mark the UAV from which the signal originates as a cooperative target and not output an evidence collection trigger signal when the remote identification analysis result is that a remote identification broadcast signal has been analyzed; or to determine the UAV from which the signal originates as an illegal flight target and output an evidence collection trigger signal when the remote identification analysis result is that no remote identification broadcast signal has been analyzed and the radio frequency fingerprint classification result is that the UAV is a non-cooperative UAV.

[0044] The video evidence collection module 205 is used to capture evidence images of the airspace where the illegal flight target is located after receiving the evidence collection trigger signal.

[0045] Furthermore, Figure 2 The system shown may also include an angle-of-arrival direction finding unit 206, a power supply module 207, and a data transmission interface 208.

[0046] The angle of arrival direction finding unit 206 is used to estimate the direction of the signal source based on the phase difference generated when different antenna channels receive the same UAV radio frequency signal, and send the target direction angle to the trigger control unit 204.

[0047] The power supply module 207 is used to provide operating power to the radio frequency receiving module 201, the radio frequency fingerprint recognition module 202, the remote identification and detection channel 203, the trigger control unit 204, the camera evidence collection module 205, the angle of arrival direction finding unit 206 and the data transmission interface 208 respectively.

[0048] The data transmission interface 208 is used to receive the identification results and alarm information generated by the trigger control unit 204, and upload the identification results, alarm information and evidence images to the monitoring center.

[0049] In this application, the radio frequency fingerprint recognition module 202, the remote identification and detection channel 203, the trigger control unit 204, and the angle of arrival direction finding unit 206 can be processed by the edge processor 212. For example, the edge processor 212 completes remote identification parsing, radio frequency fingerprint feature extraction, classification decision, direction calculation, and trigger control at the system deployment location. This method eliminates the need to send the original in-phase orthogonal sampling data to the monitoring center for classification and reasoning.

[0050] Figure 3 A schematic diagram of a scenario where a radio frequency fingerprint monitoring system for illegally flying drones is deployed on a high-voltage power transmission tower.

[0051] For example, refer to Figure 3 The radio frequency fingerprint monitoring system for unauthorized drone flights can be installed on the tower body of a high-voltage transmission tower. In the example, when a cooperative drone in the monitoring area sends a remote identification broadcast signal, the remote identification detection channel 203 parses the remote identification broadcast signal and marks the corresponding drone as a cooperative target. Drones that do not send remote identification broadcast signals will still generate radio frequency signals for remote control links or image transmission links. The radio frequency receiving module 201 receives the radio frequency signals, and the radio frequency fingerprint recognition module 202 identifies non-cooperative drones based on the radio frequency fingerprint classification results, triggering video evidence collection and information uploading.

[0052] As one specific implementation method, Figure 3 In the deployment scenario shown, the power supply module 207 can obtain electrical energy from the alternating magnetic field around the transmission line, and the data transmission interface 208 can send alarm information to the monitoring center through fourth-generation mobile communication or long-distance low-power wireless communication.

[0053] In some embodiments, the radio frequency receiving module 201 includes a radio frequency tuner and a demodulator. The radio frequency tuner performs down-conversion processing on the UAV radio frequency signal in a specified frequency band to obtain an intermediate frequency signal; the demodulator performs analog-to-digital sampling on the intermediate frequency signal and outputs the undemodulated in-phase quadrature sampling data to the radio frequency fingerprint recognition module 202.

[0054] In one specific implementation, the radio frequency tuner uses the R820T2 tuner, and the demodulator uses the RTL2832U demodulator.

[0055] Specifically, the R820T2 tuner down-converts the UAV RF signal of the selected frequency band to the intermediate frequency. The analog-to-digital converter in the RTL2832U demodulator performs analog-to-digital sampling on the intermediate frequency signal. After digital down-conversion processing, it outputs in-phase quadrature sampling data and transmits the in-phase quadrature sampling data to the edge processor 212 through the universal serial bus interface.

[0056] In this embodiment, the R820T2 tuner has a frequency coverage range of 24MHz to 1766MHz. For monitoring frequency bands exceeding this range, the R820T2 tuner can be used in conjunction with an external downconverter or an independent receiver front-end for frequency band extension. Since the demodulator outputs in-phase quadrature sampling data that has not been demodulated according to the protocol, the in-phase quadrature sampling data can retain physical layer characteristics such as instantaneous frequency offset, phase noise, and envelope waveform generated by the UAV RF front-end.

[0057] In some embodiments, the demodulation method of the RF receiving module 201 is determined by software processing, such as analyzing amplitude modulation (AM) signals, frequency modulation (FM) signals, frequency shift keying (FSK) signals, quadrature phase shift keying (QPSK) signals, and orthogonal frequency division multiplexing (OFDM) signals. Thus, the same RF receiving module 201 can provide raw sampling data for remote identification broadcast signal analysis, UAV remote control signal analysis, and UAV image transmission signal analysis.

[0058] In the solution provided in this application, the physical layer differences upon which RF fingerprint recognition is based stem from the hardware manufacturing differences in the RF front-end of the UAV. Specifically, different UAVs exhibit differences in the frequency accuracy of crystal oscillators, nonlinear distortion in power amplifiers, local oscillator leakage in mixers, and amplitude-phase imbalance in in-phase and quadrature components in quadrature modulators. These hardware differences are reflected in the instantaneous frequency, envelope, power spectrum, cyclostationary characteristics, frequency hopping sequence, and amplitude-phase imbalance characteristics of the in-phase and quadrature sampling data.

[0059] In some embodiments, the radio frequency fingerprint features consist of instantaneous frequency features, envelope features, power spectrum features, cyclic stationary features, frequency hopping sequence features, and in-phase quadrature unbalanced features.

[0060] Specifically, the RF fingerprint recognition module 202 performs instantaneous phase difference analysis on the in-phase quadrature sampling data to obtain instantaneous frequency characteristics; calculates the statistics of the amplitude sequence of the in-phase quadrature sampling data to obtain envelope characteristics; estimates the power spectral density of the in-phase quadrature sampling data to obtain power spectral characteristics; estimates the cyclic spectral density of the in-phase quadrature sampling data to obtain cyclic stationary characteristics; performs time-spectrum analysis on the in-phase quadrature sampling data to obtain frequency hopping sequence characteristics; and performs constellation diagram analysis on the in-phase quadrature sampling data to obtain in-phase quadrature imbalance characteristics.

[0061] For example, instantaneous frequency characteristics may include the mean, standard deviation, and variance of instantaneous frequency offset, used to characterize the local oscillator frequency stability and crystal oscillator accuracy. The RF fingerprint recognition module 202 determines the instantaneous phase based on in-phase quadrature sampling data and differentially calculates the instantaneous phases of adjacent sampling times to obtain the instantaneous frequency sequence and corresponding statistics.

[0062] For example, envelope features may include the rate of change of envelope entropy, rising edge slope, and falling edge slope, used to characterize the nonlinear characteristics and automatic gain control response of the power amplifier. The RF fingerprinting module 202 calculates the amplitude sequence based on the in-phase and quadrature components, and determines the envelope statistics based on the amplitude sequence.

[0063] For example, power spectrum characteristics may include negative 3dB bandwidth, out-of-band attenuation rate, and power spectrum symmetry, used to characterize nonlinear distortion generated by the transmit filter and power amplifier. The RF fingerprint recognition module 202 obtains power spectrum characteristics using Welch power spectral density estimation.

[0064] For example, cyclic stationary features may include the cross-sectional peak positions of the cyclic spectral density function, used to characterize the modulation type and symbol rate. The RF fingerprinting module 202 performs cyclic spectral density estimation on in-phase quadrature sampled data and extracts the corresponding peak positions from the cyclic spectral density estimation results.

[0065] For example, frequency hopping sequence characteristics may include the frequency hopping period, dwell time, and frequency set size, used to characterize the sequence generation rules adopted in frequency hopping communication. The radio frequency fingerprinting module 202 generates a time spectrum based on in-phase quadrature sampling data and determines the frequency hopping sequence characteristics based on the time-varying relationship of different frequency components in the time spectrum.

[0066] For example, the in-phase quadrature imbalance feature can include the amplitude ratio and phase difference between the in-phase and quadrature components, used to characterize the hardware mismatch of the quadrature modulator. The RF fingerprint recognition module 202 forms a constellation diagram based on the in-phase quadrature sampling data and determines the in-phase quadrature imbalance feature based on the amplitude and phase distribution of the constellation points.

[0067] In a preferred embodiment, a 128-dimensional radio frequency fingerprint feature vector composed of the above-described features is selected. The radio frequency fingerprint feature vector is input into the radio frequency fingerprint classification model to control the computational load of the edge processor 212 while maintaining the distinguishability between different signal categories.

[0068] In some embodiments, the radio frequency fingerprint recognition module 202 includes an edge processor 212, on which the radio frequency fingerprint classification model is deployed. The radio frequency fingerprint classification model is a lightweight one-dimensional convolutional neural network, including an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a flattening layer, a fully connected layer, and an output layer connected in sequence. After the first and second convolutional layers, a batch normalization layer and a linear rectified activation layer are respectively connected. The output layer uses a normalized exponential function to output the probability distribution of each category.

[0069] Figure 4 This is a structural diagram of a lightweight one-dimensional convolutional neural network.

[0070] For example, refer to Figure 4 The input layer receives an in-phase orthogonal sample sequence of size 1×2024. The sample sequence includes 1024 complex sampling points, each of which contains a real value of an in-phase component and a real value of an orthogonal component.

[0071] The first convolutional layer consists of 32 one-dimensional convolutional kernels of size 3, followed by a batch normalization layer and a linear rectified activation layer. The first pooling layer uses max pooling with a pooling size of 2 and a stride of 2. The second convolutional layer consists of 64 one-dimensional convolutional kernels of size 3, followed by a batch normalization layer and a linear rectified activation layer. The second pooling layer uses max pooling with a pooling size of 2 and a stride of 2.

[0072] For example, the feature data output from the second pooling layer is converted into one-dimensional feature data by a flattening layer and then input into a fully connected layer containing 64 neurons.

[0073] For example, the output layer includes 32 output nodes, which output the probability distribution corresponding to each category through a normalized exponential function.

[0074] For example, the categories in this application may include different drone models, assembled drones and non-drone wireless device categories, triggering the control unit 204 to form radio frequency fingerprint classification results based on probability distribution.

[0075] In some embodiments, the remote identification broadcast signal is carried in a vendor-specific data field of a Wi-Fi beacon frame or a neighbor sensing network broadcast frame. Specifically, the remote identification detection channel 203 identifies the vendor-specific data field in the Wi-Fi beacon frame or the neighbor sensing network broadcast frame, and, according to the remote identification data format, parses the UAV's unique identifier, location, altitude, speed, timestamp, and pilot's location from the vendor-specific data field as the remote identification parsing result.

[0076] In one specific embodiment, after the remote identification and detection channel 203 captures a Wi-Fi signal in the 2.4GHz band, it identifies the frame type of the captured frame. When the captured frame is a Wi-Fi beacon frame or a neighbor sensing network broadcast frame, the remote identification and detection channel 203 locates the manufacturer-specific data fields in the frame and parses the field content according to the remote identification data format. The parsed UAV unique identifier, location, altitude, speed, timestamp, and pilot location are combined into a remote identification parsing result and sent to the trigger control unit 204.

[0077] For example, when the remote identification parsing result indicates that the broadcast frame conforms to the remote identification data format, the trigger control unit 204 marks the corresponding UAV as a cooperative target. The RF fingerprint classification result can still be used as the output of the parallel analysis channel, but it does not participate in triggering the evidence collection for cooperative targets. Based on this, when the remote identification detection channel 203 fails to parse a legitimate remote identification broadcast signal, the trigger control unit 204 determines, based on the RF fingerprint classification result, whether the received RF signal originates from a non-cooperative UAV.

[0078] In some embodiments, the system determines the direction of the illegal flight target by the angle of arrival direction finding unit 206, and controls the camera evidence collection module 205 to face the illegal flight target according to the direction.

[0079] Figure 5 This is a schematic diagram illustrating the principle of angle-of-arrival direction finding using three antennas.

[0080] For example, refer to Figure 5 The radio frequency receiving module 201 includes a first antenna channel 209, a second antenna channel 210, and a third antenna channel 211 arranged in an equilateral triangle on the same horizontal plane. The angle of arrival direction finding unit 206 synchronously collects in-phase orthogonal sampling data from the three antenna channels, performs cross-correlation calculation on the in-phase orthogonal sampling data of any two antenna channels to obtain the received phase difference; determines the reference direction angle of signal incidence according to the received phase difference, signal wavelength, and antenna spacing; performs averaging processing on the reference direction angles determined based on the three antenna pairs to obtain the target direction angle, and outputs the target direction angle to the trigger control unit 204.

[0081] Each of the three antenna pairs consists of two antenna channels, and the antenna channel portions are identical across different antenna pairs. Specifically, in Figure 5 In the direction finding unit 206 shown, the three antenna pairs are respectively composed of the first antenna channel 209 and the second antenna channel 210, the second antenna channel 210 and the third antenna channel 211, and the third antenna channel 211 and the first antenna channel 209.

[0082] In a preferred embodiment, the first antenna channel 209, the second antenna channel 210, and the third antenna channel 211 are each connected to an omnidirectional antenna. The three omnidirectional antennas are arranged in an equilateral triangle on the same horizontal plane, with a spacing d of 20cm between adjacent antennas. Based on this, the angle-of-arrival direction-finding unit 206 synchronously acquires the three in-phase orthogonal sampling data, aligning the three sampling data in time.

[0083] For any pair of antennas, the angle-of-arrival direction-finding unit 206 performs cross-correlation calculations on the two in-phase orthogonal sampled data to obtain the received phase difference when the two antennas receive the same radio frequency signal. .

[0084] Among them, the signal incident direction angle Determined according to the following formula: ; in, This indicates the signal wavelength of the drone's radio frequency signal. This represents the received phase difference of a set of antenna pairs. This indicates the antenna spacing between the two antennas in this antenna pair. This indicates the reference direction angle determined based on this set of antenna pairs.

[0085] In this application, the angle-of-arrival direction-finding unit 206 calculates the reference azimuth angles corresponding to the three antenna pairs respectively, and performs averaging on the three reference azimuth angles to reduce random errors, thereby obtaining the target azimuth angle in the horizontal direction. Under the condition that the signal-to-noise ratio is greater than 15dB, the horizontal direction-finding error of the specific antenna structure can be controlled within ±10 degrees.

[0086] In some further embodiments, the angle of arrival direction finding unit 206 can also estimate the pitch angle based on the intensity difference of the signals received by the three antenna channels, and send the horizontal angle and the pitch angle together as the target direction angle to the trigger control unit 204.

[0087] In some embodiments, the video evidence collection module 205 includes a stepper motor gimbal and a camera module. The trigger control unit 204 responds to the input of a target orientation angle by sending an angle command to the stepper motor gimbal. The stepper motor gimbal drives the horizontal and vertical stepper motors to rotate to the target angle according to the angle command, causing the camera module to face the target orientation angle. After the stepper motor gimbal stabilizes, the camera module powers on to capture evidence images and powers off after capturing the images. The stepper motor gimbal is then reset to its zero position after capturing the images.

[0088] Specifically, after receiving the non-cooperative drone decision result, the trigger control unit 204 obtains the horizontal and pitch angles from the angle of arrival direction finding unit 206, generates an angle command based on the horizontal and pitch angles, and sends the angle command to the stepper motor gimbal.

[0089] For example, the stepper motor gimbal controls the horizontal and vertical stepper motors to rotate according to angle commands. Once the stepper motor gimbal reaches the target angle and stabilizes, it triggers the control unit 204 to power on the camera module.

[0090] In some specific embodiments, the camera module may employ an OV5640 or OV2640 low-power complementary metal-oxide-semiconductor image sensor.

[0091] In this application, the camera module remains powered off when no evidence-collecting trigger signal is received. After the evidence-collecting image is captured, the camera module is immediately powered off, and the stepper motor pan-tilt unit returns to the zero position to reduce the standby power consumption of the camera evidence-collecting module 205.

[0092] In some embodiments, the system employs a multi-level low-power management strategy.

[0093] In some specific embodiments, the radio frequency receiving module 201 operates in a duty cycle scanning mode. Specifically, when the radio frequency receiving module 201 does not detect that the signal energy exceeds the noise floor, the edge processor 212 enters a deep standby state with a timed wake-up function; when the radio frequency receiving module 201 detects that the signal energy exceeds the noise floor, it enters a full-speed operation state; the video evidence collection module 205 powers on after receiving the evidence collection trigger signal and powers off after the shooting is completed.

[0094] In one specific implementation, in deep standby mode, the RF receiving module 201 scans at a 10% duty cycle, with each scan lasting 100ms, followed by a 900ms sleep period. Simultaneously, the edge processor 212 enters deep sleep mode, retaining only the timed wake-up function. When the RF receiving module 201 detects a signal energy more than 10dB higher than the noise floor, it wakes up the edge processor 212, which then enters full-speed operation, performing in-phase quadrature sampling data preprocessing, RF fingerprint feature extraction, and classification inference. The video evidence collection module 205 and the stepper motor gimbal are powered on only when triggering shooting, completing a single shot lasting approximately several seconds. After shooting, the video evidence collection module 205 is powered off, and the system re-enters a low-power operating state combining duty cycle scanning and deep standby.

[0095] In another implementation, an ultra-low power wake-up front end is provided before the radio frequency receiver module 201.

[0096] In some specific embodiments, the ultra-low power wake-up front-end can employ the A7106 receiver chip to detect signal energy in the 2.4GHz band. When the ultra-low power wake-up front-end detects signal energy, it wakes up the main RF receiving channel, which consists of an RTL2832U demodulator and an R820T2 tuner. The main RF receiving channel then collects in-phase quadrature sampling data and performs RF fingerprint analysis and remote identification parsing. In this embodiment, the ultra-low power wake-up front-end only performs signal presence detection and does not replace the original sampling data output function of the main RF receiving channel.

[0097] In some embodiments, the system also performs remote identification deception verification based on the location and radio direction finding results carried in the remote identification broadcast signal.

[0098] Specifically, the remote identification and detection channel 203 converts the UAV position parsed from the remote identification broadcast signal into the desired azimuth and desired distance relative to the system. The arrival angle direction finding unit 206 outputs the measured azimuth and estimates the measured distance based on the signal strength. The control unit 204 then calculates the azimuth difference between the measured azimuth and the desired azimuth, as well as the distance difference between the measured distance and the desired distance. When the azimuth difference or the distance difference exceeds the corresponding judgment threshold, a remote identification deception alarm is output.

[0099] For example, the remote identification and detection channel 203 extracts the latitude, longitude, and altitude of the UAV from the remote identification analysis results, and determines the desired azimuth angle based on the relative spatial relationship between the UAV's position and the system deployment position. and expected distance .

[0100] In some specific embodiments, the angle-of-arrival direction-finding unit 206 determines the measured azimuth angle based on the received phase difference of the three antenna channels. And estimate the measured distance based on the received signal strength indication value. .

[0101] For example, the trigger control unit 204 calculates the azimuth difference according to the following formula: ; in, Indicates azimuth difference. This indicates the measured azimuth angle output by the angle-of-arrival direction finding unit 206. This represents the desired azimuth angle determined based on the drone's position in the remote identification broadcast signal.

[0102] For example, the trigger control unit 204 calculates the distance difference according to the following formula: ; in, This represents the measured distance estimated based on signal strength. This indicates the expected distance determined based on the drone's location in the remote identification broadcast signal.

[0103] In one specific implementation, the judgment threshold corresponding to the azimuth difference can be 20 degrees, and the judgment threshold corresponding to the distance difference can be 100m.

[0104] Specifically, when greater than 20 degrees or When the distance exceeds 100m, the control unit 204 is triggered to output remote identification information for suspected deception.

[0105] In this application, the trigger control unit 204 can also classify alarms according to the number of differences exceeding the judgment threshold.

[0106] For example, as a level 1 alarm, when either the azimuth difference or the distance difference exceeds the corresponding judgment threshold, a low-confidence remote identification of deception is output.

[0107] For example, as a secondary alarm, when both the azimuth difference and the distance difference exceed the corresponding judgment threshold, a high-confidence remote identification deception alarm is output.

[0108] For example, when a remote identification deception alarm is triggered, the trigger control unit 204 can link the video evidence collection module 205 to capture evidence images, and upload the remote identification analysis content, measured azimuth angle, measured distance, azimuth angle difference, distance difference and evidence images to the monitoring center through the data transmission interface 208.

[0109] In some embodiments, the power supply module 207 adopts different power supply structures depending on the system deployment scenario.

[0110] In the scenario of fixed deployment of high-voltage transmission towers, the power supply module 207 adopts the inductive power extraction method. The current transformer is connected to the grounding down conductor or low-voltage take-off line of the high-voltage transmission line, and the induced current is coupled out from the alternating magnetic field around the transmission line. The induced current is then rectified, filtered and regulated in sequence to form the system working power supply.

[0111] In mobile or temporary deployment scenarios, the power supply module 207 can use a rechargeable lithium battery.

[0112] In another embodiment, the power supply module 207 employs a composite power supply structure consisting of a solar panel and a battery. The solar panel charges the battery, and the battery provides operating power to all components of the system. This composite power supply structure is suitable for field monitoring locations where high-voltage transmission lines are not installed.

[0113] In another embodiment, the power supply module 207 adopts Ethernet power supply, which obtains both working power and data transmission channel through a wired network, and is suitable for deployment locations with wired network coverage.

[0114] In this application, the data transmission interface 208 can adopt a fourth-generation mobile communication method or a long-range low-power wireless communication method according to the network conditions of the monitoring location.

[0115] For example, in an area covered by a fourth-generation mobile communication network, the data transmission interface 208 includes a fourth-generation mobile communication CAT-1 communication module and uploads encrypted alarm messages to the monitoring center via a message queue telemetry transmission protocol. The alarm messages include the identification result, evidence-collected images, timestamps, and system deployment location.

[0116] In a specific communication configuration, the fourth-generation mobile communication CAT-1 communication module can adopt the Air724UG communication module.

[0117] In some specific implementation scenarios, in remote areas not covered by fourth-generation mobile communication networks, the data transmission interface 208 includes a long-range low-power wireless communication module and a local storage medium. The data transmission interface 208 prioritizes transmitting text information such as RF fingerprint classification results, target orientation angle, and remote identification status. Evidence-collected images are stored in a MicroSD memory card and uploaded again after network recovery or system recycling.

[0118] As another feasible implementation, in handheld mobile monitoring scenarios, the stepper motor gimbal can be eliminated. The camera module uses a wide-angle lens and is fixedly mounted on the system housing. The user then orients the system towards the target airspace based on the target direction angle output by the system and takes pictures. This implementation retains the relationships between RF signal reception, remote identification and parsing, RF fingerprint classification, and evidence collection triggering, but does not perform the automatic turning and zeroing process of the stepper motor gimbal.

[0119] As another feasible implementation, the edge processor 212 can adopt a field-programmable gate array (FPGA). The inference operation of the RF fingerprint classification model is executed through the hardware parallel logic in the FPGA. This implementation keeps the relationship between in-phase orthogonal sampling data input, RF fingerprint feature processing and classification result output unchanged.

[0120] As another feasible implementation, the RF receiver module 201 can use the HackRF One or AD936x series software-defined radio platform. The software-defined radio platform receives the UAV's RF signal and outputs raw in-phase quadrature sampling data, while the RF fingerprint recognition module 202 still extracts RF fingerprint features and performs classification decisions based on the in-phase quadrature sampling data.

[0121] In another implementation, the radio frequency fingerprint classification model can employ a support vector machine or a random forest model. Specifically, the radio frequency fingerprint recognition module 202 first extracts instantaneous frequency features, envelope features, power spectrum features, cyclostationary features, frequency hopping sequence features, and in-phase orthogonal imbalance features from the in-phase orthogonal sampling data, and then inputs the radio frequency fingerprint features into a support vector machine or random forest model to obtain the radio frequency fingerprint classification result.

[0122] In another implementation, the RF fingerprint classification model can also be obtained using knowledge distillation. Specifically, during the training phase, a teacher model guides the student model's learning, and a small student model is deployed on the edge processor 212. Furthermore, during system operation, the small student model still receives RF fingerprint features or in-phase orthogonal sampling data and outputs the classification results corresponding to each category.

[0123] In another implementation, the first antenna channel 209, the second antenna channel 210, and the third antenna channel 211 can be connected to a quadruple-arm helical antenna or a patch antenna, respectively, to adjust the antenna gain and radiation pattern characteristics. In environments with strong co-channel interference, adaptive filtering can also be performed on the in-phase quadrature sampling data to suppress interference from wireless fidelity signals and Bluetooth signals on the UAV's RF fingerprint characteristics.

[0124] For example, in an implementation that only performs remote identification broadcast signal detection without performing radio frequency fingerprinting, a wireless fidelity baseband processor or a Bluetooth baseband processor can be used to directly receive and demodulate the remote identification broadcast signal. In this implementation, the remote identification detection channel 203 still parses the remote identification information from the vendor-specific data fields of the wireless fidelity beacon frame or the neighbor sensing network broadcast frame, but does not output in-phase quadrature sampling data for the radio frequency fingerprinting module 202 to perform classification and decision.

[0125] Regarding the system in the above embodiments, the specific ways in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0126] Although the operations are described in a specific order in the accompanying drawings, this should not be construed as requiring these operations to be performed in the specific order or serial order shown, or requiring all of the operations shown to obtain the desired result. In certain environments, multitasking and parallel processing may be advantageous.

[0127] The system and method of this application can be implemented using standard programming techniques, and various method steps can be implemented using rule-based logic or other logic. It should also be noted that the terms "system" and "module" as used herein and in the claims are intended to include implementations using one or more lines of software code and / or hardware implementations and / or devices for receiving input.

[0128] Any step, operation, or procedure described herein may be performed or implemented using one or more hardware or software modules, either alone or in combination with other devices. In one embodiment, the software module is implemented using a computer program product comprising a computer-readable medium containing computer program code, which is executable by a computer processor to perform any or all of the described steps, operations, or procedures.

[0129] The foregoing description of implementations of this application has been provided for illustrative and descriptive purposes. The foregoing description is not exhaustive and is not intended to limit this application to the exact forms disclosed. Various modifications and variations may exist in accordance with the foregoing teachings, or may arise from practice of this application. These embodiments were chosen and described to illustrate the principles of this application and its practical application, enabling those skilled in the art to utilize this application in various implementations and modifications to suit the specific purpose of the concept.

[0130] Regarding the system in the above embodiments, the specific ways in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0131] It can be further understood that, unless otherwise specified, "connection" includes both direct connections where no other components exist between the two parties and indirect connections where other components exist between them.

[0132] It is further understood that although the operations are described in a specific order in the accompanying drawings in the embodiments of this application, this should not be construed as requiring these operations to be performed in the specific order or serial order shown, or requiring all the operations shown to be performed to obtain the desired result. In certain environments, multitasking and parallel processing may be advantageous.

[0133] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the field of this application that are not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0134] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

[0135] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A radio frequency fingerprint monitoring system for unauthorized drone flights, characterized in that, include: The radio frequency receiver module is used to receive radio frequency signals from the drone and output in-phase quadrature sampling data; The radio frequency fingerprint recognition module is connected to the radio frequency receiving module and is used to extract radio frequency fingerprint features from the in-phase quadrature sampling data, classify and decide using the radio frequency fingerprint classification model, and output the radio frequency fingerprint classification result to the trigger control unit. The remote identification detection channel works in parallel with the radio frequency fingerprint recognition module. It is used to listen to and parse remote identification broadcast signals that conform to the remote identification data format, and output the remote identification parsing results to the trigger control unit. The trigger control unit is configured to mark the drone from which the signal originates as a cooperative target and not output the evidence collection trigger signal when the remote identification analysis result is that the remote identification broadcast signal has been analyzed; or to determine the drone from which the signal originates as an illegal flight target and output the evidence collection trigger signal when the remote identification analysis result is that the remote identification broadcast signal has not been analyzed and the radio frequency fingerprint classification result is that the drone is a non-cooperative drone. The video evidence collection module is used to capture evidence images of the airspace where the illegal flight target is located after receiving the evidence collection trigger signal.

2. The radio frequency fingerprint monitoring system for unauthorized drone flights according to claim 1, characterized in that, The radio frequency receiving module includes a radio frequency tuner and a demodulator; The radio frequency tuner is used to down-convert the radio frequency signal of the UAV in a specified frequency band to obtain an intermediate frequency signal. The demodulator is used to perform analog-to-digital sampling on the intermediate frequency signal and output the undemodulated in-phase quadrature sampling data to the radio frequency fingerprint recognition module.

3. The radio frequency fingerprint monitoring system for unauthorized drone flights according to claim 1, characterized in that, The remote identification broadcast signal is carried in a vendor-specific data field of a wireless fidelity beacon frame or a neighbor sensing network broadcast frame, and the remote identification detection channel parses the remote identification broadcast signal in the following manner: Identify the vendor-specific data field in the wireless fidelity beacon frame or the neighbor-aware network broadcast frame; According to the remote identification data format, the drone's unique identifier, location, altitude, speed, timestamp, and pilot's location are parsed from the manufacturer-specific data fields to serve as the remote identification parsing result.

4. The radio frequency fingerprint monitoring system for unauthorized drone flights according to claim 2, characterized in that, The radio frequency fingerprint features consist of instantaneous frequency features, envelope features, power spectrum features, cyclostationary features, frequency hopping sequence features, and in-phase quadrature imbalance features; Specifically, the radio frequency fingerprint recognition module performs instantaneous phase difference analysis on the in-phase quadrature sampling data to obtain the instantaneous frequency characteristics, calculates the amplitude sequence statistics of the in-phase quadrature sampling data to obtain the envelope characteristics, estimates the power spectral density of the in-phase quadrature sampling data to obtain the power spectral characteristics, estimates the cyclic spectral density of the in-phase quadrature sampling data to obtain the cyclic stationary characteristics, performs time-spectrum analysis on the in-phase quadrature sampling data to obtain the frequency hopping sequence characteristics, and performs constellation diagram analysis on the in-phase quadrature sampling data to obtain the in-phase quadrature imbalance characteristics.

5. The radio frequency fingerprint monitoring system for unauthorized drone flights according to claim 1, characterized in that, The radio frequency fingerprint recognition module includes an edge processor, and the radio frequency fingerprint classification model is deployed on the edge processor; The radio frequency fingerprint classification model is a lightweight one-dimensional convolutional neural network, which includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a flattening layer, a fully connected layer, and an output layer connected in sequence. The first convolutional layer and the second convolutional layer are respectively connected to a batch normalization layer and a linear rectified activation layer, and the output layer uses a normalized exponential function to output the probability distribution of each category.

6. The radio frequency fingerprint monitoring system for unauthorized drone flights according to claim 3, characterized in that, The system further includes an angle-of-arrival direction-finding unit. The radio frequency receiving module includes three antenna channels arranged in an equilateral triangle on the same horizontal plane. The angle-of-arrival direction-finding unit is used for: Synchronously acquire in-phase orthogonal sampling data from the three antenna channels; Perform cross-correlation calculations on the in-phase orthogonal sampled data of any two antenna channels to obtain the received phase difference; The reference direction angle of signal incidence is determined according to the received phase difference, signal wavelength, and antenna spacing. The target direction angle is obtained by averaging the reference direction angles determined by the three sets of antennas. The target direction angle is output to the trigger control unit; Among the three antenna pairs, each antenna pair is composed of two antenna channels, and the antenna channel portions contained in different antenna pairs are the same.

7. The radio frequency fingerprint monitoring system for unauthorized drone flights according to claim 6, characterized in that, The video evidence collection module includes a stepper motor gimbal and a camera module; The trigger control unit is also used to send an angle command to the stepper motor gimbal in response to the input of the target direction angle. The stepper motor gimbal is used to drive the stepper motors in the horizontal and pitch dimensions to rotate to the target angle according to the angle command, so that the camera module faces the target direction angle; The camera module is used to power on and capture the evidence image after the stepper motor gimbal is stabilized, and to power off after the capture is completed. The stepper motor gimbal is set to the zero position after shooting is completed.

8. The radio frequency fingerprint monitoring system for unauthorized drone flights according to claim 5, characterized in that, The radio frequency receiving module operates in duty cycle scanning mode; The edge processor is configured to enter a deep standby state with timed wake-up function when the RF receiving module does not detect signal energy exceeding the noise floor, or enter a full-speed running state when the RF receiving module detects signal energy exceeding the noise floor. The video evidence collection module powers on upon receiving the evidence collection trigger signal and powers off after the recording is completed.

9. The radio frequency fingerprint monitoring system for unauthorized drone flights according to claim 6, characterized in that, The remote identification and detection channel is also used to convert the UAV position parsed from the remote identification broadcast signal into a desired azimuth and desired distance relative to the system; The angle of arrival direction finding unit is also used to output the measured azimuth angle and the measured distance estimated based on the signal strength; The trigger control unit is also used to calculate the azimuth difference between the measured azimuth and the desired azimuth, and to calculate the distance difference between the measured distance and the desired distance, and to output a remote identification deception alarm when the azimuth difference or the distance difference exceeds the corresponding judgment threshold.

10. A method for detecting radio frequency fingerprints of illegally flying drones, characterized in that, An RF fingerprint monitoring system for illegally flying drones, comprising an RF receiving module, an RF fingerprint recognition module, a remote identification and detection channel, a trigger control unit, and a camera evidence collection module, wherein the method includes: The radio frequency receiving module is controlled to receive radio frequency signals from the UAV and obtain in-phase quadrature sampling data; The remote identification detection channel and the radio frequency fingerprint recognition module are controlled to perform remote identification detection and radio frequency fingerprint recognition in parallel; wherein, the remote identification detection includes listening to and parsing remote identification broadcast signals that conform to the remote identification data format to obtain remote identification parsing results, and the radio frequency fingerprint recognition includes extracting radio frequency fingerprint features from the in-phase quadrature sampling data and using the radio frequency fingerprint classification model to make classification decisions to obtain radio frequency fingerprint classification results; The trigger control unit controls the signal source drone to mark it as a cooperative target and not output the evidence collection trigger signal when the remote identification analysis result is that the remote identification broadcast signal is resolved; or when the remote identification analysis result is that the remote identification broadcast signal is not resolved and the radio frequency fingerprint classification result is that the signal source drone is determined to be an illegal flight target and the evidence collection trigger signal is output. The camera evidence collection module is controlled to capture evidence images of the airspace where the illegal flight target is located after receiving the evidence collection trigger signal.

Images