Specific unmanned aerial vehicle model rapid identification method and system based on radio frequency fingerprint database

Abstract

The invention provides a specific unmanned aerial vehicle model rapid identification method and system based on a radio frequency fingerprint database. The method comprises the following steps: firstly, receiving a radio frequency signal generated by a target unmanned aerial vehicle during flight, carrying out dynamic noise reduction processing on the radio frequency signal to generate a noise-reduced radio frequency signal, then separating combined data comprising a signal strength fluctuation mode, a phase hopping frequency degree and a modulation distortion characteristic from the noise-reduced radio frequency signal, and finally, carrying out noise reduction processing on the combined data. The method comprises the following steps: acquiring combined data of a target unmanned aerial vehicle, performing layer-by-layer matching on the combined data and a pre-constructed radio frequency fingerprint database to generate a matching result, then extracting a model correlation parameter of the target unmanned aerial vehicle from the matching result, finally generating an identification decision containing a real-time environment adaptation factor based on the model correlation parameter, and outputting a target unmanned aerial vehicle model result corresponding to the identification decision; according to the technical scheme provided by the invention, the distinction degree of fine signal differences is greatly improved, and the stability and accuracy of unmanned aerial vehicle model identification in complex electromagnetic interference and multi-obstacle reflection scenes are realized.

Description

Technical Field

[0001] This application relates to the field of drone technology, and in particular, to a method and system for quickly identifying a specific drone model based on a radio frequency fingerprint library. Background Art

[0002] With the wide application of drones in fields such as security monitoring and logistics transportation, in open scenarios with complex electromagnetic interference and multi-obstacle reflections (such as urban building complexes, industrial factory areas, etc.), it is urgent to achieve quick and accurate identification of specific drone models. In such scenarios, the radio frequency signals of drones are easily affected by environmental interference, resulting in distorted features. Moreover, the signal differences between different models of drones are subtle, and signal processing and model matching need to be completed within a millisecond-level response time to meet the requirements of real-time supervision and countermeasures.

[0003] Currently, in response to the above requirements, a matching and identification scheme based on a static radio frequency fingerprint library is mainly adopted. This scheme pre-collects the radio frequency signals of different models of drones in a fixed experimental environment, extracts signal strength, spectral features, etc. to generate a reference fingerprint library; during actual identification, the real-time signals of the target drone are compared with the fingerprint library using a threshold, and the model with the highest matching degree is selected as the identification result.

[0004] The above scheme has significant limitations in dynamic open scenarios: First, the static fingerprint library cannot adapt to the real-time changing electromagnetic interference intensity and obstacle reflection paths in complex environments, resulting in the offset between the signal features and the reference fingerprints exceeding the preset threshold, leading to misjudgments; second, the fixed threshold matching mechanism is difficult to cope with the dynamic fluctuations of the signal distortion degree, and it is necessary to adjust the threshold multiple times or recalibrate the fingerprint library, seriously affecting the real-time performance of identification; third, no collaborative mechanism for environmental parameters (such as the distribution of interference sources, the location of obstacles) and radio frequency fingerprint correction is established, resulting in a great difference in identification efficiency under different scenarios. Summary of the Invention

[0005] This application provides a method and system for quickly identifying a specific drone model based on a radio frequency fingerprint library to solve the problems in the prior art that the static radio frequency fingerprint library cannot adapt to the signal offset caused by electromagnetic interference and obstacle reflection in dynamic open scenarios, the lack of real-time performance of the fixed threshold matching mechanism due to environmental fluctuations, and the unstable identification efficiency caused by the lack of collaborative association between environmental parameters and radio frequency fingerprint correction.

[0006] In a first aspect, this application provides a method for quickly identifying a specific drone model based on a radio frequency fingerprint library, including: Receiving the radio frequency signals generated by the target drone during flight, and performing dynamic noise reduction processing on the radio frequency signals to generate noise-reduced radio frequency signals; Separate combined data containing signal intensity fluctuation patterns, phase jump frequencies, and modulation distortion characteristics from the noise-reduced RF signal, and layer-by-layer match the combined data with a pre-constructed RF fingerprint library to generate a matching result; Extract model-related parameters of the target UAV from the matching result, where the model-related parameters include signal propagation delay differences corresponding to specific obstacle reflection paths and frequency domain distortion correlation degrees corresponding to the distribution of electromagnetic interference sources; Generate an identification decision containing a real-time environment adaptation factor based on the model-related parameters and output the target UAV model result corresponding to the identification decision.

[0007] Optionally, it is characterized in that an RF signal generated during the flight of the target UAV is received, and dynamic noise reduction processing is performed on the RF signal to generate a noise-reduced RF signal, including: Real-time capture the original RF signal of the target UAV through multi-channel signal acquisition devices deployed in the monitoring area, where the original RF signal contains noise components caused by electromagnetic interference sources and obstacle reflections in the current scene; Monitor the environmental parameters in the original RF signal, where the environmental parameters include real-time interference intensity distribution and obstacle density distribution; Select the signal filtering intensity level according to the environmental parameters. If the real-time interference intensity is higher than the set range and the obstacle density is higher than the set threshold, then multi-stage cascaded filtering is used to perform layer-by-layer noise suppression on the original RF signal. If the real-time interference intensity is lower than the set range, then single-stage filtering is used to perform noise suppression on the original RF signal to obtain a filtered RF signal; Perform signal integrity restoration processing on the filtered RF signal to generate a noise-reduced RF signal.

[0008] Optionally, separating combined data containing signal intensity fluctuation patterns, phase jump frequencies, and modulation distortion characteristics from the noise-reduced RF signal, and layer-by-layer matching the combined data with a pre-constructed RF fingerprint library to generate a matching result, including: Divide the noise-reduced RF signal into multiple signal segments according to a preset time window, and extract from each signal segment the fluctuation pattern of the signal intensity varying with the flight altitude, the frequency distribution of the phase jump intervals, and the position characteristics of the modulation waveform distortion; Fuse the fluctuation pattern, frequency distribution, and position characteristics into combined data according to a preset weight ratio; Calculate the similarity of the combined data with the first-layer signal intensity fluctuation pattern reference set stored in the RF fingerprint library in sequence, and screen out a candidate model set with a similarity higher than the first threshold; Jointly match the frequency distribution and position features in the combined data with the second-layer phase jump frequency reference set and modulation distortion reference set corresponding to the candidate model set to generate a matching result including the ranking of the matching degrees of candidate models.

[0009] Optionally, extract the model association parameters of the target UAV from the matching result. The model association parameters include the signal propagation delay difference corresponding to the reflection path of a specific obstacle and the frequency domain distortion correlation degree corresponding to the distribution of electromagnetic interference sources, including: According to the ranking of the matching degrees of candidate models in the matching result, extract the radio frequency fingerprint reference data corresponding to the candidate model with the highest ranking. Obtain the signal propagation delay difference parameter associated with the obstacle distribution in the current scene from the radio frequency fingerprint reference data. Obtain the frequency domain distortion correlation degree parameter associated with the distribution of electromagnetic interference sources from the radio frequency fingerprint reference data. Integrate the signal propagation delay difference parameter and the frequency domain distortion correlation degree parameter into the model association parameters of the target UAV.

[0010] Optionally, generate an identification decision including a real-time environment adaptation factor based on the model association parameters and output the target UAV model result corresponding to the identification decision, including: Obtain the real-time tracking data of the dynamic change of the obstacle position and the real-time fluctuation data of the intensity of the electromagnetic interference source in the current scene. Adjust the weight coefficient of the signal propagation delay difference parameter according to the real-time tracking data, and adjust the weight coefficient of the frequency domain distortion correlation degree parameter according to the real-time fluctuation data. Input the adjusted signal propagation delay difference parameter and the adjusted frequency domain distortion correlation degree parameter into the environment adaptation factor calculation model to generate a real-time environment adaptation factor for correcting the matching degrees of candidate models in the radio frequency fingerprint library. Dynamically compensate the matching degrees of candidate models in the matching result according to the real-time environment adaptation factor to generate an identification decision including the final matching model. Verify the consistency between the model identifier of the final matching model in the identification decision and the model feature description pre-stored in the radio frequency fingerprint library. When the consistency verification passes, output the final matching model as the target UAV model result.

[0011] Optionally, adjust the weight coefficient of the signal propagation delay difference parameter according to the real-time tracking data, and adjust the weight coefficient of the frequency domain distortion correlation degree parameter according to the real-time fluctuation data, including: Dynamically adjust the weight coefficient of the signal propagation delay difference parameter according to the dynamic rate of the change in the position of the obstacle in the real-time tracking data. When the dynamic rate exceeds the preset rate threshold, linearly enhance the weight coefficient according to the ratio of the dynamic rate to the preset rate threshold to generate an adjusted delay weight coefficient; Dynamically adjust the weight coefficient of the frequency-domain distortion correlation parameter according to the fluctuation amplitude of the electromagnetic interference source intensity in the real-time fluctuation data. When the fluctuation amplitude exceeds the preset amplitude threshold, increment the weight coefficient in segments according to the difference between the fluctuation amplitude and the amplitude threshold to generate an adjusted frequency-domain weight coefficient.

[0012] Optionally, dynamically compensate the matching degree of the candidate models in the matching result according to the real-time environment adaptation factor to generate an identification decision including the final matching model, including: Decompose the real-time environment adaptation factor into a delay correction factor and a frequency-domain correction factor; Extract the initial matching degree of the candidate model set from the matching result, and multiply each item of the signal propagation delay difference parameter corresponding to each candidate model in the delay correction factor and the initial matching degree item by item to generate a matching degree after delay compensation; Multiply each item of the frequency-domain distortion correlation parameter corresponding to each candidate model in the matching degree after delay compensation by the frequency-domain correction factor item by item to generate a matching degree after frequency-domain compensation; Combine and process the matching degree after delay compensation and the matching degree after frequency-domain compensation according to a preset ratio to generate a comprehensive compensation matching degree for each candidate model; Sort the candidate model set in descending order according to the comprehensive compensation matching degree, select the candidate model with the highest ranking as the final matching model, and generate an identification decision including the final matching model.

[0013] In a second aspect, the present application provides a specific UAV model rapid identification system based on a radio frequency fingerprint library, including: A receiving module, configured to receive the radio frequency signal generated by the target UAV during flight and perform dynamic noise reduction processing on the radio frequency signal to generate a noise-reduced radio frequency signal; A matching module, configured to separate combined data including a signal intensity fluctuation pattern, a phase jump frequency, and a modulation distortion feature from the noise-reduced radio frequency signal, and perform layer-by-layer matching of the combined data with a pre-constructed radio frequency fingerprint library to generate a matching result; An extraction module, configured to extract the model association parameters of the target UAV from the matching result, where the model association parameters include a signal propagation delay difference corresponding to a specific obstacle reflection path and a frequency-domain distortion correlation corresponding to the electromagnetic interference source distribution; An output module, configured to generate an identification decision including a real-time environment adaptation factor based on the model association parameter, and output a target UAV model result corresponding to the identification decision.

[0014] In a third aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for quickly identifying a specific UAV model based on a radio frequency fingerprint library as described in the first aspect above.

[0015] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, and when the computer program is executed by a computer, it implements a method for quickly identifying a specific UAV model based on a radio frequency fingerprint library as described in the first aspect.

[0016] Through the collaborative design of dynamic noise reduction processing and layer-by-layer matching of combined data, the embodiment of the present application effectively solves the core problem of distortion of UAV radio frequency signal features in complex electromagnetic interference and multi-obstacle reflection scenarios. Specifically, the dynamic noise reduction processing adaptively adjusts the filtering intensity based on real-time environment parameters (interference intensity, obstacle density), significantly suppressing the influence of noise components on key features such as the signal intensity fluctuation pattern and phase jump frequency; at the same time, by separating and fusing multi-dimensional features to generate combined data, combined with the layer-by-layer matching mechanism of the radio frequency fingerprint library (progressive screening of intensity fluctuation → phase jump → modulation distortion), the discrimination of subtle signal differences is greatly improved, so as to achieve high-precision and low-latency model identification in an open scenario.

[0017] Furthermore, by decomposing the environment adaptation factor into a delay correction factor (associated with the obstacle reflection path) and a frequency domain correction factor (associated with the electromagnetic interference distribution), the delay difference parameter and frequency domain distortion parameter of the candidate model are compensated and corrected twice, and combined with a preset ratio to generate a comprehensive matching degree, ensuring the robustness of the identification decision under environmental fluctuations. This mechanism not only avoids the real-time deficiency problem of traditional static threshold matching, but also realizes continuous optimization of scene adaptability through incremental update of the radio frequency fingerprint library, and finally achieves double improvement in the stability and accuracy of UAV model identification in a complex open environment.

[0018] These aspects or other aspects of the present application will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0020] Figure 1 It shows a flowchart of a method for quickly identifying a specific UAV model based on a radio frequency fingerprint library provided by the present application; Figure 2 It shows a schematic structural diagram of a system for quickly identifying a specific UAV model based on a radio frequency fingerprint library provided by the present application; Figure 3 It shows a schematic structural diagram of a computing device provided by the present application. Detailed implementation manners

[0021] To enable those skilled in the art to better understand the solutions of the present application, the following will clearly and completely describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application.

[0022] In some processes described in the specification, claims, and the above drawings of the present application, there are multiple operations that appear in a specific order. However, it should be clearly understood that these operations can be executed not in the order in which they appear in this document or in parallel. The operation numbers such as 101 and 102 are only used to distinguish different operations, and the numbers themselves do not represent any execution order. In addition, these processes can include more or fewer operations, and these operations can be executed in sequence or in parallel. It should be noted that the descriptions such as "first" and "second" in this document are used to distinguish different messages, devices, modules, etc., do not represent a sequence, and do not limit that "first" and "second" are of different types.

[0023] In an open scenario with complex electromagnetic interference and multi-obstacle reflections, the existing matching and recognition solutions based on static radio frequency fingerprint libraries face severe challenges: static fingerprint libraries rely on benchmark data in a fixed experimental environment and are difficult to adapt to the real-time changing electromagnetic interference intensity and obstacle reflection paths in a dynamic open scenario, resulting in the offset between the signal characteristics of the target UAV and the benchmark fingerprints exceeding the preset threshold, leading to misjudgments; at the same time, the fixed threshold matching mechanism cannot respond to the dynamic fluctuations of the signal distortion degree, and the threshold needs to be repeatedly calibrated or the fingerprint library needs to be updated, seriously restricting the recognition real-time performance; in addition, the existing solutions lack a collaborative mechanism for environmental parameters (such as the distribution of interference sources and the positions of obstacles) and radio frequency fingerprint correction, resulting in significant fluctuations in the recognition efficiency in different scenarios and making it difficult to meet the high-precision supervision requirements.

[0024] In view of the above problems, the present application proposes a method for quickly identifying a specific UAV model based on a radio frequency fingerprint library. The core lies in breaking through the inherent limitations of the static fingerprint library and the fixed threshold mechanism through a closed-loop technical link of dynamic noise reduction processing - combined data hierarchical matching - dynamic compensation of environmental adaptation factors. Specifically, the dynamic noise reduction processing is driven by real-time environmental parameters (interference intensity, obstacle density), adaptively suppressing the influence of noise components on core features such as the signal strength fluctuation pattern and the phase jump frequency. Further, the noise-reduced radio frequency signal is separated into multi-dimensional combined data, and candidate models are screened through layer-by-layer matching of the radio frequency fingerprint library (intensity fluctuation → phase jump → modulation distortion), and the matching result is dynamically compensated and corrected in combination with the environmental adaptation factor. This solution effectively eliminates the signal feature offset through the real-time collaboration between the environmental parameters and the radio frequency fingerprint library (such as adjusting the time delay difference weight based on the dynamic rate of obstacles and correcting the frequency domain distortion correlation based on the interference fluctuation amplitude), while avoiding repeated calibration operations, significantly improving the recognition real-time performance, accuracy, and environmental robustness in dynamic open scenarios, and providing reliable technical support for UAV supervision and countermeasures.

[0025] Next, the technical solutions in the embodiments of the present application will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present application.

[0026] Figure 1 The flowchart of a method for quickly identifying a specific UAV model based on a radio frequency fingerprint library provided by an embodiment of the present application is as Figure 1 shown, and the method includes: Step 101, receiving the radio frequency signal generated by the target UAV during flight, and performing dynamic noise reduction processing on the radio frequency signal to generate a noise-reduced radio frequency signal; In this step, the dynamic noise reduction processing refers to the process of adaptively adjusting the noise suppression strategy according to real-time environmental parameters (such as interference intensity distribution, obstacle density distribution); among them, the real-time interference intensity distribution represents the electromagnetic interference energy distribution in different frequency bands in the monitoring area, and the obstacle density distribution represents the density of obstacle reflection paths in a unit space; multi-stage cascaded filtering refers to performing noise suppression operations with different frequency bands and amplitudes on the radio frequency signal based on environmental parameters, and single-stage filtering refers to noise suppression with a single frequency band or amplitude; the signal integrity restoration processing refers to repairing the key signal features related to the UAV flight state that may be lost during the filtering process through a signal segment recombination and feature retention mechanism.

[0027] In this embodiment, first, a multi-channel signal acquisition device deployed in the monitoring area synchronously captures the original radio frequency signal of the target UAV. The device covers different spatial positions to capture multipath reflection signals. Subsequently, the energy distribution in different frequency bands of the original signal (such as the interference intensity in the 2.4 GHz / 5.8 GHz frequency bands) is analyzed in real time, and the obstacle distribution density parameters are obtained synchronously through lidar or infrared sensors. According to the combined relationship between the interference intensity and the obstacle density (such as a high-interference + high-density scenario), multi-stage cascaded filtering is activated: the first stage filters out broadband noise outside the main frequency band, the second stage performs amplitude threshold truncation on impulse noise, and the third stage performs frequency-domain notch filtering on the remaining noise; if the interference intensity or the obstacle density is lower than the threshold, only single-stage main-frequency band-pass filtering is performed. Finally, the continuity of the time-domain waveform of the filtered signal is detected, the signal mutation or phase jump breakpoints caused by filtering are identified, and interpolation repair is performed based on the signal characteristics of adjacent sections (such as amplitude slope, phase continuity), and the signal intensity mutation sections and phase jump sections related to the flight altitude and attitude of the UAV are retained to generate a noise-reduced radio frequency signal.

[0028] For example, in the monitoring scenario of urban building clusters, a multi-channel signal acquisition device deployed on the roof captures the original radio frequency signal of the target UAV, detects strong interference caused by a WiFi router in the 2.4 GHz frequency band (the interference intensity distribution shows that the energy peak exceeds the threshold), and at the same time, the lidar feedback shows that the obstacle density distribution is dense in the reflection path of high-rise buildings (the density value is higher than the threshold). Accordingly, the system activates multi-stage cascaded filtering: the first stage filters out the stray signals outside the 2.4 GHz frequency band, the second stage performs amplitude truncation on the impulse noise caused by building reflections, and the third stage performs frequency-domain notch filtering on the remaining narrow-band interference. After the waveform continuity detection of the filtered signal, two phase jump breakpoints are found, and linear interpolation repair is performed based on the phase slopes of the front and rear sections. Finally, a noise-reduced signal that retains the signal intensity fluctuations related to the flight altitude and the complete phase jump characteristics is generated for subsequent combined data separation and matching.

[0029] Step 102: Separate the combined data containing the signal intensity fluctuation pattern, phase jump frequency, and modulation distortion characteristics from the noise-reduced radio frequency signal, and layer-by-layer match the combined data with a pre-constructed radio frequency fingerprint library to generate a matching result; In this step, the signal strength fluctuation pattern refers to the temporal fluctuation law of the RF signal strength with the change of the UAV flight altitude and attitude; the phase jump frequency indicates the number of discontinuous phase jumps and the interval distribution characteristics of the signal phase within a unit time; the modulation distortion feature refers to the morphological differences in the distortion positions and distortion intervals of the RF signal modulation waveform relative to the standard modulation method; the combined data is a multi-dimensional data set formed by fusing the above three types of features according to a preset weight; the RF fingerprint database stores the reference feature data of different UAV models in typical scenarios; layer-by-layer matching refers to screening candidate models layer by layer in the order of signal strength → phase jump → modulation distortion.

[0030] In this embodiment, first, the noise-reduced RF signal is segmented into continuous signal segments according to a preset time window (such as 50 ms) to ensure that each signal segment contains a complete flight state cycle (such as climbing, hovering). For each signal segment, the signal strength fluctuation pattern is extracted: the signal strength time series curve is aligned by the dynamic time warping algorithm, and the fluctuation similarity with the reference curve of the height change is calculated; the phase jump frequency is extracted: the number of phase jumps within the signal segment is counted, and the frequency distribution histogram of the jump intervals is calculated through a sliding window; the modulation distortion feature is extracted: the modulation waveform of the signal segment is compared with the standard waveform point by point, and the distortion positions are marked and the proportion of the distortion interval length is calculated. Subsequently, the fluctuation pattern, frequency distribution, and distortion feature are fused into a combined data unit according to a preset weight (such as 5:3:2), and the combined data units of all signal segments are spliced into a complete combined data. Then, the combined data is input into the RF fingerprint database. In the first-layer matching, based on the similarity calculation method of dynamic time warping, a set of candidate models with a signal strength fluctuation pattern similarity higher than a threshold (such as 80%) is screened; in the second-layer matching, the comprehensive matching degree of the candidate models is generated by calculating the overlapping area of the phase jump frequency distribution and weighted summing the offset amounts of the modulation distortion positions, and the matching results are generated in descending order.

[0031] For example, in the scenario of monitoring urban building clusters, the noise-reduced RF signals generated in step 101 are segmented into 50-ms time windows. A certain signal segment corresponds to the flight stage of the UAV passing through the gaps between high-rise buildings. When extracting the signal intensity fluctuation pattern, its intensity curve shows periodic attenuation due to building occlusion. After time-series alignment with the reference curve of model C in the fingerprint database, the peak-valley difference is less than 8%. The statistical result of the phase jump frequency shows that the number of jumps in this segment is 12 times / 50 ms (the phase mutation increases due to multipath reflection). The overlapping area ratio of the frequency distribution histogram and the reference histogram of model C reaches 78%. The analysis of the modulation distortion characteristics finds that the waveform distortion positions are concentrated at the modulation falling edge (caused by the delay of the building reflection path), and the distortion interval accounts for 15%. After the combined data is fused according to the preset weights (signal intensity fluctuation 5: phase jump 3: modulation distortion 2), the first-layer matching selects models C and D as the candidate set. In the second-layer matching, the weighted score of the overlapping area of the phase jump of model C (78%) and the distortion offset (0.08) is 0.82, and the score of model D is 0.65. Finally, the result with the highest matching degree of model C is generated. Step 103: Extract the model-related parameters of the target UAV from the matching result. The model-related parameters include the signal propagation time delay difference corresponding to the reflection path of specific obstacles and the frequency-domain distortion correlation degree corresponding to the distribution of electromagnetic interference sources. In this step, the signal propagation time delay difference refers to the deviation value between the signal propagation time delay of the target UAV in the current obstacle distribution scenario and the reference time delay of the reference model in the RF fingerprint database under the same reflection path. The frequency-domain distortion correlation degree refers to the coincidence degree between the signal frequency-domain distortion position of the target UAV in the electromagnetic interference source distribution scenario and the reference distortion position of the reference model under the same interference conditions.

[0032] In this embodiment, based on the matching result (candidate model set and matching degree ranking) generated in step 102, the RF fingerprint reference data corresponding to the candidate model with the highest ranking is extracted. First, obtain the reference signal propagation time delay of this model under the reflection path of typical obstacles from the RF fingerprint reference data (for example, the reference time delay of the high-rise building reflection path is 1.2 ms). By comparing the actual time delay of the target UAV in the same reflection path in the current scenario (for example, 1.5 ms), calculate the difference between the two (0.3 ms) as the signal propagation time delay difference parameter. Second, obtain the reference frequency-domain distortion position of this model under the typical electromagnetic interference distribution from the RF fingerprint reference data (for example, the distortion interval in the 2.4-GHz frequency band caused by WiFi interference). By analyzing the actual distortion position of the target UAV under the current interference source distribution (for example, the distortion in the 2.4 - 2.42-GHz interval), calculate the coincidence ratio with the reference distortion interval (for example, 85%) as the frequency-domain distortion correlation degree parameter. Finally, integrate the time delay difference parameter and the frequency-domain correlation degree parameter into the model-related parameters for subsequent calculation of the environmental adaptation factor.

[0033] For example, in the scenario of monitoring urban building clusters, the matching result in step 102 determines that model C is the candidate model with the highest matching degree. The reference delay (1.2 ms) of model C under the reflection path of high-density buildings is extracted from its RF fingerprint reference data, and compared with the actual delay of the same path in the current scenario (due to the enhancement of multipath reflection, the measured delay is 1.55 ms), and the signal propagation delay difference parameter of 0.35 ms is calculated. At the same time, the reference distortion interval (2.4 - 2.41 GHz) of model C under the interference of 2.4 GHz band WiFi is extracted from the reference data, and the actual distortion interval of the target UAV in the current scenario is 2.4 - 2.415 GHz (the interval is extended by 5 MHz due to the change in the distribution of interference sources), and the coincidence ratio is calculated as 80% (overlapping interval 10 MHz / total distortion interval 15 MHz), which is used as the frequency domain distortion correlation parameter. After integrating the above parameters, the model correlation parameter includes a delay difference of 0.35 ms and a frequency domain correlation of 80%. Step 104, generate an identification decision including a real-time environment adaptation factor based on the model correlation parameter, and output the target UAV model result corresponding to the identification decision; In this step, the real-time environment adaptation factor refers to a weight correction coefficient generated based on the dynamic parameters of the current scenario (the change rate of the obstacle position and the fluctuation amplitude of the interference source intensity), which is used to compensate for the environmental offset of the RF fingerprint library matching result; the dynamic compensation mechanism refers to the process of secondarily adjusting the matching degree of the candidate model according to the environment adaptation factor; the consistency verification mechanism refers to the step of logically verifying the consistency between the identifier of the finally matched model and the pre-stored model features in the RF fingerprint library.

[0034] In this embodiment, first, the signal propagation delay difference and the frequency domain distortion correlation are extracted from the model correlation parameter and input into the environment adaptation factor calculation model. Based on the real-time tracking data (the change rate of the obstacle position is 5 m / s) and the fluctuation data, this model generates a delay correction factor and a frequency domain correction factor through a preset rate-weight mapping table and an amplitude-weight piecewise function. Then, the correction factors are dynamically compensated with the matching degree of the candidate model: the delay correction factor is multiplied by the initial matching degree of model C and the delay difference parameter to obtain the delay compensation matching degree; the frequency domain correction factor is multiplied by the frequency domain correlation parameter to obtain the frequency domain compensation matching degree. The two compensation results are combined according to a preset ratio to generate a comprehensive compensation matching degree. According to the ranking of the comprehensive matching degree, model C is still the highest matching result. Finally, the identifier of model C (such as serial number UA-C-001) is verified for consistency with the pre-stored model features (flight altitude-signal intensity curve, standard modulation waveform) in the RF fingerprint library, and it is confirmed that the signal intensity fluctuation mode error < 5% and the modulation distortion coincidence degree > 75%. After the verification passes, model C is output as the target UAV model result.

[0035] For example, in the scenario of monitoring urban building clusters, the model correlation parameters (time delay difference of 0.35 ms and frequency domain correlation degree of 80%) extracted in step 103 are input into the environmental adaptation factor calculation model. According to the real-time obstacle position change rate of 5 m / s (exceeding the threshold of 4 m / s), the time delay weight linear increase coefficient of 1.2 is triggered, and the basic weight of 1.0 is superimposed to generate the time delay correction factor of 1.2; at the same time, the electromagnetic interference source fluctuation amplitude is ±3 dB (exceeding the threshold of 2 dB), and the frequency domain weight piecewise increasing gradient of 0.1 / dB is triggered according to the difference of 1 dB to generate the frequency domain correction factor of 1.1. Compensation calculation is performed on the initial matching degree of 0.82 for model C: time delay compensation matching degree = 0.82×1.2×(1 - 0.35 ms / reference time delay of 1.2 ms) = 0.82×1.2×0.708≈0.698; frequency domain compensation matching degree = 0.698×1.1×(80% / 100%) = 0.698×1.1×0.8≈0.614. The combined comprehensive matching degree of 0.614 is still higher than that of other candidate models. The identifier of model C is verified with the pre-stored features in the fingerprint database: the measured signal strength fluctuation pattern error is 3% (<5%), and the modulation distortion coincidence degree is 78% (>75%). After passing the verification, model C is output as the final recognition result, which is consistent with the actual model of the drone in the scenario.

[0036] To solve the problem of radio frequency signal noise component interference caused by complex electromagnetic interference and obstacle reflection, and improve the signal noise reduction quality in dynamic open scenarios, in some embodiments, according to step 102, the radio frequency signal generated by the target drone during flight is received, and dynamic noise reduction processing is performed on the radio frequency signal to generate a noise-reduced radio frequency signal, including: Step 201, the original radio frequency signal of the target drone is captured in real time by a multi-channel signal acquisition device deployed in the monitoring area, and the original radio frequency signal contains noise components caused by electromagnetic interference sources and obstacle reflections in the current scenario; In this step, the multi-channel signal acquisition device refers to a plurality of radio frequency receiving modules distributedly deployed in the monitoring area, which are used to synchronously collect the radio frequency signals transmitted by the drone from different spatial positions; the noise components in the original radio frequency signal include broadband noise caused by electromagnetic interference sources (such as WiFi routers, base stations), and pulse noise formed by multipath interference caused by obstacle reflections.

[0037] In this embodiment, in the urban building complex monitoring scenario, multi-channel signal acquisition devices deployed on the rooftops of three adjacent buildings (covering the east, west, and south directions) are simultaneously activated to capture in real time the radio frequency signals emitted by the target unmanned aerial vehicle (UAV) during flight. The device on the east side is close to a commercial area WiFi hotspot, and the received signal is mixed with strong broadband noise in the 2.4 GHz frequency band; the device on the west side receives impulse noise propagated through multiple paths (manifested as signal glitches with sudden amplitude changes) due to the reflection of the high-rise glass curtain wall; the device on the south side has less obstruction, and the noise component in the received signal is relatively weak. After the signals of each channel are calibrated by a synchronous clock, they are aggregated to a central processing unit to form an original radio frequency signal dataset containing spatial difference characteristics.

[0038] Step 202: Monitor the environmental parameters in the original radio frequency signal, where the environmental parameters include the real-time interference intensity distribution and the obstacle density distribution. In this step, the real-time interference intensity distribution refers to the spatial distribution characteristics of electromagnetic interference energy in different frequency bands, which are extracted through spectrum analysis; the obstacle density distribution refers to the number and spatial density of obstacle reflection paths per unit area, which are measured by lidar or infrared sensors.

[0039] In this embodiment, channel-by-channel spectrum analysis is performed on the original radio frequency signals aggregated in step 201: The signal of the east side channel generates an energy distribution map in the 2.4 - 2.48 GHz frequency band through fast Fourier transform, showing that the peak of the WiFi interference intensity is concentrated at 2.412 GHz (-22 dBm); the signal of the west side channel is statistically analyzed in the time domain amplitude to identify the dense interval of impulse noise (20 - 30 mutations per second); meanwhile, the lidar scans the glass curtain wall area on the west side of the building complex, and the feedback shows that the density of obstacle reflection paths is 5 paths per square meter. After integrating the above data, the environmental parameters are generated: the real-time interference intensity distribution is marked as {2.4 GHz frequency band: strong, 5.8 GHz frequency band: weak}, and the obstacle density distribution is marked as {west side area: high, east / south side area: medium}.

[0040] Step 203: Select the signal filtering intensity level according to the environmental parameters. If the real-time interference intensity is higher than the set range and the obstacle density is higher than the set threshold, then multi-stage cascaded filtering is used to suppress the noise layer by layer for the original radio frequency signal; if the real-time interference intensity is lower than the set range, then single-stage filtering is used to suppress the noise for the original radio frequency signal, and the filtered radio frequency signal is obtained. In this step, multi-stage cascaded filtering refers to operations such as frequency band isolation, amplitude truncation, and frequency domain notch filtering performed sequentially based on different noise types (broadband noise, impulse noise, etc.); single-stage filtering refers to band-pass or notch filtering for a single noise type (such as main frequency band interference); the signal filtering intensity level refers to the strategy of dynamically selecting the filtering level and suppression amplitude according to the environmental parameters.

[0041] In this embodiment, according to the environmental parameters monitored in step 202 (the real-time interference intensity distribution is marked as strong interference in the 2.4GHz band, and the obstacle density distribution is marked as high density on the west side), it is determined that the multi-stage cascaded filtering conditions are met (interference intensity > threshold, obstacle density > threshold). First, perform the first-stage frequency band isolation filtering on the original radio frequency signal of the east channel: truncate the stray signals outside the 2.4GHz main frequency band (retain 2.4 - 2.48GHz), and suppress the broadband noise caused by WiFi; secondly, perform the second-stage amplitude truncation filtering on the west channel signal: detect the amplitude mutation interval of the pulse noise (>30dB), and replace it with the mean value of the adjacent section to suppress the multipath reflection noise; finally, perform the third-stage frequency domain notch filtering on the residual narrowband interference: generate notch windows at WiFi hotspot frequencies such as 2.412GHz and 2.422GHz to further filter out the residual interference. Since the interference intensity of the south channel is low (< threshold), only perform single-stage main frequency band band-pass filtering.

[0042] Step 204, perform signal integrity restoration processing on the filtered radio frequency signal to generate a noise-reduced radio frequency signal; In this step, the signal integrity restoration processing refers to repairing the signal waveform distortion caused by the filtering operation, including repairing the phase jump break, compensating for the lost amplitude mutation, and reconstructing the waveform continuity; the key feature retention refers to restoring the signal intensity mutation section and the phase jump section that are strongly related to the flight state of the drone (such as altitude, attitude).

[0043] In this embodiment, perform waveform integrity detection on the multi-channel signals filtered in step 203: the filtered signal of the east channel has a phase jump break at the 2.48GHz boundary due to frequency band truncation, and generate a linear interpolation signal through the phase slope calculation of adjacent frequency bands to repair the jump; the filtered signal of the west channel has some lost amplitude mutations due to amplitude truncation, and reconstruct the peak and valley values of the mutation based on the amplitude change trend of the front and rear sections; the signal integrity of the south channel is good, and only the waveform of the smoothing filter transition area needs to be processed. After the processing is completed, extract the intensity mutation section (such as the intensity drops by 20dB during the hovering stage) and the complete phase jump section (such as the jump interval is shortened to 5ms during the climbing stage) related to the flight altitude in the signals of each channel, and merge them to generate a noise-reduced radio frequency signal.

[0044] In order to improve the matching robustness of multi-dimensional signal features in complex scenarios, in some embodiments, according to what is described in step 102, separate the combined data including the signal intensity fluctuation pattern, the phase jump frequency, and the modulation distortion characteristics from the noise-reduced radio frequency signal, and perform layer-by-layer matching of the combined data with a pre-constructed radio frequency fingerprint library to generate a matching result, including: Step 301: Divide the noise-reduced RF signal into multiple signal segments according to a preset time window, and extract from each signal segment the fluctuation pattern of the signal intensity varying with the flight altitude, the frequency distribution of the phase jump intervals, and the position characteristics of the modulation waveform distortion. In this step, the time window division means dividing the continuous RF signal into signal segments of a fixed duration for analyzing the periodic characteristics of the flight state; the signal intensity fluctuation pattern refers to the law of sequential peak-valley changes in the signal intensity as the UAV flight altitude rises and falls; the phase jump frequency distribution represents the number of discontinuous phase jumps within the signal segment and the statistical characteristics of their time intervals; the position characteristics of the modulation waveform distortion refer to the position and length ratio of the distorted section relative to the standard waveform on the time axis in the signal modulation waveform.

[0045] In this embodiment, the noise-reduced RF signal is first divided into continuous signal segments of a preset time window. Each signal segment corresponds to a different stage of the UAV flight state. The signal intensity curve is matched with the flight altitude change curve through a time series alignment algorithm, and the similarity in peak-valley positions, amplitude differences, and change rates between the two is analyzed to extract the fluctuation pattern characteristics. Subsequently, the number of phase jumps within each signal segment is counted, and the distribution characteristics of the jump intervals are calculated through a sliding window to generate the frequency distribution characteristics. At the same time, the signal segment modulation waveform is compared point by point with the standard waveform, and the position and length ratio of the distorted section are marked to extract the position characteristics. Finally, each signal segment outputs a feature vector containing the fluctuation pattern, frequency distribution, and distortion position.

[0046] Step 302: Fuse the fluctuation pattern, frequency distribution, and position characteristics into combined data according to a preset weight ratio. In this step, the preset weight ratio means assigning different importance weights to the three types of characteristics according to the scenario requirements (such as a signal intensity fluctuation weight of 5, a phase jump frequency weight of 3, and a modulation distortion weight of 2); the combined data is a multi-dimensional data set obtained by weighted fusion of the three types of feature vectors and is used for the matching calculation of the RF fingerprint library.

[0047] In this embodiment, the extracted fluctuation pattern, frequency distribution, and position characteristics are respectively normalized and weighted and fused according to a preset weight ratio. Among them, the fluctuation pattern calculates the matching degree with the reference curve through the similarity difference, the frequency distribution determines the coincidence degree with the reference distribution through the overlapping area analysis, and the position characteristics evaluate the consistency of the distorted section through the deviation tolerance. After the three are weighted and summed, a combined data unit is generated. The combined data units of all signal segments are spliced in chronological order into the complete combined data and input into the RF fingerprint library for matching.

[0048] Step 303: Calculate the similarity of the combined data with the first-layer signal intensity fluctuation pattern reference set stored in the RF fingerprint library in sequence, and screen out the candidate model set with a similarity higher than the first threshold. In this step, the first-layer signal intensity fluctuation pattern reference set refers to the reference fluctuation curves of the signal intensity of each model of unmanned aerial vehicle (UAV) stored in the radio frequency fingerprint database with respect to the change in flight altitude under typical scenarios; the similarity calculation refers to quantifying the matching degree between the target signal fluctuation pattern and the reference curve through the dynamic time warping or waveform difference accumulation algorithm.

[0049] In this embodiment, the fluctuation pattern feature vectors of all signal segments in the combined data generated in step 302 are sequentially subjected to time series alignment and difference accumulation calculation with the first-layer reference fluctuation curves of each model in the radio frequency fingerprint database to generate the similarity scores for each model; the models with similarity scores higher than the preset threshold are screened to form a candidate set, and the corresponding second-layer phase jump frequency reference set and third-layer modulation distortion reference set are loaded into the next matching stage.

[0050] Step 304, jointly match the frequency distribution and position features in the combined data with the second-layer phase jump frequency reference set and modulation distortion reference set corresponding to the candidate model set to generate a matching result including the matching degree ranking of candidate models; In this step, the second-layer phase jump frequency reference set refers to the histogram of the phase jump interval distribution of candidate models under typical scenarios; the modulation distortion reference set refers to the reference data of the proportion of the modulation waveform distortion position and length of candidate models under the same interference conditions; the joint matching refers to the weighted comprehensive evaluation through the calculation of the overlapping area of the frequency distribution and the analysis of the offset of the distortion position.

[0051] In this embodiment, the second-layer reference data (phase jump frequency distribution histogram) and third-layer reference data (modulation distortion position) of each model are extracted from the candidate model set, the overlapping area between the frequency distribution feature in the target combined data and the phase jump reference histogram is calculated, and at the same time, the offset analysis is performed on the position feature and the modulation distortion reference. The two results are weighted and summed according to the preset weight to generate the comprehensive matching degree of each candidate model, and the final matching result is formed by arranging them in descending order.

[0052] To address the differential effects of obstacle reflection paths and electromagnetic interference source distributions on signal characteristics in dynamic scenarios, in some embodiments, according to step 103, the model-related parameters of the target UAV are extracted from the matching result. The model-related parameters include the signal propagation delay difference corresponding to a specific obstacle reflection path and the frequency domain distortion correlation degree corresponding to the electromagnetic interference source distribution, including: Step 401, according to the candidate model matching degree ranking in the matching result, extract the radio frequency fingerprint reference data corresponding to the candidate model with the highest ranking; In this step, the candidate model matching degree ranking refers to the model list arranged in descending order of the comprehensive matching degree generated through a hierarchical matching process (signal strength fluctuation → phase jump → modulation distortion); the radio frequency fingerprint reference data refers to the benchmark feature data of the target model in a typical scenario pre-stored in the radio frequency fingerprint library, including signal propagation delay, phase jump frequency, and modulation distortion position parameters.

[0053] In this embodiment, based on the matching results generated in step 304 (such as model C > model E > model F), the radio frequency fingerprint reference data corresponding to the candidate model with the highest ranking (model C) is extracted, including the benchmark signal propagation delay, the distribution histogram of phase jump intervals, and the benchmark interval of modulation waveform distortion positions of this model in a typical obstacle distribution scenario, providing a benchmark comparison basis for subsequent model association parameter extraction.

[0054] Step 402: Obtain the signal propagation delay difference parameter associated with the obstacle distribution in the current scenario from the radio frequency fingerprint reference data; In this step, the signal propagation delay difference parameter refers to the deviation value between the actual signal propagation delay of the target UAV under the current obstacle distribution and the benchmark delay in the same scenario in the radio frequency fingerprint reference data, which is used to quantify the influence degree of environmental dynamic changes on the signal propagation path.

[0055] In this embodiment, the benchmark signal propagation delay (such as the benchmark delay) of model C in a typical high-rise building dense scenario is obtained from the radio frequency fingerprint reference data of model C. By comparing the actual delay (such as the actual delay) under the same obstacle distribution conditions in the current scenario, the difference between the two is calculated as the delay difference parameter, reflecting the influence of the current dynamic change of obstacle positions (such as newly added temporary obstacles) on the signal propagation path.

[0056] Step 403: Obtain the frequency domain distortion correlation degree parameter associated with the electromagnetic interference source distribution from the radio frequency fingerprint reference data; In this step, the frequency domain distortion correlation degree parameter refers to the coincidence ratio between the actual signal frequency domain distortion position of the target UAV under the current electromagnetic interference source distribution and the benchmark distortion position under the same interference conditions in the radio frequency fingerprint reference data, which is used to quantify the interference degree of electromagnetic interference dynamic changes on the signal frequency domain characteristics.

[0057] In this embodiment, the benchmark frequency domain distortion position (such as a specific frequency band interval) of model C in a typical electromagnetic interference scenario (such as dense distribution of WiFi hotspots) is extracted from the radio frequency fingerprint reference data of model C. By comparing the actual distortion position (such as frequency band expansion or offset) caused by the electromagnetic interference source distribution (such as newly added temporary communication base stations) in the current scenario, the coincidence ratio between the actual distortion interval and the benchmark interval is calculated to generate the frequency domain distortion correlation degree parameter, reflecting the offset influence of the current interference fluctuation on the signal characteristics.

[0058] Step 404: Integrate the signal propagation delay difference parameter and the frequency-domain distortion correlation parameter into a model association parameter for the target UAV; In this step, the model association parameter refers to a composite parameter that fuses the signal propagation delay difference (reflecting the dynamic impact of obstacles) and the frequency-domain distortion correlation (reflecting the dynamic impact of electromagnetic interference), and is used to quantify the combined effect of the current scene environment fluctuations on the signal characteristics of the target UAV.

[0059] In this embodiment, the signal propagation delay difference parameter extracted in step 402 (such as the delay deviation caused by the addition of obstacles) and the frequency-domain distortion correlation parameter generated in step 403 (such as the decrease in coincidence degree caused by interference spread) are integrated in a preset format (such as a vector or matrix) to form a composite parameter set containing the delay difference weight and the frequency-domain correlation weight, providing input for the dynamic calculation of the subsequent real-time environment adaptation factor, and correlating with the reference data correction logic of the target model in the RF fingerprint library.

[0060] To improve the real-time performance and accuracy of the recognition decision under dynamic environmental fluctuations, in some embodiments, according to step 104, an recognition decision including a real-time environment adaptation factor is generated based on the model association parameter, and the target UAV model result corresponding to the recognition decision is output, including: Step 501: Obtain real-time tracking data on the dynamic change of the obstacle position and real-time fluctuation data on the intensity of the electromagnetic interference source in the current scene; In this step, the real-time tracking data refers to the information on the moving speed and distribution density change of the obstacle position collected in real time by multi-source sensors (such as lidar, camera); the real-time fluctuation data refers to the dynamic change characteristics of the energy intensity of the electromagnetic interference source in the frequency domain and time domain monitored by a spectrum analysis device.

[0061] In this embodiment, a lidar array deployed in the monitoring area scans the obstacle position in real time to generate a heat map containing the moving speed (such as the displacement amount of the obstacle per unit time) and density change; at the same time, a distributed spectrum monitoring network samples the frequency band energy of the electromagnetic interference source (such as a WiFi router, base station) to generate the fluctuation amplitude (such as the peak energy change range) and the time series fluctuation trend of the interference intensity at each frequency point; after the above data is aligned in space and time, an environment parameter set that is strongly correlated with the current scene dynamics is formed, providing input for the adjustment of the weight coefficient.

[0062] Step 502: Adjust the weight coefficient of the signal propagation delay difference parameter according to the real-time tracking data, and adjust the weight coefficient of the frequency-domain distortion correlation parameter according to the real-time fluctuation data; In this step, the weight coefficient adjustment refers to adaptively modifying the decision weights of the model association parameters (time delay difference, frequency domain correlation degree) based on the environmental dynamic parameters (obstacle movement rate, interference fluctuation amplitude) to enhance the response sensitivity to scene changes.

[0063] In this embodiment, according to the comparison relationship between the obstacle movement rate in the real-time tracking data and the preset rate threshold, a linear enhancement rule is used to adjust the weight coefficient of the time delay difference parameter: if the rate exceeds the threshold, the weight is increased according to the ratio of the rate to the threshold; at the same time, according to the difference between the interference intensity fluctuation amplitude in the real-time fluctuation data and the preset amplitude threshold, a piecewise increasing rule is used to adjust the weight coefficient of the frequency domain correlation degree parameter: if the amplitude exceeds the threshold, the weight coefficient gradient is increased in segments according to the difference, and finally an updated weight adapted to the environment is generated.

[0064] Step 503, input the adjusted signal propagation time delay difference parameter and the adjusted frequency domain distortion correlation degree parameter into the environmental adaptation factor calculation model to generate a real-time environmental adaptation factor for correcting the matching degree of candidate models in the radio frequency fingerprint library; In this step, the environmental adaptation factor calculation model refers to the calculation logic of weighted fusion of the signal propagation time delay difference parameter and the frequency domain distortion correlation degree parameter based on the dynamic weight coefficients (time delay difference weight, frequency domain correlation degree weight) to generate a dynamic correction factor; the real-time environmental adaptation factor is a correction coefficient that quantifies the influence of the current environmental fluctuation on the matching result.

[0065] In this embodiment, the adjusted time delay difference weight coefficient and the frequency domain correlation degree weight coefficient in step 502 are respectively weighted and superimposed with the corresponding time delay difference parameter (such as the time delay deviation caused by obstacle movement) and the frequency domain distortion correlation degree parameter (such as the decrease in coincidence degree caused by interference spread) to generate a time delay correction factor and a frequency domain correction factor; the two are combined into a real-time environmental adaptation factor through a preset fusion rule (such as the time delay correction factor accounting for 60% and the frequency domain correction factor accounting for 40%) to dynamically correct the matching degree of candidate models in the radio frequency fingerprint library.

[0066] Step 504, dynamically compensate the matching degree of the candidate models in the matching result according to the real-time environmental adaptation factor to generate an identification decision including the final matching model; In this step, the dynamic compensation refers to the associated calculation of the real-time environmental adaptation factor and the initial matching degree of the candidate model to eliminate the matching degree deviation caused by environmental fluctuations; the identification decision is the final model determination result generated by sorting the comprehensive matching degree after compensation.

[0067] In this embodiment, the initial matching degree of the candidate model is extracted from the matching results (e.g., the initial matching degree of model C is 0.82). The real-time environment adaptation factor is multiplied step by step with the initial matching degree, the time-delay difference parameter, and the frequency-domain correlation degree parameter to generate the time-delay compensated matching degree and the frequency-domain compensated matching degree. After the two are combined into the comprehensive compensated matching degree according to a preset ratio, the candidate models are sorted in descending order, and the model with the highest matching degree is selected as the final result, and its consistency is verified with the model features in the radio frequency fingerprint library to ensure the reliability of the output.

[0068] Step 505: Verify the consistency between the model identifier of the final matching model in the recognition decision and the pre-stored model feature description in the radio frequency fingerprint library. When the consistency verification passes, output the final matching model as the target UAV model result. In this step, the consistency verification refers to the process of comparing the measured feature data of the final matching model (such as the signal strength fluctuation pattern, modulation distortion position) with the reference features of this model pre-stored in the radio frequency fingerprint library in multiple dimensions to verify its logical consistency. The model feature description refers to the unique identifier of the UAV model stored in the radio frequency fingerprint library and the corresponding set of reference feature parameters (such as the standard fluctuation curve, modulation waveform template).

[0069] In this embodiment, the pre-stored model feature description of the final matching model is extracted from the radio frequency fingerprint library, including the standard signal strength fluctuation curve, the phase jump interval distribution histogram, and the modulation waveform template. The measured signal strength fluctuation pattern of model C is aligned with the standard fluctuation curve in time series and the peak-valley position error is calculated. At the same time, the interval overlap degree analysis is performed on the measured modulation distortion position and the template. If the fluctuation pattern error is less than the preset threshold and the modulation distortion overlap degree is higher than the minimum requirement, it is determined that the consistency verification passes, and the model result is output; otherwise, an abnormal alarm is triggered and the matching process is executed again.

[0070] To address the different requirements for the weight coefficients due to the dynamic change of the obstacle position and the fluctuation of the interference source intensity, in some embodiments, according to step 502, the weight coefficient of the signal propagation time-delay difference parameter is adjusted according to the real-time tracking data, and the weight coefficient of the frequency-domain distortion correlation degree parameter is adjusted according to the real-time fluctuation data, including: Step 601: Dynamically adjust the weight coefficient of the signal propagation time-delay difference parameter according to the dynamic rate of the obstacle position change in the real-time tracking data. When the dynamic rate exceeds the preset rate threshold, linearly enhance the weight coefficient according to the ratio of the dynamic rate to the preset rate threshold to generate the adjusted time-delay weight coefficient. In this embodiment, the dynamic rate of the obstacle position change (such as the displacement of the obstacle within a unit time) is extracted through real-time tracking data and compared with a preset rate threshold. If the dynamic rate exceeds the threshold, the weight coefficient of the time-delay difference parameter is linearly enhanced according to the ratio of the rate to the threshold. For example, when the rate exceeds the threshold, the weight coefficient is adjusted positively according to the proportion by which the rate exceeds the threshold. If it does not exceed the threshold, the original weight coefficient remains unchanged, ensuring that the weight of the time-delay difference parameter dynamically reflects the actual impact of the obstacle movement on the signal propagation path.

[0071] Step 602: Dynamically adjust the weight coefficient of the frequency-domain distortion correlation degree parameter according to the fluctuation amplitude of the electromagnetic interference source intensity in the real-time fluctuation data. When the fluctuation amplitude exceeds a preset amplitude threshold, increment the weight coefficient in segments according to the difference between the fluctuation amplitude and the amplitude threshold to generate an adjusted frequency-domain weight coefficient. In this embodiment, the fluctuation amplitude of the electromagnetic interference source intensity (such as the difference between the peak and valley of the interference energy) is extracted through real-time fluctuation data and compared with a preset amplitude threshold. If the fluctuation amplitude exceeds the threshold, the weight coefficient of the frequency-domain distortion correlation degree parameter is incremented in segments according to the difference between the amplitude and the threshold. For example, for each amplitude interval exceeding the threshold, the weight coefficient is incremented according to a gradient. If it does not exceed the threshold, the original weight coefficient is maintained, ensuring that the weight of the frequency-domain parameter is adaptively adjusted according to the degree of interference fluctuation.

[0072] To address the limitation of the single-dimensional correction of the environmental adaptation factor on the matching result, in some embodiments, according to what is described in step 504, the candidate model matching degree in the matching result is dynamically compensated according to the real-time environmental adaptation factor to generate an identification decision including the final matching model, including: Step 701: Decompose the real-time environmental adaptation factor into a time-delay correction factor and a frequency-domain correction factor. In this embodiment, the real-time environmental adaptation factor is decomposed into a time-delay correction factor and a frequency-domain correction factor through a preset weight separation rule. Among them, the time-delay correction factor is generated from the dynamic weight adjustment result of the signal propagation time-delay difference parameter, reflecting the impact of the dynamic change of the obstacle position on the matching degree. The frequency-domain correction factor is generated from the dynamic weight adjustment result of the frequency-domain distortion correlation degree parameter, reflecting the correction requirement of the electromagnetic interference fluctuation on the matching degree. The two are decoupled from the environmental adaptation factor through a weighted ratio allocation mechanism to form independent correction parameters to support the step-by-step compensation logic.

[0073] Step 702: Extract the initial matching degree of the candidate model set from the matching result, and multiply each item of the time-delay correction factor by the signal propagation time-delay difference parameter corresponding to each candidate model in the initial matching degree to generate a time-delay compensated matching degree. In this embodiment, the initial matching degree of the candidate model is extracted from the matching results (the hierarchical matching score based on the signal strength fluctuation pattern and the phase jump frequency), and the time delay correction factor is multiplied item by item with the signal propagation time delay difference parameters corresponding to each candidate model (such as the time delay deviation value caused by the dynamic obstacle) to generate the matching degree after time delay compensation. The greater the time delay difference parameter (the significant influence of the obstacle movement), the stronger the compensation effect of the time delay correction factor on the matching degree, thereby improving the matching priority of the models sensitive to environmental dynamics.

[0074] Step 703: Multiply the frequency domain correction factor item by item with the frequency domain distortion correlation degree parameters corresponding to each candidate model in the matching degree after time delay compensation to generate the matching degree after frequency domain compensation. In this embodiment, the frequency domain distortion correlation degree parameters of each model are extracted from the candidate model set (reflecting the coincidence degree between the signal distortion and the reference under the electromagnetic interference fluctuation), and the frequency domain correction factor (the weight adjustment result triggered by the interference fluctuation amplitude) is multiplied item by item with the frequency domain correlation degree parameters of each model to generate the matching degree after frequency domain compensation. The lower the frequency domain correlation degree parameter (the decrease in the coincidence degree of the distortion caused by the interference), the stronger the compensation effect of the frequency domain correction factor on the matching degree, thereby preferentially improving the matching weight of the anti-interference models.

[0075] Step 704: Combine and process the matching degree after time delay compensation and the matching degree after frequency domain compensation according to a preset ratio to generate the comprehensive compensation matching degree of each candidate model. In this embodiment, according to the dynamic characteristics of the scenario, the combined weight ratio of time delay compensation and frequency domain compensation is preset (such as 60% for time delay and 40% for frequency domain), and the matching degree after time delay compensation (reflecting the influence of obstacle movement) and the matching degree after frequency domain compensation (reflecting the influence of interference fluctuation) are weighted and combined to generate the comprehensive compensation matching degree. The combined correction ability for complex environmental dynamics is enhanced through the weighted combination mechanism, and finally, the candidate models are sorted in descending order according to the comprehensive matching degree, and the optimal recognition result is output.

[0076] Step 705: Sort the candidate model set in descending order according to the comprehensive compensation matching degree, select the candidate model with the highest ranking as the final matching model, and generate an identification decision including the final matching model. In this step, the comprehensive compensation matching degree refers to the comprehensive score of the candidate model generated by combining the time delay correction factor and the frequency domain correction factor through step-by-step compensation and preset ratio combination. The descending order sorting means sorting the candidate models from high to low according to the comprehensive compensation matching degree. The identification decision is a determination result data set including the final matching model identification and the matching degree verification information.

[0077] In this embodiment, the system numerically compares the comprehensive compensation matching degrees of each model in the candidate model set, generates a sorted list according to the descending order rule, and selects the model with the highest ranking as the final matching model. At the same time, the identifier, comprehensive compensation matching degree, time delay / frequency domain compensation parameters, and environmental adaptation factors of the final matching model are integrated into structured recognition decision data for subsequent consistency verification and result output, ensuring the traceability and reliability of the decision-making process.

[0078] Figure 2 FIG. is a schematic structural diagram of a specific UAV model rapid recognition system based on a radio frequency fingerprint library provided by an embodiment of the present application. As Figure 2 shown, the system includes: A receiving module, configured to receive the radio frequency signals generated during the flight of the target UAV, and perform dynamic noise reduction processing on the radio frequency signals to generate noise-reduced radio frequency signals; A matching module, configured to separate combined data including signal intensity fluctuation patterns, phase jump frequencies, and modulation distortion characteristics from the noise-reduced radio frequency signals, and perform layer-by-layer matching of the combined data with a pre-constructed radio frequency fingerprint library to generate a matching result; An extraction module, configured to extract model-related parameters of the target UAV from the matching result, where the model-related parameters include signal propagation time delay differences corresponding to specific obstacle reflection paths and frequency domain distortion correlation degrees corresponding to the distribution of electromagnetic interference sources; An output module, configured to generate an identification decision including a real-time environmental adaptation factor based on the model-related parameters, and output the target UAV model result corresponding to the identification decision.

[0079] Figure 2 The specific UAV model rapid recognition system based on a radio frequency fingerprint library can execute Figure 1 the specific UAV model rapid recognition method described in the embodiment shown. The implementation principle and technical effects will not be elaborated here. For the specific UAV model rapid recognition system based on a radio frequency fingerprint library in the above embodiment, the specific manners in which each module and unit perform operations have been described in detail in the embodiment related to the method, and will not be elaborated here.

[0080] In a possible design, Figure 2 the specific UAV model rapid recognition system described in the embodiment shown can be implemented as a computing device. As Figure 3 shown, the computing device may include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, where the one or more computer instructions are called and executed by the processing component 32.

[0081] The processing component 32 is used for the above-mentioned Figure 1 A method for quickly identifying a specific UAV model based on a radio frequency fingerprint library in the above-mentioned embodiment.

[0082] Among them, the processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-mentioned method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components for executing the above-mentioned method.

[0083] The storage component 31 is configured to store various types of data to support the operation of the terminal. The storage component may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0084] Of course, the computing device may necessarily further include other components, such as an input / output interface, a display component, a communication component, etc.

[0085] The input / output interface provides an interface between the processing component and the peripheral interface module, and the above-mentioned peripheral interface module may be an output device, an input device, etc.

[0086] The communication component is configured to facilitate communication between the computing device and other devices in a wired or wireless manner, etc.

[0087] Among them, the computing device may be a physical device or an elastic computing host provided by a cloud computing platform, etc. At this time, the computing device may refer to a cloud server, and the above-mentioned processing component, storage component, etc. may be basic server resources rented or purchased from a cloud computing platform.

[0088] The embodiment of the present application also provides a computer storage medium storing a computer program, and when the computer program is executed by a computer, it can implement the above-mentioned Figure 1 A method for quickly identifying a specific UAV model based on a radio frequency fingerprint library in the above-mentioned embodiment.

[0089] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, devices and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0090] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0091] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0092] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of each embodiment of the present application.

Claims

1. A rapid identification method for a specific UAV model based on a radio frequency fingerprint library, characterized in that Including: Receiving the radio frequency signal generated by the target UAV during flight, and performing dynamic noise reduction processing on the radio frequency signal to generate a noise-reduced radio frequency signal; Separating combined data including signal strength fluctuation patterns, phase jump frequencies, and modulation distortion characteristics from the noise-reduced radio frequency signal, and performing layer-by-layer matching of the combined data with a pre-constructed radio frequency fingerprint library to generate a matching result; Extracting model-related parameters of the target UAV from the matching result, where the model-related parameters include signal propagation delay differences corresponding to specific obstacle reflection paths and frequency domain distortion correlation degrees corresponding to the distribution of electromagnetic interference sources; Generating an identification decision including a real-time environment adaptation factor based on the model-related parameters, and outputting the target UAV model result corresponding to the identification decision.

2. The method according to claim 1, wherein Receiving the radio frequency signal generated by the target UAV during flight, and performing dynamic noise reduction processing on the radio frequency signal to generate a noise-reduced radio frequency signal, including: Real-time capturing of the original radio frequency signal of the target UAV through a multi-channel signal acquisition device deployed in the monitoring area, where the original radio frequency signal contains noise components caused by electromagnetic interference sources and obstacle reflections in the current scene; Monitoring environmental parameters in the original radio frequency signal, where the environmental parameters include real-time interference intensity distribution and obstacle density distribution; Selecting a signal filtering intensity level according to the environmental parameters. If the real-time interference intensity is higher than the set range and the obstacle density is higher than the set threshold, then multi-stage cascaded filtering is used to perform layer-by-layer noise suppression on the original radio frequency signal. If the real-time interference intensity is lower than the set range, then single-stage filtering is used to perform noise suppression on the original radio frequency signal to obtain a filtered radio frequency signal; Performing signal integrity restoration processing on the filtered radio frequency signal to generate a noise-reduced radio frequency signal.

3. The method according to claim 1, wherein Separating combined data including signal strength fluctuation patterns, phase jump frequencies, and modulation distortion characteristics from the noise-reduced radio frequency signal, and performing layer-by-layer matching of the combined data with a pre-constructed radio frequency fingerprint library to generate a matching result, including: Dividing the noise-reduced radio frequency signal into multiple signal segments according to a preset time window, and extracting the fluctuation pattern of signal strength changing with flight altitude, the frequency distribution of phase jump intervals, and the position characteristics of modulation waveform distortion from each signal segment; Fusing the fluctuation pattern, frequency distribution, and position characteristics into combined data according to a preset weight ratio; Calculating the similarity of the combined data with the first-layer signal strength fluctuation pattern reference set stored in the radio frequency fingerprint library in sequence, and screening out a candidate model set with a similarity higher than the first threshold; Performing joint matching of the frequency distribution and position characteristics in the combined data with the second-layer phase jump frequency reference set and modulation distortion reference set corresponding to the candidate model set to generate a matching result including the ranking of candidate model matching degrees.

4. The method according to claim 1, characterized in that Extracting model-related parameters of the target UAV from the matching result, where the model-related parameters include signal propagation delay differences corresponding to specific obstacle reflection paths and frequency domain distortion correlation degrees corresponding to the distribution of electromagnetic interference sources, including: Sort according to the matching degree of candidate models in the matching result, and extract the radio frequency fingerprint reference data corresponding to the candidate model with the highest ranking; Obtain the signal propagation delay difference parameter associated with the obstacle distribution in the current scenario from the radio frequency fingerprint reference data; Obtain the frequency domain distortion correlation parameter associated with the electromagnetic interference source distribution from the radio frequency fingerprint reference data; Integrate the signal propagation delay difference parameter and the frequency domain distortion correlation parameter into the model association parameter of the target UAV.

5. The method according to claim 1, characterized in that, Generate an identification decision including a real-time environment adaptation factor based on the model association parameter, and output the target UAV model result corresponding to the identification decision, including: Obtain the real-time tracking data of the dynamic change of the obstacle position and the real-time fluctuation data of the electromagnetic interference source intensity in the current scenario; Adjust the weight coefficient of the signal propagation delay difference parameter according to the real-time tracking data, and adjust the weight coefficient of the frequency domain distortion correlation parameter according to the real-time fluctuation data; Input the adjusted signal propagation delay difference parameter and the adjusted frequency domain distortion correlation parameter into the environment adaptation factor calculation model to generate a real-time environment adaptation factor for correcting the matching degree of candidate models in the radio frequency fingerprint library; Dynamically compensate the matching degree of candidate models in the matching result according to the real-time environment adaptation factor to generate an identification decision including the final matching model; Verify the consistency between the model identifier of the final matching model in the identification decision and the model feature description pre-stored in the radio frequency fingerprint library. When the consistency verification passes, output the final matching model as the target UAV model result.

6. The method according to claim 5, wherein Adjust the weight coefficient of the signal propagation delay difference parameter according to the real-time tracking data, and adjust the weight coefficient of the frequency domain distortion correlation parameter according to the real-time fluctuation data, including: Dynamically adjust the weight coefficient of the signal propagation delay difference parameter according to the dynamic rate of the obstacle position change in the real-time tracking data. When the dynamic rate exceeds the preset rate threshold, linearly enhance the weight coefficient according to the ratio of the dynamic rate to the preset rate threshold to generate the adjusted delay weight coefficient; Dynamically adjust the weight coefficient of the frequency domain distortion correlation parameter according to the fluctuation amplitude of the electromagnetic interference source intensity in the real-time fluctuation data. When the fluctuation amplitude exceeds the preset amplitude threshold, increment the weight coefficient in segments according to the difference between the fluctuation amplitude and the amplitude threshold to generate the adjusted frequency domain weight coefficient.

7. The method according to claim 5, wherein Dynamically compensate the matching degree of candidate models in the matching result according to the real-time environment adaptation factor to generate an identification decision including the final matching model, including: Decompose the real-time environment adaptation factor into a delay correction factor and a frequency domain correction factor; Extract the initial matching degree of the candidate model set from the matching result, and multiply the delay correction factor item by item with the signal propagation delay difference parameter corresponding to each candidate model in the initial matching degree to generate the matching degree after delay compensation; Multiply the frequency domain correction factor item by item with the frequency domain distortion correlation parameter corresponding to each candidate model in the matching degree after delay compensation to generate the matching degree after frequency domain compensation; Combine the matching degree after the time delay compensation and the matching degree after the frequency domain compensation according to a preset ratio to generate the comprehensive compensation matching degree of each candidate model; Sort the candidate model set in descending order according to the comprehensive compensation matching degree, select the candidate model with the highest ranking as the final matching model, and generate an identification decision including the final matching model.

8. A fast recognition system for a specific UAV model based on a radio frequency fingerprint library, characterized in that Comprising: A receiving module, configured to receive a radio frequency signal generated during the flight of a target unmanned aerial vehicle, and perform dynamic noise reduction processing on the radio frequency signal to generate a noise-reduced radio frequency signal; A matching module, configured to separate combined data including a signal intensity fluctuation pattern, a phase hopping frequency, and a modulation distortion feature from the noise-reduced radio frequency signal, and perform layer-by-layer matching of the combined data with a pre-constructed radio frequency fingerprint library to generate a matching result; An extraction module, configured to extract model-related parameters of the target unmanned aerial vehicle from the matching result, where the model-related parameters include a signal propagation time delay difference corresponding to a reflection path of a specific obstacle and a frequency domain distortion correlation degree corresponding to the distribution of electromagnetic interference sources; An output module, configured to generate an identification decision including a real-time environment adaptation factor based on the model-related parameters, and output a target unmanned aerial vehicle model result corresponding to the identification decision.

9. A computing device, characterized in that, Comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for quickly identifying a specific unmanned aerial vehicle model based on a radio frequency fingerprint library according to any one of claims 1 to 7.

10. A computer storage medium, characterized in that, Stores a computer program, and when the computer program is executed by a computer, it implements a method for quickly identifying a specific unmanned aerial vehicle model based on a radio frequency fingerprint library according to any one of claims 1 to 7.

Images