A UAV target detection method based on a spatio-temporal synchronization correlation mechanism

Abstract

This application relates to a method for detecting unmanned aerial vehicle (UAV) targets based on a spatio-temporal synchronization correlation mechanism, belonging to the field of wireless detection technology. The method includes the following steps: S1. Establish a unified time reference system to ensure precise timestamp alignment of the IQ data streams collected by the two technologies, with an accuracy reaching the microsecond level; S2. Collect the UAV signal data obtained by protocol reverse parsing and radio frequency fingerprint recognition technologies; S3. Construct a three-dimensional correlation matrix including a time window, frequency domain position, and spatial angle of arrival, and perform spatio-temporal matching on the detection events within the synchronization window; S4. Analyze the matching results to identify real UAV targets. This application can effectively and accurately analyze UAV signals, improve the accuracy and reliability of UAV target detection, and reduce the possibility of multi-target false alarms.

Description

Technical Field

[0001] This application relates to the field of wireless detection technology, and in particular to a method for detecting drone targets based on a spatio-temporal synchronization correlation mechanism. The purpose of this application is to improve the accuracy of drone target detection and reduce the occurrence rate of multi-target false alarms. Background Art

[0002] In a radio spectrum monitoring system, the detection of drone targets usually adopts a dual technical path of protocol reverse parsing and radio frequency fingerprint recognition.

[0003] The protocol reverse parsing technology is based on the in-depth parsing of the drone communication protocol stack. By demodulating the physical layer signal, parsing the data link layer frame structure, decoding the application layer protocol, etc., core information such as the drone identity code (UID) and telemetry data can be obtained; the radio frequency fingerprint recognition technology realizes the classification and recognition of drone signals by analyzing the instantaneous spectrum characteristics of the signal (including center frequency, bandwidth, modulation type), long-term spectrum occupancy patterns (such as hopping period, TDMA time slot structure), and transmitter hardware characteristics (such as IQ imbalance, phase noise, etc., radio frequency fingerprints).

[0004] When a single drone target appears, due to the two technologies independently implementing signal parsing from the protocol layer and the physical layer respectively, multi-target false alarms may occur due to asynchronous detection mechanisms or lack of feature correlation; therefore, there is an urgent need for an effective mechanism to solve the multi-target false alarm problem. Summary of the Invention

[0005] The purpose of this application is to provide a method for detecting drone targets based on a spatio-temporal synchronization correlation mechanism. Through precise time synchronization and the construction of a three-dimensional correlation matrix, real drone targets can be effectively identified, and the probability of multi-target false alarms can be reduced.

[0006] A method for detecting drone targets based on a spatio-temporal synchronization correlation mechanism provided by this application adopts the following technical solutions: including the following steps:

[0007] S1. Establish a unified time reference system to ensure precise timestamp alignment of the IQ data streams collected by the two technologies, with an accuracy reaching the microsecond level;

[0008] S2. Collect the drone signal data obtained by the protocol reverse parsing and radio frequency fingerprint recognition technologies, including but not limited to the drone identity code, telemetry data, signal strength indication, and modulation method;

[0009] S3. Construct a three-dimensional correlation matrix including a time window, a frequency domain position, and a spatial angle of arrival, and perform spatio-temporal matching on the detection events within the synchronization window;

[0010] S4. Analyze the matching results to identify real drone targets and reduce the probability of multi-target false alarms;

[0011] The acquisition of the space angle of arrival is achieved through the following steps:

[0012] I. Use signal difference of at least two receiving antennas and utilize time difference positioning technology;

[0013] II. Use the phase difference measurement method to calculate the phase difference of the signals received by two antennas;

[0014] III. Combine the signal propagation model to deduce the incident angle of the signal and consider multi-path signals to improve the accuracy;

[0015] The analysis steps of the matching results include:

[0016] a. Apply machine learning algorithms to classify matching events and identify the identity characteristics of drones;

[0017] b. Generate alarm information and promptly notify the monitoring personnel for subsequent processing. The alarm information includes drone type, location, flight status, and potential threat level.

[0018] Preferably, the time reference system includes a high-precision clock source capable of providing a unified time reference for multiple signal acquisition units and a time synchronization protocol for realizing time synchronization between acquisition devices to ensure the time consistency of each signal acquisition device.

[0019] Preferably, the implementation scheme of the time synchronization protocol includes Network Time Protocol (NTP), Precision Time Protocol (PTP), time synchronization algorithms, redundant synchronization mechanisms, and synchronization status monitoring.

[0020] Preferably, the Network Time Protocol includes:

[0021] Deploy NTP server: Deploy a high-precision NTP server in the monitoring system as the time reference source to provide unified time information;

[0022] Time stamp synchronization: Each signal acquisition device regularly requests time stamps from the NTP server through an NTP client to ensure the time consistency of each device, with an accuracy reaching the microsecond level;

[0023] The Precision Time Protocol includes:

[0024] PTP network architecture: Establish a PTP network between devices and utilize the IEEE 1588 standard to achieve high-precision time synchronization;

[0025] Master-slave clock configuration: Select one device as the master clock and the rest as slave clocks. The master clock broadcasts time information regularly, and the slave clocks correct their times after receiving it to maintain synchronization at the microsecond level;

[0026] The time synchronization algorithm includes:

[0027] Time deviation calculation: Calculate the request response time at each timestamp request, adjust the deviation, and ensure synchronization accuracy;

[0028] Delay compensation: Consider network delay and perform delay compensation on synchronization information to further improve the accuracy and reliability of time synchronization;

[0029] The redundant synchronization mechanism includes:

[0030] Multi-source time synchronization: Support synchronization from multiple time sources, such as GPS time signals, atomic clocks, etc., to enhance the robustness of the system;

[0031] Fault detection and switching: Real-time monitor the time synchronization status. When the primary time source fails, automatically switch to the backup time source to ensure the continuous and stable operation of the system;

[0032] The synchronization status monitoring includes:

[0033] Status report: Regularly generate time synchronization status reports, monitor the synchronization accuracy and status of each device, and ensure that the system operates under optimal conditions;

[0034] Alarm mechanism: When it is detected that the time synchronization deviation exceeds the set range, immediately trigger an alarm to remind the maintenance personnel to conduct inspections and adjustments.

[0035] Preferably, the acquisition step of the IQ data stream includes real-time monitoring of the UAV signal and preliminary processing of the data to extract effective feature information.

[0036] Preferably, the construction of the three-dimensional correlation matrix further includes:

[0037] Normalize the signal strength within the time window to ensure comparability between different signal sources;

[0038] Perform clustering analysis in the frequency domain position to identify signal features with similar frequencies and enhance the accuracy of signal recognition;

[0039] Measure the angle of arrival in space multiple times and take the average value to improve the accuracy.

[0040] Preferably, the setting of the time window is dynamically adjusted according to the actual application to adapt to the signal characteristics in different environments. The dynamic adjustment includes:

[0041] Automatically adjust the width of the time window according to the environmental noise level to optimize signal detection;

[0042] Automatically optimize the time window settings according to the signal characteristics of different types of drones to improve detection accuracy.

[0043] Preferably, the dynamic adjustment algorithm of the time window is as follows:

[0044] Input parameters:

[0045] Environmental noise level (N): The background noise level measured by the receiver, in decibels (dB);

[0046] Signal strength (S): The received strength of the drone signal, in decibel-milliwatts (dBm);

[0047] Signal stability indicator (SSI): A stability index calculated by comparing historical signal strengths. The smaller the value, the smaller the signal fluctuation;

[0048] Initial time window (T_initial): The initial time window set by the system, in milliseconds (ms);

[0049] Steps of the dynamic adjustment algorithm:

[0050] 1. Monitor environmental noise and signal strength

[0051] Continuously monitor the current environmental noise level (N) and the drone signal strength (S);

[0052] 2. Calculate the signal-to-noise ratio (SNR)

[0053] Use the formula to calculate the signal-to-noise ratio: ;

[0054] 3. Set the dynamic adjustment threshold

[0055] Set the signal-to-noise ratio threshold (SNR_threshold), such as 10 dB. When the SNR is below this threshold, it means that the background noise is strong and may affect signal detection;

[0056] 4. Adjust the time window

[0057] a) When SNR ≥ SNR_threshold:

[0058] Increase the time window to improve the probability of signal capture. The adjustment formula is:

[0059] ;

[0060] where k is the adjustment coefficient, which can be set to 0.1;

[0061] b) When SNR < SNR_threshold:

[0062] Reduce the time window to reduce the impact of background noise on signal analysis. The adjustment formula is:

[0063] ;

[0064] 5. Introduce the Signal Stability Indicator (SSI)

[0065] Further adjust the time window according to the value of the Signal Stability Indicator (SSI):

[0066] If SSI < 5dB (indicating that the signal is very stable), increase the time window;

[0067] If SSI > 15dB (indicating that the signal fluctuates greatly), reduce the time window;

[0068] Perform the final adjustment in combination with the current time window:

[0069] ;

[0070] where m is the stability adjustment coefficient, which can be set to 0.05;

[0071] 6. Output the adjusted time window

[0072] Update the adjusted time window (T_final) to the system for subsequent signal acquisition and processing.

[0073] Preferably, the machine learning algorithm includes but is not limited to support vector machines, decision trees, and neural networks to improve the classification accuracy and efficiency of matching events.

[0074] Preferably, the training process of the machine learning algorithm adopts an incremental learning method to adapt to new input data and environmental changes.

[0075] Preferably, the generation method of the alarm information includes real-time data sending, SMS notification, and email alert to ensure that the monitoring personnel can obtain information in a timely manner.

[0076] In summary, the present application includes at least one of the following beneficial technical effects:

[0077] 1. Adopt a spatio-temporal synchronization correlation mechanism to establish a unified time reference system, perform precise timestamp alignment on the IQ data streams collected by the two technologies, and then construct a three-dimensional correlation matrix: time window, frequency domain position, and spatial angle of arrival. Perform spatio-temporal matching on the detection events within the synchronization window, which can effectively and accurately analyze the UAV signals, improve the accuracy and reliability of UAV target detection, reduce the possibility of multi-target false alarms, and provide more accurate technical support for UAV monitoring;

[0078] 2. The use of a three-dimensional correlation matrix for spatio-temporal matching, combined with protocol reverse parsing and radio frequency fingerprint recognition technology, can effectively reduce the multi-target false alarm phenomenon caused by asynchronous detection or lack of feature correlation;

[0079] 3. By continuously monitoring the environmental noise level and signal strength and dynamically adjusting the time window settings, the system can adapt to different environmental conditions, improving the robustness and flexibility of signal detection;

[0080] 4. It can generate alarm information based on the matching results and notify the monitoring personnel in a variety of ways, such as text messages and emails, to ensure the timely response and handling of UAV activities;

[0081] 5. The application of machine learning algorithms and their incremental learning capabilities enable the system to quickly adapt to new input data and environmental changes, further improving the efficiency and accuracy of UAV target detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 is a flowchart of the UAV target detection method of this application;

[0083] Figure 2 is a flowchart of the acquisition step of the angle of arrival in space in step S3 of this application;

[0084] Figure 3 is a flowchart of the analysis step of the matching result in step S4 of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0085] The following is a further detailed description of this application in conjunction with the attached Figure 1 - attached Figure 3 , to further elaborate on this application.

[0086] This application provides a UAV target detection method based on a spatio-temporal synchronization correlation mechanism, including the following steps:

[0087] S1. Establish a unified time reference system to ensure precise timestamp alignment of the IQ data streams collected by the two technologies, with an accuracy reaching the microsecond level;

[0088] S2. Collect the UAV signal data obtained by protocol reverse parsing and radio frequency fingerprint recognition technology, including but not limited to the identity identification code, telemetry data, signal strength indication, and modulation method of the UAV;

[0089] S3. Construct a three-dimensional correlation matrix including a time window, a frequency domain position, and an angle of arrival in space, and perform spatio-temporal matching on the detection events within the synchronization window;

[0090] S4. Analyze the matching results to identify real drone targets and reduce the probability of multi-target false alarms;

[0091] The acquisition of the angle of arrival in space is achieved through the following steps:

[0092] I. Use signal difference of at least two receiving antennas and utilize time difference positioning technology;

[0093] II. Use the phase difference measurement method to calculate the phase difference of the signals received by two antennas;

[0094] III. Combine the signal propagation model to deduce the incident angle of the signal and consider multi-path signals to improve accuracy;

[0095] The analysis steps of the matching results include:

[0096] a. Apply machine learning algorithms to classify matching events and identify the identity characteristics of drones;

[0097] b. Generate alarm information and promptly notify the monitoring personnel for subsequent processing. The alarm information includes drone type, location, flight status, and potential threat level.

[0098] Among them, the time reference system includes a high-precision clock source that can provide a unified time reference for multiple signal acquisition units and a time synchronization protocol for realizing time synchronization between acquisition devices to ensure the time consistency of each signal acquisition device.

[0099] Among them, the implementation solutions of the time synchronization protocol include Network Time Protocol (NTP), Precision Time Protocol (PTP), time synchronization algorithms, redundant synchronization mechanisms, and synchronization status monitoring.

[0100] Among them, the Network Time Protocol includes:

[0101] Deploy NTP server: Deploy a high-precision NTP server in the monitoring system as the time reference source to provide unified time information;

[0102] Time stamp synchronization: Each signal acquisition device regularly requests time stamps from the NTP server through the NTP client to ensure the time consistency of each device, with an accuracy reaching the microsecond level;

[0103] The Precision Time Protocol includes:

[0104] PTP network architecture: Establish a PTP network between devices and utilize the IEEE 1588 standard to achieve high-precision time synchronization;

[0105] Master-slave clock configuration: Select one device as the master clock and the remaining devices as slave clocks. The master clock regularly broadcasts time information, and the slave clocks perform time correction after receiving it to maintain microsecond-level synchronization;

[0106] The time synchronization algorithm includes:

[0107] Time deviation calculation: At each timestamp request, calculate the request response time, adjust the deviation, and ensure the synchronization accuracy;

[0108] Delay compensation: Consider network delay, perform delay compensation on the synchronization information, and further improve the accuracy and reliability of time synchronization;

[0109] The redundant synchronization mechanism includes:

[0110] Multi-source time synchronization: Support synchronization from multiple time sources, such as GPS time signals, atomic clocks, etc., to enhance the robustness of the system;

[0111] Fault detection and switching: Real-time monitor the time synchronization status. When the primary time source fails, automatically switch to the backup time source to ensure the continuous and stable operation of the system;

[0112] The synchronization status monitoring includes:

[0113] Status report: Regularly generate time synchronization status reports, monitor the synchronization accuracy and status of each device, and ensure the system operates under optimal conditions;

[0114] Alarm mechanism: When it is detected that the time synchronization deviation exceeds the set range, immediately trigger an alarm to remind the maintenance personnel to conduct inspections and adjustments.

[0115] Among them, the acquisition steps of the IQ data stream include real-time monitoring of the drone signal and preliminary processing of the data to extract effective feature information.

[0116] Among them, the construction of the three-dimensional correlation matrix further includes:

[0117] Normalize the signal strength within the time window to ensure comparability between different signal sources;

[0118] Perform clustering analysis in the frequency domain position to identify signal features with similar frequencies and enhance the accuracy of signal recognition;

[0119] Measure the spatial angle of arrival multiple times and take the average value to improve the accuracy.

[0120] Among them, the setting of the time window is dynamically adjusted according to the actual application to adapt to the signal characteristics in different environments. The dynamic adjustment includes:

[0121] Automatically adjust the width of the time window according to the environmental noise level to optimize signal detection;

[0122] Automatically optimize the setting of the time window according to the signal characteristics of different types of drones to improve the detection accuracy.

[0123] Among them, the steps of the dynamic adjustment algorithm for the time window are as follows:

[0124] Input parameters:

[0125] Environmental noise level (N): The background noise level measured by the receiver, in decibels (dB);

[0126] Signal strength (S): The received strength of the UAV signal, in decibel-milliwatts (dBm);

[0127] Signal stability indicator (SSI): A stability index calculated by comparing historical signal strengths, where a smaller value indicates less signal fluctuation;

[0128] Initial time window (T_initial): The initial time window set by the system, in milliseconds (ms);

[0129] Steps of the dynamic adjustment algorithm:

[0130] 1. Monitor the environmental noise and signal strength

[0131] Continuously monitor the current environmental noise level (N) and the UAV signal strength (S);

[0132] 2. Calculate the signal-to-noise ratio (SNR)

[0133] Use the formula to calculate the signal-to-noise ratio: ;

[0134] 3. Set the dynamic adjustment threshold

[0135] Set the signal-to-noise ratio threshold (SNR_threshold), such as 10 dB. When the SNR is lower than this threshold, it means that the background noise is strong and may affect signal detection;

[0136] 4. Adjust the time window

[0137] a) When SNR ≥ SNR_threshold:

[0138] Increase the time window to improve the probability of signal capture. The adjustment formula is:

[0139] ;

[0140] Among them, k is the adjustment coefficient, which can be set to 0.1;

[0141] b) When SNR < SNR_threshold:

[0142] Reduce the time window to reduce the impact of background noise on signal analysis. The adjustment formula is:

[0143] ;

[0144] 5. Introduce Signal Stability Indicator (SSI)

[0145] Further adjust the time window according to the value of the Signal Stability Indicator (SSI):

[0146] If SSI < 5 dB (indicating that the signal is very stable), increase the time window;

[0147] If SSI > 15 dB (indicating that the signal fluctuates greatly), decrease the time window;

[0148] Perform the final adjustment in combination with the current time window:

[0149] ;

[0150] where m is the stability adjustment coefficient, which can be set to 0.05;

[0151] 6. Output the adjusted time window

[0152] Update the adjusted time window (T_final) to the system for subsequent signal acquisition and processing.

[0153] Among them, the machine learning algorithms include but are not limited to support vector machines, decision trees, and neural networks to improve the classification accuracy and efficiency of matching events;

[0154] The support vector machine takes the extracted signal features, such as signal strength, center frequency, and modulation method, as inputs. The SVM can effectively distinguish different types of drones. In the training stage, the model learns the features of the sample data, finds the optimal separating surface, and optimizes the classification accuracy;

[0155] The decision tree classifies based on the extracted features such as signal strength and frequency pattern, generates simple and easy-to-understand classification rules, is suitable for quick decision-making in real-time monitoring, and can quickly provide classification results;

[0156] The neural network uses a deep learning model to be able to handle more complex pattern recognition tasks, especially in the extraction and classification of signal features. It automatically extracts signal features through a multi-layer network structure for more efficient classification.

[0157] Among them, the training process of the machine learning algorithm adopts the incremental learning method, which allows the model to be updated when new data is received without having to retrain from scratch to adapt to new input data and environmental changes. When the system detects new types of drones or signal features, incremental learning can enable the model to quickly adapt, reducing the training time and computing resources.

[0158] Among them, the implementation steps of incremental learning are as follows:

[0159] Data collection and preprocessing

[0160] Real-time data acquisition: The system continuously monitors the UAV signals and collects new data in real time, including information such as signal strength, frequency characteristics, modulation mode, etc.;

[0161] Data annotation: The newly collected data needs to be annotated for incremental learning. The annotation process can be carried out through expert review or automated tools;

[0162] Feature extraction: Extract features from the newly collected signal data to generate feature vectors for subsequent model updates;

[0163] Model update mechanism

[0164] Incremental training: After the new data is annotated and feature-extracted, it is directly added to the existing training dataset; Select appropriate machine learning algorithms such as SVM, decision tree or neural network for incremental training; At this time, the model will be updated based on the new data without having to train from scratch;

[0165] Training parameter adjustment: During the incremental learning process, adjust the training parameters of the model, such as learning rate and regularization parameter, to avoid overfitting and maintain the generalization ability of the model;

[0166] Model verification: After incremental training, use the reserved test set to verify the updated model, evaluate the performance changes of the model, and ensure the accuracy and stability of the model on the new dataset;

[0167] Dynamic update strategy

[0168] Adaptive update frequency: Dynamically adjust the model update frequency according to the frequency of UAV activities and the speed of new data appearance. For example, increase the update frequency when UAV activities are frequent; decrease the update frequency when UAV activities are less;

[0169] Error feedback mechanism: Compare the classification results of the monitoring system with the actual situation. If it is found that the model produces misclassifications in some cases, these error instances can be used for targeted incremental learning to further optimize the model;

[0170] Performance monitoring and optimization

[0171] Real-time monitoring: During the operation of the system, continuously monitor the performance metrics of the model, such as accuracy, recall rate and F1-score, to ensure the adaptability of the model in different environments;

[0172] Regular evaluation: Conduct a comprehensive model evaluation regularly, and retrain in combination with new data to ensure that the model can adapt to long-term environmental changes and the evolution of UAV signal characteristics;

[0173] User feedback integration: Collect feedback from monitoring personnel and use this feedback information to adjust and optimize the model to improve the actual application effect of the system.

[0174] Among them, the generation methods of alarm information include real-time data sending, SMS notification, and email alert to ensure that monitoring personnel can obtain information in a timely manner.

[0175] The embodiments of this application are as follows:

[0176] Embodiment 1: Drone target detection in urban environment

[0177] Background: In the urban environment, the flight activities of drones are becoming increasingly frequent, which may pose threats to public safety and privacy. This embodiment shows how to use the method of this application to detect drone targets in urban areas.

[0178] Steps:

[0179] 1. Establish a time reference system:

[0180] Deploy a high-precision NTP server and ensure that all signal acquisition devices, such as receiving antennas and signal processing units, regularly request timestamps through NTP clients to maintain time synchronization at the microsecond level;

[0181] 2. Signal data acquisition:

[0182] Use protocol reverse parsing and radio frequency fingerprint recognition technologies to collect signal data from drones flying in the air, including identity identification codes, telemetry data, and signal strength indicators;

[0183] 3. Construct a three-dimensional correlation matrix:

[0184] Set the time window to 100 ms and construct a three-dimensional correlation matrix containing the time window, frequency domain position, and spatial angle of arrival; use at least two receiving antennas for signal difference to obtain the spatial angle of arrival of the drone;

[0185] 4. Matching result analysis:

[0186] Apply machine learning algorithms and use support vector machines to classify signal data to identify the identity characteristics of drones;

[0187] Generate alarm information, including drone type, location, flight status, and potential threat level, and notify monitoring personnel via SMS and email;

[0188] 5. Dynamically adjust the time window:

[0189] Dynamically adjust according to the ambient noise level of real-time monitoring of urban traffic noise. If the noise level is high, shorten the time window to 50 ms to optimize signal detection;

[0190] Result: Through the above steps, the flight trajectories of multiple drones were successfully identified, and alarm information was sent in a timely manner to ensure the safety of the urban area.

[0191] Example 2: Drone detection in farmland monitoring

[0192] Background: In the agricultural field, drones are commonly used for crop monitoring and spraying pesticides. This example demonstrates how to effectively monitor drones in a farmland environment using the method of this application.

[0193] Steps:

[0194] 1. Establish a time reference system:

[0195] Configure a Precision Time Protocol (PTP) network, set one device as the master clock, and regularly broadcast time information to ensure synchronization of each signal acquisition device;

[0196] 2. Signal data acquisition:

[0197] Through the telemetry system of the drone, real-time acquisition of drone signal data, including identity identification codes and signal strength;

[0198] 3. Construct a three-dimensional correlation matrix:

[0199] Set the time window to 150 ms, construct a three-dimensional correlation matrix, including the time window, frequency domain position, and spatial angle of arrival; use the phase difference measurement method to calculate the signal phase difference and deduce the incident angle of the signal;

[0200] 4. Matching result analysis:

[0201] Use the decision tree algorithm to classify the identity characteristics of the drone and generate alarm information to notify the farmland management personnel in a timely manner;

[0202] 5. Dynamically adjust the time window:

[0203] Automatically adjust the time window according to weather changes and ambient noise, such as wind noise and bird calls. If the wind speed is high, shorten the time window to 75 ms to ensure accurate signal capture;

[0204] Result: The drones flying over the farmland were successfully identified, realizing real-time monitoring and management, and improving agricultural production efficiency.

[0205] Example 3: Drone monitoring in public activities

[0206] Background: In large-scale public events, such as music festivals and sports events, it is crucial to monitor the flight of drones. This embodiment demonstrates how to apply the method of this application for drone target detection in public events;

[0207] Steps:

[0208] 1. Establish a time reference system:

[0209] Deploy multiple high-precision clock sources at the event site to ensure that all signal acquisition units can achieve microsecond-level time synchronization;

[0210] 2. Signal data acquisition:

[0211] Utilize protocol reverse parsing and radio frequency fingerprint recognition technologies to collect drone signal data in real time, and record the identity identification code, telemetry data, and signal strength;

[0212] 3. Construct a three-dimensional correlation matrix:

[0213] Set the time window to 200 ms, and construct a three-dimensional correlation matrix including the time window, frequency domain position, and spatial angle of arrival; obtain the spatial angle of arrival of the drone by using multiple receiving antennas for signal difference;

[0214] 4. Matching result analysis:

[0215] Apply a neural network algorithm to classify the identity characteristics of the drone and generate an alarm message, the content of which includes the type of the drone, flight status, and potential threat level;

[0216] 5. Dynamically adjust the time window:

[0217] Dynamically adjust the time window according to the environmental noise at the event site (such as the noise of the audience). When the noise level is high, shorten the time window to 100 ms to optimize signal detection;

[0218] Result: By implementing the above steps, multiple drones were successfully detected and identified in public events, and alarms were issued in a timely manner to ensure the safe progress of the event.

[0219] The embodiments of this specific implementation manner are all preferred embodiments of this application, and do not limit the protection scope of this application accordingly. The same components are represented by the same reference numerals. Therefore, all equivalent changes made according to the structure, shape, and principle of this application should be covered within the protection scope of this application.

Claims

1. A method for detecting unmanned aerial vehicle targets based on a spatiotemporal synchronous association mechanism, characterized in that: The following steps are involved: S1. Establish a unified time reference system to ensure precise time stamp alignment of the IQ data streams collected by the two technologies with microsecond accuracy; S2. Collect drone signal data obtained through protocol reverse analysis and radio frequency fingerprinting technology, including but not limited to the drone’s identification code, telemetry data, signal strength indication, and modulation mode; S3, constructing a three-dimensional correlation matrix including time window, frequency domain position and spatial arrival angle, and performing spatiotemporal matching of detection events within the synchronization window; S4. Analyze the matching results to identify the real UAV target and reduce the probability of multiple target false alarms; The acquisition of the spatial arrival angle is achieved by the following steps: I. Using the signal difference of at least two receiving antennas and using the time difference positioning technology; II. Use the phase difference measurement method to calculate the phase difference of the signals received by the two antennas; III. Combined with the signal propagation model, the incident angle of the signal is derived, and multipath signals are considered to improve accuracy; The matching result analysis step includes: a. Apply machine learning algorithms to classify matching events and identify the identity characteristics of drones; b. Generate alarm information and promptly notify monitoring personnel for follow-up processing. The alarm information includes the type, location, flight status and potential threat level of the drone.

2. The method for detecting unmanned aerial vehicle targets based on a spatiotemporal synchronization association mechanism according to claim 1, characterized in that: The time reference system includes a high-precision clock source capable of providing a unified time reference for multiple signal acquisition units and a time synchronization protocol for achieving time synchronization between acquisition devices, so as to ensure the time consistency of each signal acquisition device.

3. The method for detecting unmanned aerial vehicle targets based on a spatiotemporal synchronization association mechanism according to claim 2, characterized in that: The implementation scheme of the time synchronization protocol includes a network time protocol, a precision time protocol, a time synchronization algorithm, a redundant synchronization mechanism and synchronization status monitoring.

4. The method for detecting unmanned aerial vehicle targets based on a spatiotemporal synchronization association mechanism according to claim 3 is characterized in that: The network time protocol includes: Deploy NTP server: Deploy a high-precision NTP server in the monitoring system as a time reference source to provide unified time information; Timestamp synchronization: Each signal acquisition device periodically requests a timestamp from the NTP server through the NTP client to ensure the time consistency of each device with microsecond accuracy; The precision time protocol includes: PTP network architecture: Establish a PTP network between devices and use the IEEE 1588 standard to achieve high-precision time synchronization; Master-slave clock configuration: select one device as the master clock and the rest as slave clocks. The master clock broadcasts time information regularly, and the slave clocks receive the time and correct it to maintain microsecond synchronization. The time synchronization algorithm includes: Time deviation calculation: For each timestamp request, the request response time is calculated and the deviation is adjusted to ensure synchronization accuracy; Delay compensation: Considering network delay, delay compensation is performed on synchronization information to further improve the accuracy and reliability of time synchronization; The redundant synchronization mechanism includes: Multi-source time synchronization: supports synchronization from multiple time sources to improve the robustness of the system; Fault detection and switching: Real-time monitoring of time synchronization status, automatic switching to the backup time source when the main time source fails, ensuring continuous and stable operation of the system; The synchronization status monitoring includes: Status report: Generate time synchronization status reports regularly to monitor the synchronization accuracy and status of each device to ensure that the system runs under optimal conditions; Alarm mechanism: When it is detected that the time synchronization deviation exceeds the set range, an alarm is triggered immediately to remind maintenance personnel to check and adjust.

5. The method for detecting unmanned aerial vehicle targets based on a spatiotemporal synchronization association mechanism according to claim 1, characterized in that: The IQ data stream collection step includes real-time monitoring of drone signals and preliminary processing of the data to extract effective feature information.

6. The method for detecting unmanned aerial vehicle targets based on a spatiotemporal synchronization association mechanism according to claim 1, characterized in that: The construction of the three-dimensional incidence matrix further includes: The signal intensity within the time window is normalized to ensure comparability between different signal sources; Perform cluster analysis in the frequency domain to identify signal features with similar frequencies and enhance the accuracy of signal recognition; Multiple measurements of the spatial angle of arrival are taken and averaged to improve accuracy.

7. The method for detecting unmanned aerial vehicle targets based on a spatiotemporal synchronization association mechanism according to claim 1, characterized in that: The setting of the time window is dynamically adjusted according to actual applications to adapt to signal characteristics in different environments. The dynamic adjustment includes: Automatically adjust the width of the time window according to the ambient noise level to optimize signal detection; The time window setting is automatically optimized according to the signal characteristics of different types of drones to improve detection accuracy.

8. The method for detecting unmanned aerial vehicle targets based on a spatiotemporal synchronization association mechanism according to claim 1, characterized in that: The machine learning algorithms include but are not limited to support vector machines, decision trees, and neural networks to improve the classification accuracy and efficiency of matching events.

9. The method for detecting unmanned aerial vehicle targets based on a spatiotemporal synchronization association mechanism according to claim 1, characterized in that: The training process of the machine learning algorithm adopts an incremental learning method to adapt to new input data and environmental changes.

10. The method for detecting unmanned aerial vehicle targets based on a spatiotemporal synchronization association mechanism according to claim 1, characterized in that: The alarm information is generated in a manner including real-time data transmission, SMS notification and email alert, ensuring that monitoring personnel obtain information in a timely manner.

Images