Unmanned aerial vehicle detection countering method, system, device and medium

Abstract

The invention discloses an unmanned aerial vehicle detection countering method, system and equipment and a medium, and the method comprises the steps: obtaining a task instruction and sensing signals of a plurality of sensing devices in a monitoring region in real time, carrying out the preprocessing of the task instruction and the sensing signals, and obtaining the standardized modal data; performing feature extraction and high-level semantic extraction on the task instruction and the standardized modal data corresponding to each sensing signal in parallel to obtain a task context parameter and a high-dimensional feature vector of each sensing signal; fusing the high-dimensional feature vectors based on the task context parameters by using a space-time alignment multi-modal fusion algorithm to obtain multi-modal fusion features; using a neural network model to predict and generate a flight trajectory of the unmanned aerial vehicle in a future time period according to the historical trajectory data and the multi-modal fusion features; and performing threat level judgment and behavior anomaly detection based on the task context parameters and the flight trajectory, generating alarm information, and executing a security countering strategy generated based on the alarm information.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent sensing and processing, and more particularly to a UAV detection countermeasure method, system, device and medium. BACKGROUND

[0002] With the rapid development of UAV technology, UAVs are increasingly widely used in civilian, commercial and military fields. However, since UAVs can fly at low altitudes and are small in size, flexible and highly maneuverable, if misused (such as illegal intrusion, privacy invasion or security threats, etc.), it will pose a certain security risk. Therefore, the counter-UAV system (C-UAS) has become a research hotspot in recent years, aiming to realize the detection, tracking, identification and interference of UAVs. However, the existing counter-UAV system mainly relies on a single sensor (such as radar or radio frequency detection) for detection, and has limited processing capability, and there are certain limitations in the detection capability of small UAVs, which is prone to false negatives or false positives, and it is difficult to cope with high dynamic UAV threats in complex environments. SUMMARY

[0003] The present application provides a UAV detection countermeasure method, system, device and medium to solve the technical problem that the existing counter-UAV system is prone to false negatives or false positives, and is difficult to cope with high dynamic UAV threats in complex environments.

[0004] In a first aspect, a UAV detection countermeasure method is provided, comprising: real-time acquisition of task instructions and sensing signals of a plurality of different sensing devices in a monitoring area, and preprocessing of the task instructions and the sensing signals to obtain standardized modal data corresponding to the task instructions and each sensing signal; parallel feature extraction and high-level semantic extraction of the standardized modal data corresponding to the task instructions and each sensing signal to obtain task context parameters and high-dimensional feature vectors corresponding to each sensing signal; using a spatio-temporal alignment multi-modal fusion algorithm, dynamically adjusting a confidence score strategy based on the task context parameters, and fusing the high-dimensional feature vectors corresponding to each sensing signal to obtain multi-modal fusion features containing UAV states; using a pre-trained neural network model to predict and generate a flight trajectory of a UAV within a preset time period in the future according to historical trajectory data of the UAV and the multi-modal fusion features; based on the task context parameters and the flight trajectory, threat level determination and behavior anomaly detection are performed to generate an alarm information, and a security countermeasure strategy is generated based on the alarm information and executed.

[0005] In a second aspect, a UAV detection countermeasure system is provided, comprising: a data preprocessing unit configured to acquire task instructions and sensing signals of a plurality of different sensing devices in a monitoring area in real time, and preprocess the task instructions and the sensing signals to obtain standardized modal data corresponding to the task instructions and each sensing signal; a feature extraction unit configured to perform feature extraction and high-level semantic extraction on the standardized modal data corresponding to the task instructions and each sensing signal in parallel to obtain a task context parameter and a high-dimensional feature vector corresponding to each sensing signal; a feature fusion unit configured to perform fusion on the high-dimensional feature vector corresponding to each sensing signal based on the task context parameter by using a spatiotemporal alignment multi-modal fusion algorithm to dynamically adjust a confidence score strategy, and obtain multi-modal fusion features containing a state of the unmanned aerial vehicle; a trajectory prediction unit configured to predict a flight trajectory of the unmanned aerial vehicle in a future preset time period according to historical trajectory data of the unmanned aerial vehicle and the multi-modal fusion features by using a pre-trained neural network model; a detection countermeasure unit configured to perform threat level determination and behavior anomaly detection based on the task context parameter and the flight trajectory, generate an alarm information, and generate a security countermeasure strategy based on the alarm information and perform the security countermeasure strategy.

[0006] The application further provides an unmanned aerial vehicle detection countermeasure system, comprising a server, a plurality of sensing devices of different types and arranged in a monitoring area, a plurality of interference devices of different types and arranged in the monitoring area, and a terminal. The plurality of sensing devices are configured to collect information of an unmanned aerial vehicle invading the monitoring area in real time from different dimensions to generate sensing signals. The server is configured to acquire task instructions and sensing signals of a plurality of sensing devices in a monitoring area in real time, preprocess the task instructions and the sensing signals to obtain standardized modal data corresponding to the task instructions and each sensing signal, perform feature extraction and high-level semantic extraction on the standardized modal data corresponding to the task instructions and each sensing signal in parallel to obtain a task context parameter and a high-dimensional feature vector corresponding to each sensing signal, perform fusion on the high-dimensional feature vector corresponding to each sensing signal based on the task context parameter by using a spatiotemporal alignment multi-modal fusion algorithm to dynamically adjust a confidence score strategy, and obtain multi-modal fusion features containing a state of the unmanned aerial vehicle, and predict a flight trajectory of the unmanned aerial vehicle in a future preset time period according to historical trajectory data of the unmanned aerial vehicle and the multi-modal fusion features by using a pre-trained neural network model, perform threat level determination and behavior anomaly detection based on the task context parameter and the flight trajectory, generate an alarm information, and generate a security countermeasure strategy based on the alarm information. The terminal is configured to send the task instructions and display the alarm information. A number of interference devices are connected with the server through an MQTT or HTTP interface, and are used to execute the security countermeasure strategy pushed by the server.

[0007] In a third aspect, a computer device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the UAV detection countermeasure method when executing the computer program.

[0008] In a fourth aspect, a computer readable storage medium is provided, which stores a computer program, and the computer program implements the steps of the UAV detection countermeasure method when executed by a processor.

[0009] Compared with the prior art, the UAV detection countermeasure method can obtain standardized modal data by uniformly preprocessing the real-time acquired task instructions and the sensing signals of the plurality of different sensing devices in the monitoring area, and can obtain the task context parameters and the high-dimensional feature vectors corresponding to each sensing signal by performing feature extraction and high-level semantic extraction on the standardized modal data of different modalities in parallel, and can obtain the multi-modal fusion features containing the UAV state by using the spatiotemporal alignment multi-modal fusion algorithm, dynamically adjusting the confidence score strategy based on the task context parameters, and fusing the high-dimensional feature vectors corresponding to each sensing signal. The flight trajectory of the UAV in the future preset time period can be predicted by using the pre-trained neural network model according to the historical trajectory data of the UAV and the multi-modal fusion features. Finally, the threat level determination and behavior anomaly detection are performed based on the task context parameters and the flight trajectory, the alarm information is generated, the security countermeasure strategy is generated based on the alarm information, and the security countermeasure strategy is executed. It can be seen that the UAV detection countermeasure method can detect the invading UAV according to the multi-source data, can improve the adaptability of complex environment detection, can avoid the situation of missing report or false report, and can improve the data transmission efficiency by performing parallel feature extraction of multi-source data. In the detection process, the task instructions are also used as the control parameters of the UAV detection to guide the feature fusion and decision of different sensing signals, form a dynamic, intelligent and taskable perception-decision system, and can improve the accuracy and robustness of UAV detection in complex environments. BRIEF DESCRIPTION OF DRAWINGS

[0010] Figure 1 A flowchart of a UAV detection countermeasure method provided for the first embodiment of the present application is shown in the figure; Figure 2 A flowchart of a specific implementation of step S130 in the first embodiment is shown in the figure; Figure 1 A flowchart of a specific implementation of step S140 in the first embodiment is shown in the figure; Figure 3 Figure 1 A flowchart of a specific implementation of step S140 in the first embodiment is shown in the figure; Figure 4 A flowchart of a specific implementation of step S140 in the first embodiment is shown in the figure;​Figure 1 A schematic diagram of a specific implementation of step S160; Figure 5 This is a flowchart illustrating the drone detection and countermeasure method provided in the second embodiment of the present invention. Figure 6 A flowchart illustrating the UAV detection and countermeasure method provided in the third embodiment of the present invention; Figure 7 This is a schematic diagram of the structure of the UAV detection and countermeasure system provided in the first embodiment of the present invention; Figure 8 This is a schematic diagram of the structure of the UAV detection and countermeasure system provided in the second embodiment of the present invention; Figure 9 yes Figure 8 The diagram shows the internal architecture of the server in the drone detection and countermeasure system. Figure 10 yes Figure 8 The diagram shows the interface framework of the database in the server of the UAV detection and countermeasure system. Figure 11 This is a schematic diagram of the structure of a computer device in an embodiment provided by the present invention. Detailed Implementation

[0011] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] It should also be understood that the term “and / or” as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0013] Please see Figure 1 , Figure 1 This is a schematic flowchart illustrating the drone detection and countermeasure method provided in the first embodiment of the present invention. In the embodiment shown in the figures, the drone detection and countermeasure method includes the following steps S110-160: S110. Real-time acquisition of task instructions and sensing signals from multiple different sensing devices in the monitoring area, and preprocessing of the task instructions and sensing signals to obtain standardized modal data corresponding to the task instructions and each sensing signal.

[0014] In this invention, the sensing devices can be radar, radio transmitters and receivers, detection photoelectric devices, cameras, position sensors, and climate sensors, etc. The sensing signals can be electromagnetic signals (radar signals, TDOA signals, UAV communication signals), perception signals (images, videos, and sounds acquired by the camera), environmental signals (geographic coordinates collected by the position sensor, wind speed and rainfall collected by the climate sensor), etc. The task command can be a user-input command, which can be voice data or text data. If the input is voice data, it can be converted to text before processing. In this embodiment, the task command can be used as the control parameters for UAV detection, and can include the core target and / or rules of this detection and countermeasure mission. For example, the task command can be "No-fly zone: 5 kilometers around the airport runway, altitude 0-1000 meters. Any UAV that intrudes into this area is the highest threat (level 5), and countermeasures should be taken." The UAV detection and countermeasure method of this invention can dynamically and rationally determine the threat level of the intruding UAV based on the core target and rules of the task command.

[0015] Specifically, in this step, the acquired task instructions and various sensor signals are first cleaned, converted, or standardized. For example, radar signals can be subjected to Fast Fourier Transform (FFT) to filter out noise, extract target frequency and amplitude, and convert them into structured data; video streams are downsampled (from 30 FPS to 15 FPS) to reduce bandwidth requirements, and frame extraction is performed to generate metadata for each frame; environmental signals are formatted, wind speed is converted from m / s to km / h (5 m / s→18 km / h), the validity of geographic coordinates is verified, and structured data is generated; for task instructions, the task instruction JSON can be parsed first, key fields can be extracted, the integrity of the instructions can be verified, and structured data can be generated; then the above-processed data is converted into a unified format (JSON or Protobuf) to obtain their respective standardized modal data for subsequent use.

[0016] S120. Using a distributed stream processing framework, the task instructions and standardized modal data corresponding to multiple sensing devices are distributed to different data topics according to preset topic rules.

[0017] In this step, standardized modal data corresponding to electromagnetic, image, audio, and text can be pushed to different data topics (such as predefined electromagnetic topics, sensing topics, situation topics, etc.) in the distributed stream processing framework to achieve asynchronous message transmission. For example, standardized modal data of radar signals and standardized modal data of video can be pushed to the "electromagnetic topic" and "sensing topic" in the distributed stream processing framework, respectively, while standardized modal data corresponding to environmental signals and task instructions can be pushed to the "situation topic".

[0018] This invention introduces a distributed stream processing framework (such as Kafka), which can be combined with various databases (such as in-memory databases, time-series databases, vector databases, etc.) to support real-time and high-concurrency streaming data processing. The modules and models of each consumer object can directly pull data from Kafka topics, can be independently expanded and upgraded, and the data will be persistently stored in the topic for a period of time (e.g., 7 days). When retraining is needed, the data stream at any point in the past can be replayed, which is more conducive to the data transmission and processing of dynamic and rapidly changing drone scenarios.

[0019] S130. Perform feature extraction and high-level semantic extraction on the standardized modal data corresponding to the task instructions issued by the data topic in the distributed stream processing framework and each sensor signal, respectively, to obtain the task context parameters and the high-dimensional feature vectors corresponding to each sensor signal.

[0020] In this step, the standardized modal data issued by each data topic can be processed in parallel. Specifically, feature extraction and high-level semantic extraction are performed on the standardized modal data of the task instructions to obtain the task context parameters. At the same time, feature extraction and high-level semantic extraction are performed on the standardized modal data corresponding to each sensor signal in parallel to obtain the high-dimensional feature vectors corresponding to each sensor signal.

[0021] like Figure 2 As shown, step S130 specifically includes the following steps S131-S132: S131. Perform feature extraction on the standardized modal data corresponding to the task instructions and each sensing signal in parallel to obtain multiple corresponding modal features.

[0022] In this step, feature extraction is performed on each standardized modal data. Specifically, electromagnetic signals can be analyzed for frequency structure and signal strength characteristics using Fast Fourier Transform (FFT) and wavelet transform to generate a 128-dimensional spectral vector (i.e., the extracted modal features); image and video data in the sensing signals can be used for target detection (extracting the bounding box information of the UAV) and feature extraction using the YOLOv8 algorithm to obtain the modal features of the sensing signals; audio data can be converted into 40-dimensional spectral time-series features (i.e., the modal features of the audio data) using the MFCC (Mel-Frequency Cepstral Coefficients) algorithm; and text data (task instructions) can be processed by the LLaMA large language model to generate a 1024-dimensional semantic embedding vector (i.e., the modal features of the text data).

[0023] Preferably, the modal features of each data or signal obtained by the above feature extraction can be processed in a unified format and then written into the feature topic of Kafka as the input for subsequent fusion.

[0024] S132. Extract high-level semantic information from all the modal features to obtain high-dimensional feature vectors and task context parameters corresponding to each sensing signal.

[0025] In this step, high-level semantic information is extracted from the modal features of electromagnetic signals, images, audio, and text to generate unified features, thereby improving the ability to understand the behavior of drones.

[0026] For electromagnetic signals, their modal features (128-dimensional spectral vectors) are input into a one-dimensional convolutional neural network (1D-CNN) to extract 64-dimensional semantic features. An attention mechanism is then used, combined with the formula... Highlight key frequencies (such as remote control signal bands); among them, It is the i-th dimension of the spectrum vector. For attention parameters, As weight.

[0027] For image and video data in the sensing signal, the ResNet50 model is used to extract high-dimensional image semantic features from its modal features, with a dimension of approximately 512. These features are then compressed to 256 dimensions using a Transformer encoder. Attention is focused on key areas of the UAV (such as the rotor), reducing background interference and enhancing semantic context.

[0028] For audio data, its modal features (40-dimensional spectral temporal features) can be used to obtain the 32-dimensional voiceprint features of the drone through a temporal convolutional network, thereby improving the distinguishability.

[0029] For text data, the LLaMA large model can be used according to the formula. Feature words (such as "intercept", "DJI", "danger", etc.) are extracted and compressed to 128 dimensions to obtain the task context parameters after high-level semantic extraction; among them, For the original embedding, and These are parameters related to the task.

[0030] Preferably, in order to simultaneously measure the contribution of all modalities, this embodiment uses a consistent loss function to guide the training of the respective models in all modalities. This loss function is: ; in, For the mapping features of mode m, For the mapping characteristics of mode n, Let the modal distribution divergence be... For regularization weights.

[0031] S140. Using a spatiotemporally aligned multimodal fusion algorithm, the confidence scoring strategy is dynamically adjusted based on the task context parameters, and the high-dimensional feature vectors corresponding to each sensor signal are fused to obtain multimodal fusion features that include the UAV state.

[0032] Specifically, such as Figure 3 As shown, step S140 includes the following steps S141-S143: S141. Use the Kalman filter algorithm to perform spatiotemporal alignment processing on all high-dimensional feature vectors.

[0033] In this step, the high-dimensional feature vectors corresponding to each sensor signal after semantic extraction, such as the high-dimensional feature vectors corresponding to electromagnetic signals, sensing signals and environmental signals, can be processed for time consistency. The high-dimensional feature vectors of the electromagnetic signals, sensing signals and environmental signals after time consistency processing are then mapped to the same spatial coordinate system to complete spatiotemporal alignment, thereby ensuring the consistency of information in time and space.

[0034] S142. The Bayesian confidence estimation method is used to score the confidence of the high-dimensional feature vectors corresponding to each sensor signal after spatiotemporal alignment based on the task context parameters.

[0035] In this embodiment, task context parameters (such as "night" and "city") can be used as prior knowledge. The Bayesian confidence estimation method calculates the posterior probability, i.e., the confidence level of the sensing device in providing accurate data in the current environment, based on this prior knowledge and the likelihood probability of the sensor data. Specifically, using the formula... The confidence level of each modality data is quantified to dynamically adjust the fusion weights, thereby optimizing the target state vector (such as position, velocity, threat level, etc.); among which, The modal data (i.e., the high-dimensional feature vectors corresponding to each sensor signal) is represented by S, which represents sensor characteristics, such as radar accuracy, and E represents environmental factors, such as urban areas and farmland. Likelihood probabilities can be used to model this using a Gaussian mixture model: ; Understandably, due to the different characteristics of sensing devices, the detection accuracy varies in different mission scenarios. Therefore, the weights will be dynamically adjusted during fusion. For example, radio detection is good at detecting long-range signals (such as drone remote control signals) and is more suitable for detecting targets in the initial stage of detection through spectrum identification and communication analysis at night or in foggy weather. In this scenario and / or stage, the weight will increase. Radar provides high-precision distance and speed information and is more suitable for positioning and tracking in open areas. In this scenario and / or stage, the weight will increase. Audio is usually used to detect drone acoustic signatures, and video provides target classification. The weight of video is usually increased in close-range confirmation and complex environments (such as densely populated urban areas with tall buildings).

[0036] S143. Filter out high-dimensional feature vectors with confidence scores lower than a preset threshold from the high-dimensional feature vectors corresponding to each sensor signal, and use the extended Kalman filter algorithm to fuse the high-dimensional feature vectors retained after filtering according to the confidence scores to obtain multimodal fusion features containing UAV state.

[0037] In this step, filtering is performed using a preset threshold (e.g., 0.8) to remove noisy data. After filtering out high-dimensional feature vectors with confidence scores lower than the preset threshold from the high-dimensional feature vectors corresponding to each sensor signal, the high-dimensional feature vectors retained in each sensor signal can be fused using the extended Kalman filter algorithm. The drone status can include key status indicators such as the drone's real-time position, speed, heading angle, and threat level. The multimodal fusion feature is a comprehensive description of the drone's current status and can also be written into the Kafka trajectory topic for subsequent use.

[0038] S150. Using a pre-trained neural network model, predict and generate the flight trajectory of the UAV within a preset time period based on the UAV's historical trajectory data and the multimodal fusion features.

[0039] In this embodiment, the neural network model can be an LSTM (Long Short-Term Memory) model, which can call up historical trajectory data stored in the time series database and use the pre-trained LSTM model to predict the flight trajectory of the UAV within a preset time period based on the historical trajectory data and the multimodal fusion features.

[0040] S160. Based on the mission context parameters and flight trajectory, determine the threat level and detect abnormal behavior, generate alarm information, generate a security countermeasure strategy based on the alarm information, and execute it.

[0041] In this step, by combining multidimensional analysis (threat level and abnormal behavior detection) and a rapid response mechanism, alarm information can be quickly detected and generated to alert users and provide warnings. Security countermeasures can be generated based on the alarm information, and the interference devices in the monitoring area can be controlled to execute the security countermeasures generated based on the alarm information, ensuring high efficiency in complex scenarios.

[0042] like Figure 4 As shown, step S160, which involves determining the threat level and detecting abnormal behavior based on the mission context parameters and flight trajectory, and generating alarm information, specifically includes the following steps S161-S163: S161. Based on the task context parameters and flight trajectory, the threat level of the intruding drone is obtained using the Drools rule engine, thereby determining the threat level.

[0043] In this invention, the threat level can reflect the number, location, and flight speed of the intruding drones.

[0044] In this step, the threat level of the intruding drone's flight data is obtained using Drools rules based on preset rules or rules contained in the task context parameters, thereby determining the threat level. For example, if the drone's speed is greater than 50 meters per second and its flight altitude is less than 100 meters, it is determined to be a high-threat target.

[0045] Specifically, fuzzy logic, based on the Drools rule engine, can be used to combine real-time flight trajectory parameters (position, speed, heading angle, etc.) and historical behavior patterns, according to formulas... Generate a threat level; where x is the overall threat index (which can be a weighted sum of speed and altitude), and a and b are the threat level thresholds. Membership degree is used to quantify the degree of ambiguity between "low threat" and "high threat." When the threat level or anomaly score exceeds a threshold, for example... If the value is greater than 0.8, relevant alarm information will be pushed.

[0046] S162. The isolated forest algorithm is used to identify abnormal flight behavior on the flight trajectory to obtain abnormal flight behavior data. A pre-trained classifier is then used to identify the pattern intent of the abnormal flight behavior data to obtain the UAV's behavioral intent.

[0047] In this step, the behavioral patterns can include hovering (the drone repeatedly flies in a circular or elliptical trajectory within a specific area, exhibiting a periodic trajectory pattern), hovering (maintaining relative stillness at a specific location), cruising (the drone flies at a stable speed along a fixed direction or path, with a trajectory that is a uniform straight line or a smooth curve), slamming (the drone moves at high speed towards a specific target or direction, with high speed and strong directionality, such as approaching at high speed or moving away quickly), and maneuvering (performing complex trajectory changes, such as sharp turns, evasion, or formation adjustments, with irregular and highly dynamic trajectories). The hovering mode is usually used for continuous surveillance or standby and is related to reconnaissance intentions; the hovering mode is usually used for reconnaissance, surveillance, or waiting for instructions; while cruising is often used for long-distance movement or area coverage; slamming is usually related to intentions such as surprise attack or escape; and maneuvering is usually related to intentions such as evasion, detection, or coordinated action. During anomaly detection, if the drone is a non-cooperative drone and the drone's behavioral pattern intention is abnormal, such as hovering for reconnaissance or rapid escape, it is identified as abnormal behavior information and an alarm can be triggered.

[0048] S163. If the threat level reaches a preset level, and / or the drone's behavior is abnormal, generate an alarm message.

[0049] In this embodiment, the alarm information includes threat level and abnormal behavior information (i.e., the intention of drone behavior identified as abnormal).

[0050] Understandably, in this invention, the order of steps S161 for determining the threat level and step S162 for identifying abnormal flight behavior can be interchanged.

[0051] Furthermore, in some embodiments, after generating alarm information, the threat level and abnormal behavior information in the alarm information can be written into the corresponding topic of the distributed stream processing framework, such as the threat level topic and behavior analysis topic, respectively, as a reference for countermeasure control strategies. The threat level and abnormal behavior information in the distributed stream processing framework can also be pushed externally via the WebSocket protocol for display in the human-machine interface alarm pop-up to alert the user, and control commands containing security countermeasure strategies can be pushed to external jamming devices via MQTT or HTTP interfaces for countermeasures. This embodiment uses the WebSocket / MQTT protocol for data transmission, and the fast jamming mechanism based on the WebSocket / MQTT protocol can shorten the response time and effectively deal with the threat of high-speed moving drones.

[0052] As can be seen from the above scheme, the UAV detection and countermeasure method in this embodiment obtains standardized modal data by uniformly preprocessing the real-time acquired task commands and sensor signals from multiple different sensor devices in the monitoring area. A distributed stream processing framework is then used to distribute the standardized modal data to different data topics according to preset topic rules. Subsequently, the standardized modal data from different data topics are sent to downstream consumer objects for parallel feature extraction and high-level semantic extraction, resulting in task context parameters and high-dimensional feature vectors corresponding to each sensor signal. Furthermore, a spatiotemporally aligned multimodal fusion algorithm is used to dynamically adjust the confidence scoring strategy based on the task context parameters, fusing the high-dimensional feature vectors corresponding to each sensor signal to obtain multimodal fusion features containing the UAV's state. Finally, a pre-trained neural network model is used to predict future UAV movements based on the UAV's historical trajectory data and the multimodal fusion features. The flight trajectory of the drone within a preset time period is predicted; finally, threat level determination and abnormal behavior detection are performed based on task context parameters and flight trajectory, generating alarm information. Based on the alarm information, a security countermeasure strategy is generated and executed. It can be seen that the drone detection and countermeasure method in this embodiment detects intruding drones based on multi-source data, which can improve the adaptability of detection in complex environments and avoid missed or false alarms. Moreover, the feature extraction of multi-source data is carried out in parallel, which can improve data transmission efficiency. Furthermore, the introduction of a distributed stream processing framework greatly improves the efficiency of data acquisition and transmission, which is more conducive to the data transmission and processing of dynamic and rapidly changing drone scenarios. During the detection process, the task command is also used as the control parameter for drone detection to guide the feature fusion and decision-making of different sensor signals, forming a dynamic, intelligent and task-oriented perception-decision system, which can improve the accuracy and robustness of drone detection in complex environments.

[0053] It should be noted that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0054] Reference Figure 5 , Figure 5 This is a flowchart illustrating the UAV detection and countermeasure method provided in the second embodiment of the present invention. As shown in the figure, the method includes the following steps S210-S270: S210. Real-time acquisition of task instructions and sensing signals from multiple different sensing devices in the monitoring area, and preprocessing of the task instructions and sensing signals to obtain standardized modal data corresponding to the task instructions and each sensing signal.

[0055] This step is the same as or similar to step S110, and will not be repeated here.

[0056] S220. Using a distributed stream processing framework, the task instructions and standardized modal data corresponding to multiple sensing devices are distributed to different data topics according to preset topic rules.

[0057] This step is the same as or similar to step S120, and will not be described again here.

[0058] S230. Perform feature extraction and high-level semantic extraction on the standardized modal data corresponding to the task instructions issued by the data topic in the distributed stream processing framework and each sensor signal, respectively, to obtain the task context parameters and the high-dimensional feature vectors corresponding to each sensor signal.

[0059] This step is the same as or similar to step S130, and will not be repeated here.

[0060] S240. Using a spatiotemporally aligned multimodal fusion algorithm, the confidence scoring strategy is dynamically adjusted based on the task context parameters, and the high-dimensional feature vectors corresponding to each sensor signal are fused to obtain multimodal fusion features that include the UAV state.

[0061] This step is the same as or similar to step S140, and will not be described again here.

[0062] S250. Using a pre-trained neural network model, predict and generate the flight trajectory of the UAV within a preset time period based on the UAV's historical trajectory data and the multimodal fusion features.

[0063] This step is the same as or similar to step S150, and will not be repeated here.

[0064] S260. Based on each UAV as a node, the multimodal fusion features and predicted flight trajectories of the UAVs are recorded in a graph network structure to obtain a UAV relationship graph. The graph neural network is then used to analyze the UAV relationship graph to infer the cooperative relationships of multiple UAVs and generate a global situation map of the monitoring area.

[0065] In this embodiment, each UAV is considered a node, and the node attributes are multimodal fusion features and predicted flight trajectories. A relationship graph of multiple UAVs in the monitoring area is constructed. The connections and interactions between nodes in the UAV relationship graph can be analyzed by a graph neural network (GNN) to calculate the cooperation score or relationship strength between pairs of UAVs (e.g., through attention mechanism weights). High-scoring relationship edges mean that there may be cooperative behaviors such as formation, following, and communication between them. A three-dimensional spatial coordinate is established, and the position coordinates and cooperation relationships of each UAV are mapped to the three-dimensional spatial coordinate to generate a global situation map. The global situation map can be used to more intuitively understand whether multiple UAVs have cooperative reconnaissance (the behavior of multiple UAVs flying in a dispersed manner to cover a large area), formation attack (the behavior of multiple UAVs maintaining a tight array and moving at high speed toward the same target), or dispersed interference behavior. It is suitable for the detection and countermeasures of multiple targets in complex application scenarios (such as cluster cooperation, potential cooperative intrusion, etc.).

[0066] S270. Based on the mission context parameters and flight trajectory, determine the threat level and detect abnormal behavior. Generate alarm information based on the threat level, abnormal behavior information and global situation map. Generate and execute security countermeasures based on the alarm information.

[0067] In this step, alarm information is generated based on threat level, abnormal behavior information, and global situation map. It can not only detect and identify the intentions of individual drones, but also identify the relationships and swarm behavior of multiple drones. This provides decision-makers with comprehensive and clear situational awareness, making it easier to quickly and accurately formulate security countermeasures (such as interference, evidence collection, physical strikes, etc.).

[0068] Furthermore, in some embodiments, the UAV detection and countermeasure method may also include: adaptively adjusting the weights of each feature in the multimodal fusion features based on the scene corresponding to the historical trajectory data of the UAV. That is, when fusing the high-dimensional feature vectors corresponding to each sensor signal, in addition to dynamically adjusting the fusion of the high-dimensional feature vectors corresponding to each sensor signal according to the task context parameters, the feature weights can also be adaptively adjusted according to the scene characteristics of the historical data to enhance adaptability to complex environments. Specifically, the weights of each modality (radar, radio, audio and video) can also be adaptively adjusted according to the scene characteristics of the historical data (e.g., open areas or densely populated urban areas). For example, the weights of audio and video are increased in urban areas. The specific implementation formula is as follows: ; in, The weights of mode m, To score modal credibility, This is a scene relevance function (obtained based on rules or machine learning models).

[0069] As can be seen, in the above scheme, the UAV detection and countermeasure method of this embodiment detects intruding UAVs based on multi-source data, which can improve the adaptability of detection in complex environments. Moreover, the parallel extraction of multi-source data features can improve data transmission efficiency. Furthermore, the introduction of a distributed stream processing framework greatly improves the efficiency of data acquisition and transmission, which is more conducive to the data transmission and processing of dynamic and rapidly changing UAV scenarios. During the detection process, task instructions are also used as control parameters for UAV detection to guide the feature fusion and decision-making of different sensor signals, forming a dynamic, intelligent and task-oriented perception-decision system, which can improve the accuracy and robustness of UAV detection in complex environments. In addition, a global situation map is generated by establishing a UAV relationship graph to analyze collaborative relationships, identify the relationships and swarm behavior of multiple UAVs, and provide decision-makers with comprehensive and clear situational awareness, which facilitates the rapid and accurate formulation of security countermeasure strategies.

[0070] Figure 6 This is a flowchart illustrating the UAV detection and countermeasure method provided in the third embodiment of the present invention, as shown below. Figure 6 As shown, the UAV detection and countermeasure method provided in this embodiment adds the following steps S310-S320 after step S160 of the first embodiment or after step S270 of the second embodiment: S310. Based on the historical trajectory data of the UAV, the trajectory is modeled using the HMM model to obtain a reconstructed trajectory containing the UAV behavior pattern sequence.

[0071] In this step, an in-depth analysis of the UAV's historical trajectory data is performed using an HMM (Hidden Markov Model) to infer the UAV's behavioral patterns at different points in time (e.g., hovering, cruising, loitering, slamming, maneuvering, etc.). Then, based on the behavioral pattern sequence composed of behavioral patterns at different points in time (e.g., "hovering → cruising → loitering"), a reconstructed trajectory that better reflects the UAV's intentions is generated. For example, "loitering" behavior may be related to reconnaissance intentions, and the reconstructed trajectory can highlight its periodic characteristics.

[0072] Specifically, in this embodiment, the state transition probability and observation probability of the HMM are defined as follows: ; in, Let t be the hidden state, i.e., the behavioral pattern. The state transition probability matrix is... For the probability of observation, For observation data (e.g., position, velocity). Understandably, the HMM model used in this step has been pre-trained using a large amount of historical trajectory data to learn the state transition matrix A and observation probabilities B; for example, the "cruising" state may correspond to observation data of a uniform straight trajectory, while the "maneuvering" state corresponds to highly dynamic trajectory changes.

[0073] S320. Based on the historical data corresponding to the historical trajectory data of the UAV and the reconstructed trajectory, the pre-trained neural network model is incrementally trained using an incremental learning mechanism.

[0074] In this step, the reconstructed trajectory containing the mined behavioral pattern sequence is used as an additional input to the LSTM model. The LSTM model is incrementally trained in combination with historical data, so that the model can dynamically update the historical pattern library according to new data, while retaining old data and reducing the computational cost of retraining the model.

[0075] During incremental training, the model learns to focus on key behavioral patterns and establishes a strong correlation between specific behavioral patterns and their subsequent flight trajectories to enhance the accuracy of future trajectory predictions. For example, when new data shows that "circling" is more often followed by "high-speed evacuation," the LSTM model will adjust the prediction probability to enhance the accuracy of future trajectory predictions. Historical data includes historical behavioral pattern labels, multimodal fusion features, historical predicted flight data, and historical future real flight data.

[0076] As can be seen, the UAV detection and countermeasure method in this embodiment can also perform trajectory modeling based on historical trajectory data detected by the UAV to obtain a reconstructed trajectory. The reconstructed trajectory and historical data are then combined to incrementally train the LSTM model to optimize the model and enhance the accuracy of future trajectory prediction. Understandably, historical data includes behavioral pattern labels, multimodal fusion features, predicted flight data, and actual flight data obtained from each UAV detection. During incremental training, the data can be retrieved to optimize the model when the stored data reaches a preset amount after multiple detections, or it can be optimized based on the data obtained after each detection.

[0077] Reference Figure 7 , Figure 7 This is a schematic diagram of a drone detection and countermeasure system according to a first embodiment of the present invention. In the embodiment shown in the figure, the drone detection and countermeasure system includes a data preprocessing unit 101, a data communication unit 102, a feature extraction unit 103, a feature fusion unit 104, a trajectory prediction unit 105, a detection and countermeasure unit 106, and an incremental training unit 107. Detailed descriptions of each functional unit are as follows: The data preprocessing unit 101 is used to acquire task instructions and sensing signals from multiple different sensing devices in the monitoring area in real time, and to preprocess the task instructions and the sensing signals to obtain standardized modal data corresponding to the task instructions and each sensing signal. Data communication unit 102 is used to distribute the task instructions and standardized modal data corresponding to multiple sensing devices to different data topics according to preset topic rules using a distributed stream processing framework; Feature extraction unit 103 is used to perform feature extraction and high-level semantic extraction on the standardized modal data corresponding to the task instructions issued by the data topic in the distributed stream processing framework and each sensor signal, respectively, to obtain task context parameters and high-dimensional feature vectors corresponding to each sensor signal. The feature fusion unit 104 is used to utilize a spatiotemporally aligned multimodal fusion algorithm to dynamically adjust the confidence scoring strategy based on the task context parameters, and fuse the high-dimensional feature vectors corresponding to each sensor signal to obtain multimodal fusion features that include the UAV state. The trajectory prediction unit 105 is used to predict and generate the flight trajectory of the UAV within a preset time period in the future based on the historical trajectory data of the UAV and the multimodal fusion features using a pre-trained neural network model. The detection and countermeasure unit 106 is used to determine the threat level and detect abnormal behavior based on the mission context parameters and flight trajectory, generate alarm information, generate a security countermeasure strategy based on the alarm information, and execute it. The incremental training unit 107 is used to perform trajectory modeling using an HMM model based on the historical trajectory data of the UAV, to obtain a reconstructed trajectory containing a sequence of UAV behavior patterns, and to incrementally train the pre-trained neural network model using an incremental learning mechanism based on the historical data corresponding to the historical trajectory data of the UAV and the reconstructed trajectory.

[0078] In some embodiments, the data communication unit 102 is further configured to: The threat level and abnormal behavior information in the alarm information are written into the topic corresponding to the distributed stream processing framework; The threat level and abnormal behavior information in the distributed stream processing framework are pushed to the outside for display using the WebSocket protocol, so as to be used for alarm pop-ups in the human-computer interaction interface.

[0079] In some embodiments, the feature fusion unit 104 is specifically used for: The Kalman filter algorithm is used to perform spatiotemporal alignment on all high-dimensional feature vectors. The Bayesian confidence estimation method is used to score the confidence of the high-dimensional feature vectors corresponding to each spatiotemporally aligned sensor signal based on the task context parameters. High-dimensional feature vectors with confidence scores below a preset threshold are filtered out from the high-dimensional feature vectors corresponding to each sensor signal. Then, the extended Kalman filter algorithm is used to fuse the remaining high-dimensional feature vectors after filtering based on the confidence scores to obtain multimodal fusion features that include the UAV state. The sensor signals include electromagnetic signals, perception signals, and environmental signals. The UAV state includes the UAV's real-time position, speed, and heading angle.

[0080] In some embodiments, the detection and countermeasure unit 106 is specifically used for: Based on the task context parameters and flight trajectory, the Drools rule engine is used to obtain the threat level of the intruding drones, thereby determining the threat level; wherein, the threat level reflects the number, location, and flight speed of the intruding drones; The isolated forest algorithm is used to identify abnormal flight behavior on the flight trajectory to obtain abnormal flight behavior data. A pre-trained classifier is then used to identify the pattern intent of the abnormal flight behavior data to obtain the UAV's behavioral intent. If the threat level reaches a preset level, and / or the drone's behavior is abnormal, an alarm message will be generated.

[0081] In some embodiments, the detection countermeasure unit 106 is further specifically used for: Based on each UAV as a node, the multimodal fusion features and predicted flight trajectories of the UAVs are recorded in a graph network structure to obtain a UAV relationship graph. The graph neural network is then used to analyze the UAV relationship graph to infer the cooperative relationships of multiple UAVs and generate a global situation map of the monitoring area. Based on the threat level, abnormal behavior information, and global situation map, alarm information is generated, and a security countermeasure strategy is generated and executed based on the alarm information.

[0082] In some embodiments, the feature fusion unit 104 is further specifically used for: The weights of each feature in the multimodal fusion feature are adaptively adjusted based on the scene corresponding to the historical trajectory data of the drone.

[0083] This invention provides a drone detection and countermeasure system that improves data processing efficiency and enhances adaptability to complex environments through multimodal data fusion, distributed processing, and a rapid response mechanism. Furthermore, by reconstructing and verifying behavior to optimize the LSTM model, it improves the accuracy of drone behavior prediction and supports dynamic analysis and real-time decision optimization of drone behavior in complex scenarios.

[0084] Reference Figure 8 , Figure 8 This is a schematic block diagram of a drone detection and countermeasure system provided in the second embodiment of the present invention.Figure 8 As shown, the drone detection and countermeasure system of this embodiment includes a server 301, several different types of sensing devices 302 installed in the monitoring area, several different types of jamming devices 303 installed in the monitoring area, and a terminal 304. The aforementioned sensing devices 302 are used to collect information about drones in the intrusion monitoring area in real time from different dimensions and generate sensing signals. The sensing devices 302 may include radar, radio transmitters and receivers, cameras, position sensors and climate sensors, etc. The sensing signals may include electromagnetic signals (radar signals, TDOA signals, drone communication signals), perception signals (images, videos and sounds acquired by the camera), environmental signals (geographic coordinates collected by the position sensor, wind speed and rainfall collected by the climate sensor), etc. The server 301 is used to acquire task instructions and sensing signals from multiple sensing devices 302 in the monitoring area in real time, and preprocess the task instructions and sensing signals to obtain standardized modal data corresponding to the task instructions and each sensing signal; perform feature extraction and high-level semantic extraction in parallel on the standardized modal data corresponding to the task instructions and each sensing signal to obtain task context parameters and high-dimensional feature vectors corresponding to each sensing signal; and use a spatiotemporally aligned multimodal fusion algorithm to dynamically adjust the confidence scoring strategy based on the task context parameters, and fuse the high-dimensional feature vectors corresponding to each sensing signal to obtain multimodal fusion features containing the UAV state; and use a pre-trained neural network model to predict and generate the UAV's flight trajectory within a preset time period in the future based on the UAV's historical trajectory data and the multimodal fusion features; perform threat level determination and abnormal behavior detection based on the task context parameters and flight trajectory, generate alarm information, and generate security countermeasures based on the alarm information; The terminal 304 is used to send task instructions and display alarm information for decision-makers to view; Several jamming devices 303 are connected to the server 301 via MQTT or HTTP interfaces to execute the security countermeasures pushed by the server 301. The jamming devices 303 can be radio frequency jammers, decoys, etc. Through the fast jamming mechanism of the MQTT protocol, the system response time can be shortened, improving the response speed by about 40% compared with the existing technology, effectively dealing with the threat of high-speed moving drones.

[0085] Specifically, see Figure 9 , Figure 9This is a schematic diagram of the internal architecture of server 301 in the UAV detection and countermeasure system of this embodiment. In this embodiment, server 301 mainly includes an input layer, a processing layer, and a storage layer. The functional flow executed by server 301 is implemented in the processing layer. At the beginning, the input layer can obtain sensor signals (electromagnetic signals, perception signals, environmental signals) and task instructions through dedicated hardware interfaces, TCP or API. After preprocessing, the signals are pushed to different topics (e.g., electromagnetic topics, perception topics, situational topics) in the distributed stream processing framework of the processing layer. Feature extraction is performed in parallel in the processing layer, and then feature fusion is performed to obtain multimodal fusion features. Then, the flight trajectory is predicted by the model to perform threat level determination, abnormal behavior detection, and global situational map generation. The model can also be incrementally trained by historical backtracking, that is, incremental training is performed using historical data.

[0086] In this embodiment, server 301 is equipped with multiple databases (in-memory database, time-series database, vector database, SQL structured database, document database, spatial database, etc.) to work with the distributed stream processing framework. It performs hierarchical storage for hot data (real-time trajectory), warm data (time-series data) and cold data (historical trajectory, semantic embedding), thereby optimizing query efficiency and data management.

[0087] In this embodiment, a unified database semantic module is also designed to facilitate the flow of data between different databases. By using injection inversion technology, the read, write, batch read, and batch write interfaces of different databases are unified, thus shielding the heterogeneity between different databases.

[0088] When data flows between databases, such as Figure 10 As shown, the application sends data access requests (such as reading real-time sensor values ​​or writing drone trajectories), which are passed to the database semantic module via "data request". Request types can include single read / write and batch read / write operations. The semantic module receives the request, parses it into standardized operation instructions ("semantic parsing" arrows), and injects them via IoC (Inversion of Control). The semantic module dynamically selects the corresponding database operation interface, shielding the differences between underlying SQL / non-SQL databases. For example, reading hot data (Redis) or cold data (MySQL) uses the same read interface, regardless of the database type. The abstract interface layer converts the semantic module's instructions into a general data operation protocol for driver invocation, ensuring compatibility with the driver layer. The general data operation protocol abstracts the details of database operations (such as SQL statements and database commands), providing a unified entry point for the driver layer. The driver layer then calls the corresponding database driver, such as JDBC or a database client, based on the operation type.

[0089] Since the data input from each sensor 302 is the latest data, it is stored in memory as hot data, supporting high-frequency read and write. This includes data such as real-time sensor values, real-time drone trajectory, and drone status. To reduce the pressure of hot layer data on memory, some data is periodically stored in a time-series database. Data migration can be performed based on timestamp timeouts. Data from different types of drones can be retained for different periods. The time-series database can be stored on solid-state drive-like hardware devices to accelerate IO efficiency. Alternatively, the time-series database's own caching technology can be used to provide 1-2GB of cache space in memory. Through column queries and data compression, query hit rates are improved, accelerating query efficiency. Traditional structured data, historical data, and document data are persistently stored using traditional cold layer technology. Data in the cold layer includes historical trajectories, logs, storage semantic embedding, HNSW indexes, geofencing, map information, and task data.

[0090] In the above technical solution, the UAV detection and countermeasure system provided in this embodiment can improve data processing efficiency and enhance adaptability to complex environments through multimodal data input, multimodal data fusion, distributed processing and rapid response mechanisms. Furthermore, by reconstructing and verifying behavior to optimize the LSTM model, it can improve the accuracy of UAV behavior prediction, support dynamic analysis and real-time decision optimization of UAV behavior in complex scenarios, and unify the read, write, batch read and batch write interfaces of different databases through injection inversion technology, support long-term continuous high-frequency read and write, improve data access speed by about 50%, and ensure system operation continuity through cross-database migration disaster recovery capability, greatly shortening fault recovery time.

[0091] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned UAV detection and countermeasure system and its various units or components can be referred to the corresponding descriptions in the foregoing method embodiments. For the sake of convenience and brevity, these details will not be repeated here.

[0092] In one embodiment, a computer device is provided, the internal structure of which can be shown as follows: Figure 11 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external devices via a network connection. Furthermore, the computer device may also include a display screen and input devices (e.g., mouse, keyboard, etc.) for interactive purposes.

[0093] Specifically, when the processor in the computer device executes the computer program, it implements the steps of the UAV detection and countermeasure method provided in the first, second, and third embodiments described above.

[0094] In one embodiment, the present invention may also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the UAV detection and countermeasure methods provided in the first, second, and third embodiments described above.

[0095] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0096] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0097] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for detecting and countering unmanned aerial vehicles (UAVs), characterized in that, The drone detection and countermeasures method includes: The task instructions and sensing signals from multiple different sensing devices in the monitoring area are acquired in real time, and the task instructions and sensing signals are preprocessed to obtain standardized modal data corresponding to the task instructions and each sensing signal. Feature extraction and high-level semantic extraction are performed in parallel on the standardized modal data corresponding to the task instructions and each sensor signal, respectively, to obtain the task context parameters and the high-dimensional feature vectors corresponding to each sensor signal. Using a spatiotemporally aligned multimodal fusion algorithm, the confidence scoring strategy is dynamically adjusted based on the task context parameters, and the high-dimensional feature vectors corresponding to each sensor signal are fused to obtain multimodal fusion features that include the UAV state. The flight trajectory of the UAV within a preset time period is predicted and generated using a pre-trained neural network model based on the UAV's historical trajectory data and the multimodal fusion features. Threat level determination and abnormal behavior detection are performed based on the mission context parameters and flight trajectory, alarm information is generated, and security countermeasures are generated and executed based on the alarm information.

2. The UAV detection and countermeasure method as described in claim 1, characterized in that, After obtaining the standardized modal data corresponding to the task instructions and each sensor signal, the process further includes: Using a distributed stream processing framework, the task instructions and standardized modal data corresponding to multiple sensing devices are distributed to different data topics according to preset topic rules.

3. The UAV detection and countermeasure method as described in claim 2, characterized in that, After generating the alarm information, the following is also included: The threat level and abnormal behavior information in the alarm information are written into the topic corresponding to the distributed stream processing framework; The threat level and abnormal behavior information in the distributed stream processing framework are pushed to the outside for display using the WebSocket protocol, so as to be used for alarm pop-ups in the human-computer interaction interface.

4. The UAV detection and countermeasure method as described in claim 1, characterized in that, The spatiotemporally aligned multimodal fusion algorithm dynamically adjusts the confidence scoring strategy based on the task context parameters, fuses the high-dimensional feature vectors corresponding to each sensor signal, and obtains multimodal fusion features that include the UAV state, including: The Kalman filter algorithm is used to perform spatiotemporal alignment on all high-dimensional feature vectors. The Bayesian confidence estimation method is used to score the confidence of the high-dimensional feature vectors corresponding to each spatiotemporally aligned sensor signal based on the task context parameters. High-dimensional feature vectors with confidence scores below a preset threshold are filtered out from the high-dimensional feature vectors corresponding to each sensor signal. Then, the extended Kalman filter algorithm is used to fuse the remaining high-dimensional feature vectors after filtering based on the confidence scores to obtain multimodal fusion features that include the UAV state. The sensor signals include electromagnetic signals, perception signals, and environmental signals. The UAV state includes the UAV's real-time position, speed, and heading angle.

5. The UAV detection and countermeasure method as described in claim 1, characterized in that, The threat level determination and abnormal behavior detection based on the task context parameters and flight trajectory, and the generation of alarm information, include: Based on the task context parameters and flight trajectory, the Drools rule engine is used to obtain the threat level of the intruding drones, thereby determining the threat level; wherein, the threat level reflects the number, location, and flight speed of the intruding drones; The isolated forest algorithm is used to identify abnormal flight behavior on the flight trajectory to obtain abnormal flight behavior data. A pre-trained classifier is then used to identify the pattern intent of the abnormal flight behavior data to obtain the UAV's behavioral intent. If the threat level reaches a preset level, and / or the drone's behavior is abnormal, an alarm message will be generated.

6. The UAV detection and countermeasure method as described in claim 1, characterized in that, The process of predicting and generating the drone's flight trajectory within a preset future time period using a pre-trained neural network model based on the drone's historical trajectory data and the multimodal fusion features further includes: Based on each UAV as a node, a graph network structure is used to record the multimodal fusion features and predicted flight trajectories of the UAVs, thus obtaining a UAV relationship graph. By analyzing the UAV relationship graph using a graph neural network, the cooperative relationships of multiple UAVs are inferred, and a global situation map of the monitoring area is generated.

7. The UAV detection and countermeasure method as described in claim 1, characterized in that, The drone detection and countermeasures method also includes: The weights of each feature in the multimodal fusion feature are adaptively adjusted based on the scene corresponding to the historical trajectory data of the drone.

8. The method for detecting and countering unmanned aerial vehicles (UAVs) as described in any one of claims 1-7, characterized in that, The drone detection and countermeasures method also includes: Based on historical trajectory data of UAVs, trajectory modeling is performed using an HMM model to obtain a reconstructed trajectory containing a sequence of UAV behavior patterns. Based on the historical trajectory data of the UAV and the reconstructed trajectory, the pre-trained neural network model is incrementally trained using an incremental learning mechanism; wherein, the historical data includes historical behavior pattern labels, multimodal fusion features, historical predicted flight data, and historical future real flight data.

9. A drone detection and countermeasure system, characterized in that, The drone detection and countermeasure system includes: The data preprocessing unit is used to acquire task instructions and sensing signals from multiple different sensing devices in the monitoring area in real time, and to preprocess the task instructions and the sensing signals to obtain standardized modal data corresponding to the task instructions and each sensing signal. The feature extraction unit is used to perform feature extraction and high-level semantic extraction in parallel on the standardized modal data corresponding to the task instructions and each sensor signal, respectively, to obtain the task context parameters and the high-dimensional feature vectors corresponding to each sensor signal. The feature fusion unit is used to utilize a spatiotemporally aligned multimodal fusion algorithm to dynamically adjust the confidence scoring strategy based on the task context parameters, and fuse the high-dimensional feature vectors corresponding to each sensor signal to obtain multimodal fusion features that include the UAV state. The trajectory prediction unit is used to predict and generate the flight trajectory of the UAV within a preset time period in the future based on the historical trajectory data of the UAV and the multimodal fusion features using a pre-trained neural network model. The detection and countermeasure unit is used to determine the threat level and detect abnormal behavior based on the mission context parameters and flight trajectory, generate alarm information, generate security countermeasure strategies based on the alarm information, and execute them.

10. The UAV detection and countermeasure system as described in claim 9, characterized in that, The drone detection and countermeasure system also includes: The data communication unit is used to distribute the task instructions and standardized modal data corresponding to multiple sensing devices to different data topics according to preset topic rules using a distributed stream processing framework; and / or The incremental training unit is used to perform trajectory modeling using an HMM model based on the historical trajectory data of the UAV, to obtain a reconstructed trajectory containing a sequence of UAV behavior patterns, and to incrementally train the pre-trained neural network model using an incremental learning mechanism based on the historical data corresponding to the historical trajectory data of the UAV and the reconstructed trajectory.

11. A drone detection and countermeasure system, characterized in that, It includes a server, several different types of sensors installed in the monitoring area, several different types of jamming devices and terminals installed in the monitoring area, among which, Several of the aforementioned sensing devices are used to collect information about drones in the intrusion monitoring area in real time from different dimensions and generate sensing signals; The server is used to acquire task instructions and sensor signals from multiple sensors in the monitoring area in real time, and preprocess the task instructions and sensor signals to obtain standardized modal data corresponding to the task instructions and each sensor signal. It then performs feature extraction and high-level semantic extraction in parallel on the standardized modal data corresponding to the task instructions and each sensor signal to obtain task context parameters and high-dimensional feature vectors corresponding to each sensor signal. Using a spatiotemporally aligned multimodal fusion algorithm, it dynamically adjusts the confidence scoring strategy based on the task context parameters and fuses the high-dimensional feature vectors corresponding to each sensor signal to obtain multimodal fusion features containing the UAV's state. Furthermore, it uses a pre-trained neural network model to predict and generate the UAV's flight trajectory within a preset time period based on the UAV's historical trajectory data and the multimodal fusion features. Finally, it performs threat level determination and abnormal behavior detection based on the task context parameters and flight trajectory, generates alarm information, and generates security countermeasures based on the alarm information. The terminal is used to send task instructions and display alarm information; Several jamming devices are connected to the server via MQTT or HTTP interfaces to execute the security countermeasures pushed by the server.

12. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the unmanned aerial vehicle (UAV) detection and countermeasure method as described in any one of claims 1 to 8.

13. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the unmanned aerial vehicle (UAV) detection and countermeasure method as described in any one of claims 1 to 8.

Images