A space-time joint perception unmanned aerial vehicle dual-domain countermeasure method and system

Abstract

The present application relates to the technical field of unmanned aerial vehicle countermeasure and low-altitude safety, and particularly relates to a kind of space-time joint sensing unmanned aerial vehicle dual-domain countermeasure method and system, first by simultaneously collecting the GNSS navigation signal and radio frequency signal of multiple target unmanned aerial vehicles;Second, the two kinds of signals collected are extracted and fused into a unified space-time feature matrix;Then, the space-time feature matrix is calculated using the Transformer model for space-time joint sensing, and the correlation features between the dual-link signals are captured through the multi-head attention mechanism;Next, the countermeasure resources of the navigation link and the remote control link are dynamically allocated according to the target state information and the countermeasure instructions are generated;Finally, the GNSS navigation deception signal and the radio frequency interference signal are generated synchronously and simultaneously act on the navigation link and the remote control link of the target unmanned aerial vehicle, thereby realizing the synchronization and cooperation of the unmanned aerial vehicle navigation and remote control dual-link countermeasure.

Description

Technical Field

[0001] This invention relates to the field of drone countermeasures and low-altitude security technology, and in particular to a dual-domain countermeasure method and system for drones based on joint spatial and temporal sensing. Background Technology

[0002] With the rapid iteration of drone technology, high-end industrial drones (such as industrial mapping drones and high-end security drones) are generally equipped with advanced features such as anti-jamming communication and backup navigation switching, posing a severe challenge to low-altitude safety management. Existing drone countermeasure technologies are mainly divided into two categories: GNSS navigation jamming technology and radio frequency countermeasure technology. However, whether using navigation deception or radio frequency suppression, traditional countermeasure systems usually only target a single link (i.e., the navigation link or the remote control link). High-end drones widely adopt a dual-link redundancy design. When an anomaly is detected in a certain link (such as GNSS navigation signal), it can immediately and automatically switch to a backup navigation mode (such as inertial navigation) or a backup communication link to continue the flight mission or return safely. This single-link countermeasure method is easily evaded by high-end drones, resulting in a countermeasure success rate of generally less than 70%, which cannot meet the stringent reliability requirements of high-end security scenarios such as airports and border defense. Summary of the Invention

[0003] The technical problem to be solved by this invention is: how to achieve synchronization and coordinated countermeasures between the navigation and remote control links of unmanned aerial vehicles (UAVs).

[0004] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: A dual-domain countermeasure method for unmanned aerial vehicles (UAVs) based on joint spatiotemporal sensing includes the following steps: S1. Collect GNSS navigation signals and radio frequency signals from multiple target UAVs; S2. Perform signal feature extraction on the GNSS navigation signal and radio frequency signal to obtain a space-time feature matrix; S3. Perform joint spatial-temporal sensing calculations on the spatio-temporal feature matrix using the Transformer model to obtain target state information of multiple target UAVs; S4. Based on the target status information, dynamically allocate countermeasure resources for the target UAV's navigation link and remote control link, and generate countermeasure commands; S5. Based on the countermeasure command, generate GNSS navigation decoy signal and radio frequency interference signal, and apply the GNSS navigation decoy signal and radio frequency interference signal to the navigation link and remote control link of the target UAV.

[0005] Another technical solution adopted in this invention is: A space-time joint sensing UAV dual-domain countermeasure system for executing the above-described space-time joint sensing UAV dual-domain countermeasure method includes: Antenna module, used to collect GNSS navigation signals and radio frequency signals from multiple target UAVs; The space-time joint sensing module is electrically connected to the antenna module and is used to perform signal feature extraction operations on the GNSS navigation signal and radio frequency signal to obtain a space-time feature matrix. The perception computing module is electrically connected to the spatiotemporal joint perception module. It uses the Transformer model to perform spatiotemporal joint perception computing on the spatiotemporal feature matrix to obtain target state information of multiple target UAVs. The multi-target scheduling module, electrically connected to the perception computing module, is used to dynamically allocate countermeasure resources for the target UAV navigation link and remote control link according to the target status information, and generate countermeasure commands. The dual-domain collaborative countermeasure module is electrically connected to the antenna module and the multi-target scheduling module, respectively. It is used to generate GNSS navigation decoy signals and radio frequency interference signals according to the countermeasure command, and apply the GNSS navigation decoy signals and radio frequency interference signals to the navigation link and remote control link of the target UAV through the antenna module.

[0006] The beneficial effects of this invention are as follows: This solution first simultaneously collects GNSS navigation and radio frequency signals from multiple target UAVs, ensuring from the source that subsequent countermeasures can cover all communication and navigation channels of the UAVs, avoiding signal omissions caused by single-link acquisition. Second, it extracts features from the two types of signals and fuses them into a unified spatiotemporal feature matrix, solving the problem of temporal and spatial misalignment between GNSS positioning data and radio frequency features, providing a precise spatiotemporal reference for simultaneous dual-link strikes. Then, it uses a Transformer model to perform joint spatiotemporal sensing calculations on the spatiotemporal feature matrix, capturing the correlation features between the dual-link signals through a multi-head attention mechanism, thereby accurately outputting the signal characteristics of each UAV in complex environments with multiple targets and strong interference. The method first identifies the link type and real-time status information of the UAV, solving the problem that traditional sensing methods struggle to simultaneously track the dual-link status of multiple targets. Then, it dynamically allocates countermeasure resources for navigation and remote control links based on target status information and generates countermeasure commands. This addresses the resource waste and signal interference issues caused by simply transmitting the same dual signals to all targets in a multi-aircraft environment, ensuring efficient and orderly allocation of countermeasure resources. Finally, it synchronously generates GNSS navigation decoy signals and radio frequency jamming signals, simultaneously applying them to the target UAV's navigation and remote control links. This prevents high-end UAVs from evading via inertial navigation or backup communication links, and also prevents them from manually correcting their trajectories after a single link is interfered with, achieving synchronous and coordinated dual-link strikes at the physical level. These steps are sequentially linked, from signal acquisition to feature fusion, from intelligent sensing to resource scheduling, and finally to dual-domain collaborative execution, forming a complete dual-link synchronous and coordinated countermeasure method that does not rely on a single link for countermeasures. This effectively addresses the problem of high-end UAVs using link redundancy for evasion, improving the success rate of countermeasures and ensuring the reliability and effectiveness of low-altitude security control. Attached Figure Description

[0007] Figure 1 This is a flowchart of the steps of the UAV dual-domain countermeasure method based on spatiotemporal joint sensing of the present invention; Figure 2 This is a connection block diagram of the UAV dual-domain countermeasure system with joint space-time perception according to the present invention; Label Explanation: 1. Antenna module; 2. Space-time joint sensing module; 3. Sensing computing module; 4. Multi-target scheduling module; 5. Dual-domain collaborative countermeasure module; 6. Parameter adaptive adjustment module; 7. Backend management module. Detailed Implementation

[0008] To explain in detail the technical content, objectives, and effects of the present invention, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0009] Please refer to Figure 1 One technical solution adopted in this invention is as follows: A dual-domain countermeasure method for unmanned aerial vehicles (UAVs) based on joint spatiotemporal sensing includes the following steps: S1. Collect GNSS navigation signals and radio frequency signals from multiple target UAVs; S2. Perform signal feature extraction on the GNSS navigation signal and radio frequency signal to obtain a space-time feature matrix; S3. Perform joint spatial-temporal sensing calculations on the spatio-temporal feature matrix using the Transformer model to obtain target state information of multiple target UAVs; S4. Based on the target status information, dynamically allocate countermeasure resources for the target UAV's navigation link and remote control link, and generate countermeasure commands; S5. Based on the countermeasure command, generate GNSS navigation decoy signal and radio frequency interference signal, and apply the GNSS navigation decoy signal and radio frequency interference signal to the navigation link and remote control link of the target UAV.

[0010] As can be seen from the above description, the beneficial effects of the present invention are as follows: This solution first simultaneously collects GNSS navigation and radio frequency signals from multiple target UAVs, ensuring from the source that subsequent countermeasures can cover all communication and navigation channels of the UAVs, avoiding signal omissions caused by single-link acquisition. Second, it extracts features from the two types of signals and fuses them into a unified spatiotemporal feature matrix, solving the problem of temporal and spatial misalignment between GNSS positioning data and radio frequency features, providing a precise spatiotemporal reference for simultaneous dual-link strikes. Then, it uses a Transformer model to perform joint spatiotemporal sensing calculations on the spatiotemporal feature matrix, capturing the correlation features between the dual-link signals through a multi-head attention mechanism, thereby accurately outputting the signal characteristics of each UAV in complex environments with multiple targets and strong interference. The method first identifies the link type and real-time status information of the UAV, solving the problem that traditional sensing methods struggle to simultaneously track the dual-link status of multiple targets. Then, it dynamically allocates countermeasure resources for navigation and remote control links based on target status information and generates countermeasure commands. This addresses the resource waste and signal interference issues caused by simply transmitting the same dual signals to all targets in a multi-aircraft environment, ensuring efficient and orderly allocation of countermeasure resources. Finally, it synchronously generates GNSS navigation decoy signals and radio frequency jamming signals, simultaneously applying them to the target UAV's navigation and remote control links. This prevents high-end UAVs from evading via inertial navigation or backup communication links, and also prevents them from manually correcting their trajectories after a single link is interfered with, achieving synchronous and coordinated dual-link strikes at the physical level. These steps are sequentially linked, from signal acquisition to feature fusion, from intelligent sensing to resource scheduling, and finally to dual-domain collaborative execution, forming a complete dual-link synchronous and coordinated countermeasure method that does not rely on a single link for countermeasures. This effectively addresses the problem of high-end UAVs using link redundancy for evasion, improving the success rate of countermeasures and ensuring the reliability and effectiveness of low-altitude security control.

[0011] Furthermore, step S2 specifically includes the following steps: The GNSS navigation signal is parsed to obtain the position coordinates of the target UAV. Parallel acquisition of radio frequency signals is performed to extract the amplitude, phase, and frequency characteristics of the radio frequency signals; A clock synchronization algorithm is used to synchronize the position coordinates, amplitude, phase and frequency features to generate a space-time feature matrix.

[0012] As described above, by parsing the GNSS navigation signal to obtain the position coordinates of the target UAV, a spatial reference is provided for dual-link countermeasures. At the same time, the radio frequency signal is acquired in parallel to extract amplitude, phase, and frequency features, solving the problem that a single channel cannot obtain multi-dimensional radio frequency features. Then, a clock synchronization algorithm is used to synchronously fuse the position coordinates and radio frequency features to generate a spatiotemporal feature matrix, solving the problem of spatiotemporal misalignment of multi-source data. These operations provide high-quality input for the Transformer model, ensuring the perception accuracy of subsequent dual-link collaborative countermeasures.

[0013] Furthermore, step S4 specifically includes the following steps: Based on the flight status in the target status information, multiple target drones are classified into different threat levels; An improved genetic algorithm is used to dynamically allocate countermeasure resources for the target UAV's navigation and remote control links. The fitness function of the improved genetic algorithm is: ; in, , and All are weighting coefficients, and , This is the scheduling delay value. The current number of counter-drones, To maximize the number of counter-drones, To disrupt resource utilization, To disrupt the maximum utilization of resources; Based on the threat level, the countermeasure resources are preferentially allocated to high-threat target drones, and countermeasure commands are generated.

[0014] As described above, multiple target UAVs are classified into different threat levels based on their flight status in the target status information, thus solving the problem of resource allocation priority in a multi-UAV environment. An improved genetic algorithm is used to dynamically allocate countermeasure resources for navigation and remote control links. Its fitness function comprehensively considers scheduling delay, countermeasure quantity, and resource utilization, solving the problems of interference and low resource utilization caused by fixed allocation methods. Furthermore, countermeasure resources are preferentially allocated to high-threat targets based on threat level. This scheduling method ensures the orderliness and efficiency of synchronous and coordinated countermeasures.

[0015] Furthermore, the Transformer model in step S3 includes an input layer, a spatiotemporal feature extraction layer, a multi-head attention layer, a fully connected layer, and an output layer connected in sequence. The input layer receives the spatiotemporal feature matrix, converts it into a feature vector, and outputs it to the spatiotemporal feature extraction layer; The spatiotemporal feature extraction layer contains multiple neurons, which perform spatiotemporal local feature extraction on the feature vector and output it to the multi-head attention layer; The multi-head attention layer employs the Scaled Dot-Product Attention mechanism, and its attention calculation formula is as follows: ; ; in, For querying the matrix, The key matrix, For value matrices, To query the dimensions of the key matrix and the key matrix, To focus on the number of heads, To output the weight matrix; The multi-head attention layer outputs global features to the fully connected layer; The fully connected layer contains multiple layers of neurons, and the output of the multiple neurons is fused and dimensionality reduced before being sent to the output layer. The output layer outputs target status information, which includes the target UAV's position coordinates, movement trajectory, threat level, and link type.

[0016] As described above, the Transformer model receives the spatiotemporal feature matrix through the input layer and converts it into feature vectors. It then extracts local features through the spatiotemporal feature extraction layer and captures the global dependencies of the spatiotemporal features through the multi-head attention layer. This solves the problem that traditional networks cannot simultaneously focus on the features associated with multiple targets and links. Finally, it performs feature fusion and dimensionality reduction through the fully connected layer and outputs target status information (location, trajectory, threat level, link type). This end-to-end perception network provides accurate target indication for synchronous and coordinated countermeasures.

[0017] Furthermore, the position coordinates of the target UAV are obtained using the BeiDou differential positioning algorithm, and the BeiDou differential positioning formula is as follows: ; ; in, This is the differential correction amount. To observe pseudorange, Given the pseudorange of the base station, This is the error correction amount. This represents the true pseudo-distance.

[0018] As can be seen from the above description, the BeiDou differential positioning algorithm is used to obtain position coordinates. Common errors such as satellite clock error and atmospheric delay are eliminated by the differential correction formula. Then, the corrected true pseudorange is used to obtain high-precision positioning, which solves the problem of low accuracy of ordinary single-point positioning. It provides a high-precision position reference for the accurate generation of navigation decoy signals and dual-link collaborative countermeasures.

[0019] Furthermore, it also includes the following steps: The location coordinates, motion trajectory, and link type of the target UAV are collected, and the interference parameters are dynamically adjusted.

[0020] As described above, by collecting the target UAV's position coordinates, motion trajectory, and link type, a feedback loop for the countermeasure effect is established, and the interference parameters are dynamically adjusted accordingly. This enables the system to adapt to changes in target distance, activation of anti-interference mode, and other state changes, maintaining optimal countermeasure effectiveness at all times, thus solving the problem of poor adaptability of fixed-parameter countermeasures.

[0021] Furthermore, a fuzzy control algorithm is used to dynamically adjust the interference parameters. The adaptive adjustment formula for the parameters of the fuzzy control algorithm is as follows: ; in, The adjusted interference power, As the reference interference power, For adjustment coefficients, The distance between the drone and the countermeasures system, For the minimum countermeasure distance, This represents the maximum countermeasure distance.

[0022] As can be seen from the above description, by adopting a fuzzy control algorithm to dynamically adjust the interference power, and by automatically adjusting the output power according to the distance between the UAV and the countermeasure system using a formula, the system can achieve on-demand output: reducing power when close to avoid interfering with legitimate equipment, and increasing power when far away or in anti-interference mode to ensure the countermeasure effect, thus effectively balancing the countermeasure effectiveness and electromagnetic compatibility.

[0023] Please refer to Figure 2 Another technical solution adopted in this invention is: A space-time joint sensing UAV dual-domain countermeasure system for executing the above-described space-time joint sensing UAV dual-domain countermeasure method includes: Antenna module 1 is used to collect GNSS navigation signals and radio frequency signals from multiple target UAVs; The space-time joint sensing module 2 is electrically connected to the antenna module 1 and is used to perform signal feature extraction operations on the GNSS navigation signal and radio frequency signal to obtain a space-time feature matrix. The perception computing module 3 is electrically connected to the spatiotemporal joint perception module 2. It uses the Transformer model to perform spatiotemporal joint perception computing on the spatiotemporal feature matrix to obtain target state information of multiple target UAVs. The multi-target scheduling module 4 is electrically connected to the perception computing module 3 and is used to dynamically allocate countermeasure resources for the target UAV navigation link and remote control link according to the target status information, and generate countermeasure commands. The dual-domain collaborative countermeasure module 5 is electrically connected to the antenna module 1 and the multi-target scheduling module 4, respectively. It is used to generate GNSS navigation decoy signals and radio frequency interference signals according to the countermeasure command, and apply the GNSS navigation decoy signals and radio frequency interference signals to the navigation link and remote control link of the target UAV through the antenna module 1.

[0024] As can be seen from the above description, the beneficial effects of the present invention are as follows: In this scheme, antenna module 1 is responsible for collecting GNSS navigation and radio frequency signals from multiple target UAVs and accurately transmitting countermeasure signals to the target's dual links. Space-time joint sensing module 2 is electrically connected to antenna module 1, extracting features from the collected signals and generating a space-time feature matrix to provide a unified spatiotemporal reference for subsequent sensing. Sensing computing module 3 is electrically connected to space-time joint sensing module 2, using a Transformer model to perform space-time joint sensing calculations on the space-time feature matrix, outputting target status information (position, trajectory, threat level, link type) for multiple UAVs. Multi-target scheduling module 4 is electrically connected to sensing computing module 3, dynamically allocating countermeasure resources for navigation and remote control links based on target status information and generating countermeasure commands. Dual-domain collaborative countermeasure module 5 is electrically connected to multi-target scheduling module 4 and also to antenna module 1, synchronously generating GNSS navigation decoy signals and radio frequency interference signals based on the countermeasure commands, and simultaneously applying them to the target UAV's navigation and remote control links through antenna module 1. The modules are electrically connected sequentially according to the signal flow, forming a complete, low-latency countermeasure data chain from signal acquisition, feature fusion, intelligent sensing, resource scheduling to dual-domain execution. This modular architecture not only solidifies the synchronous and collaborative countermeasure method into a reliable hardware system, but also facilitates independent upgrades of each module (such as replacing with a higher-precision GNSS receiver or upgrading the Transformer model). It also supports outdoor mobile and fixed deployments, providing a ready-to-use engineering solution for high-end security scenarios such as airports and border defense.

[0025] Furthermore, it also includes a parameter adaptive adjustment module 6, which is electrically connected to the dual-domain collaborative countermeasure module 5, and is used to collect the position coordinates, motion trajectory and link type of the target UAV, and dynamically adjust the interference parameters.

[0026] As can be seen from the above description, by setting the parameter adaptive adjustment module 6 and electrically connecting it with the dual-domain collaborative countermeasure module 5, the system is upgraded from open-loop control to closed-loop adaptive control, which can provide real-time feedback on the target status and dynamically adjust the interference parameters, greatly improving the robustness and countermeasure continuity of the system in dynamic environments.

[0027] Furthermore, it also includes a background management module 7, which is electrically connected to the multi-target scheduling module 4 and the parameter adaptive adjustment module 6, respectively, and is used to configure system parameters, monitor the working status of each module, and store countermeasure data.

[0028] As can be seen from the above description, by setting up a background management module 7, which is electrically connected to the multi-target scheduling module 4 and the parameter adaptive adjustment module 6 respectively, the system is used to configure system parameters, monitor the working status of each module, and store countermeasure data. This enables the system to have good configurability, monitorability, and traceability, meeting the operational needs of high-end security scenarios for system control and data traceability.

[0029] Please refer to Figure 1 Embodiment 1 of the present invention is as follows: A dual-domain countermeasure method for unmanned aerial vehicles (UAVs) based on joint spatiotemporal sensing includes the following steps: S1. Collect GNSS navigation signals and radio frequency signals from multiple target UAVs; Deploy antennas (e.g., high-gain dual-band array antennas with a gain of 8dBi) around the airport runway to simultaneously receive signals from all intruding drones; collect GNSS navigation signals using a BeiDou / GPS dual-mode receiver, supporting BeiDou B1 / B3 bands and GPS L1 / L2 bands; and collect radio frequency signals in parallel using a multi-channel SDR device with a sampling rate of 100MHz and a sampling accuracy of 16bit, covering the 2.4GHz, 5.8GHz, and 900MHz frequency bands.

[0030] S2. Perform signal feature extraction on the GNSS navigation signal and radio frequency signal to obtain a space-time feature matrix; Step S2 specifically includes the following steps: The GNSS navigation signal is analyzed, and the BeiDou differential positioning algorithm is used to obtain the position coordinates of the target UAV with a positioning accuracy of ≤1m. Parallel acquisition of radio frequency signals can be performed, and short-time Fourier transform can be used to extract the amplitude, phase, and frequency characteristics of the radio frequency signals. A clock synchronization algorithm is used to synchronize the position coordinates, amplitude, phase and frequency features to generate a space-time feature matrix.

[0031] S3. Perform joint spatial-temporal sensing calculations on the spatio-temporal feature matrix using the Transformer model to obtain target state information of multiple target UAVs; The Transformer model in step S3 includes an input layer, a spatiotemporal feature extraction layer, a multi-head attention layer, a fully connected layer, and an output layer connected in sequence. The input layer receives the spatiotemporal feature matrix, converts it into a feature vector, and outputs it to the spatiotemporal feature extraction layer; The spatiotemporal feature extraction layer contains multiple neurons (256 neurons, with ReLU activation function) to extract spatiotemporal local features from the feature vector and output them to the multi-head attention layer (8 heads, 512 hidden layer dimensions). The multi-head attention layer employs the Scaled Dot-Product Attention mechanism, and its attention calculation formula is as follows: ; ; in, For querying the matrix, The key matrix, For value matrices, To query the dimensions of the key matrix and the key matrix, For the number of attention heads (in this embodiment) =8), The output weight matrix is ​​defined as follows: target lock response time ≤ 200ms, trajectory tracking error ≤ 0.5m.

[0032] The multi-head attention layer outputs global features to the fully connected layer; The fully connected layer contains multiple layers of neurons (containing 128 and 64 neurons, with ReLU as the activation function). After feature fusion and dimensionality reduction, the multiple layers of neurons are output to the output layer. The output layer outputs target status information, which includes the target UAV's position coordinates, movement trajectory, threat level, and link type.

[0033] The position coordinates of the target UAV were obtained using the BeiDou differential positioning algorithm, and the BeiDou differential positioning formula is as follows: ; ; in, This is the differential correction amount. To observe pseudorange, Given the pseudorange of the base station, This is the error correction amount. This represents the true pseudo-distance.

[0034] S4. Based on the target status information, dynamically allocate countermeasure resources for the target UAV's navigation link and remote control link, and generate countermeasure commands; Step S4 specifically includes the following steps: Based on the flight status in the target status information, multiple target drones are classified into different threat levels (high, medium and low levels respectively). An improved genetic algorithm is used to dynamically allocate countermeasure resources for the target UAV's navigation and remote control links. The fitness function of the improved genetic algorithm is: ; in, , and All are weighting coefficients, and , This is the scheduling delay value. This represents the current number of countermeasure drones (8 in this example). To maximize the number of counter-drones, To disrupt resource utilization, To disrupt the maximum utilization of resources; Based on the threat level, the countermeasure resources are preferentially allocated to high-threat target drones, and countermeasure commands are generated.

[0035] S5. Based on the countermeasure command, generate GNSS navigation decoy signal and radio frequency interference signal, and apply the GNSS navigation decoy signal and radio frequency interference signal to the navigation link and remote control link of the target UAV.

[0036] The system employs "on-chip simulation" technology for satellite navigation signals to generate navigation decoy signals with an adjustable power range of 0dBm to 30dBm and a decoy accuracy of ≤0.5m. It also employs an adaptive interference algorithm to generate narrowband targeting jamming signals with an adjustable interference power range of 0dBm to 40dBm and an adjustable interference bandwidth range of 1MHz to 20MHz.

[0037] It also includes the following steps: The location coordinates, motion trajectory, and link type of the target UAV are collected, and the interference parameters are dynamically adjusted.

[0038] The interference parameters are dynamically adjusted using a fuzzy control algorithm. The adaptive adjustment formula for the parameters of the fuzzy control algorithm is as follows: ; in, The adjusted interference power, As the reference interference power, For adjustment coefficients, The distance between the drone and the countermeasures system, For the minimum countermeasure distance, This represents the maximum countermeasure distance.

[0039] Please refer to Figure 2 Embodiment two of the present invention is as follows: A space-time joint sensing UAV dual-domain countermeasure system for executing the above-described space-time joint sensing UAV dual-domain countermeasure method includes: Antenna module 1 is used to collect GNSS navigation signals and radio frequency signals from multiple target UAVs; The space-time joint sensing module 2 is electrically connected to the antenna module 1 and is used to perform signal feature extraction operations on the GNSS navigation signal and radio frequency signal to obtain a space-time feature matrix. The Space-Time Joint Sensing Module 2, as the core sensing unit of the system, integrates a GNSS receiver and a multi-channel SDR module to construct a "millimeter-level static and centimeter-level dynamic" spatiotemporal digital base, enabling precise perception of multiple targets on UAVs. The GNSS receiver supports dual-mode positioning with BeiDou and GPS, employing the BeiDou differential positioning algorithm and combining "5G+BeiDou" low-altitude sensing network technology, achieving a positioning accuracy of ≤1m, with static positioning accuracy reaching millimeter level and dynamic positioning accuracy reaching centimeter level. It can collect GNSS navigation signals from UAVs and analyze their position coordinates. The multi-channel SDR module has a sampling rate of 100MHz, a sampling accuracy of ≥16bit, supports parallel acquisition of ≥8 signals, and a signal acquisition delay of ≤3ms. It can collect remote control and image transmission radio frequency signals from UAVs and extract signal features. The module synchronously transmits the collected GNSS positioning data and radio frequency signal features to the Transformer Space-Time Joint Sensing Model, providing data support for multi-target locking.

[0040] The perception computing module 3 is electrically connected to the spatiotemporal joint perception module 2. It uses the Transformer model to perform spatiotemporal joint perception computing on the spatiotemporal feature matrix to obtain target state information of multiple target UAVs. The perception computing module 3 employs deep learning algorithms to optimize the accuracy of spatiotemporal joint perception, enabling precise locking and trajectory prediction of multi-target UAVs. The specific workflow of the model is as follows: the input layer receives GNSS positioning data (position coordinates, velocity) and radio frequency signal features (amplitude, phase, frequency) output from the spatiotemporal joint perception module 2, converting them into a unified-dimensional spatiotemporal feature matrix; the spatiotemporal feature extraction layer contains 256 neurons, using ReLU activation to extract features from the spatiotemporal feature matrix and capture the spatiotemporal variation patterns of the UAV; the 8-head attention layer uses Scaled Dot-Product. The attention mechanism, with a hidden layer dimension of 512, focuses on the key features of multi-target UAVs, capturing the global dependencies of spatiotemporal features to achieve multi-target differentiation and precise locking. The two fully connected layers contain 128 and 64 neurons respectively, and both use ReLU activation functions to fuse and reduce the dimensionality of the extracted spatiotemporal features. The output layer outputs the locking information of multi-target UAVs (position coordinates, motion trajectory, threat level, link type) and transmits it to the multi-target scheduling module 4. The model has an initial learning rate of 0.0005 and is trained using the AdamW optimizer. It can lock ≥8 targets, track the trajectory error ≤0.5m, and lock the target response time ≤200ms, and can work stably in complex electromagnetic environments.

[0041] The multi-target scheduling module 4 is electrically connected to the perception computing module 3 and is used to dynamically allocate countermeasure resources for the target UAV navigation link and remote control link according to the target status information, and generate countermeasure commands. The multi-target scheduling module 4 employs an improved genetic algorithm and embeds anti-task allocation logic to achieve dynamic allocation of countermeasure resources for multi-target UAVs. The module receives multi-target locking information output by the Transformer model, classifies UAVs into threat levels (high, medium, and low), and prioritizes the allocation of interference resources to high-threat targets (such as UAVs near the core area). It supports parallel countermeasures against ≥8 UAVs simultaneously, dynamically allocates GNSS navigation decoy resources and radio frequency interference resources to avoid interference conflicts during multi-target countermeasures, and has a scheduling delay of ≤500ms to ensure the orderliness and accuracy of countermeasures in multi-UAV environments. At the same time, the module is linked with the backend management module 7 and can receive manual intervention commands to adjust the countermeasure strategy.

[0042] The dual-domain collaborative countermeasure module 5 is electrically connected to the antenna module 1 and the multi-target scheduling module 4, respectively. It is used to generate GNSS navigation decoy signals and radio frequency interference signals according to the countermeasure command, and apply the GNSS navigation decoy signals and radio frequency interference signals to the navigation link and remote control link of the target UAV through the antenna module 1.

[0043] The dual-domain collaborative countermeasure module 5, as the core countermeasure unit of the system, includes a GNSS navigation decoy unit and a radio frequency jamming unit. It achieves dual-domain collaborative countermeasures using GNSS and radio frequency, simultaneously attacking the UAV's navigation and remote control links. The GNSS navigation decoy unit employs on-chip satellite navigation signal simulation technology to generate BeiDou and GPS navigation decoy signals. The signal power is adjustable from 0dBm to 30dBm, with a decoy accuracy of ≤0.5m. It is compatible with the anti-jamming navigation modules of high-end UAVs, generating "false but reliable" coordinate information to interfere with the UAV's navigation and positioning, forcing the UAV to deviate from its flight path, return to base, or make an emergency landing. The radio frequency jamming unit employs... An adaptive jamming algorithm generates narrowband targeted jamming signals, covering the entire link of UAV remote control and image transmission. The jamming power is adjustable from 0dBm to 40dBm, and the jamming bandwidth is adjustable from 1MHz to 20MHz. It can effectively suppress anti-jamming links such as frequency hopping communication and encrypted communication of high-end UAVs and prevent UAVs from receiving ground control commands. The dual-domain collaborative countermeasure mechanism adopts a "synchronous start and cooperative action" mode. The GNSS navigation decoy unit first outputs a decoy signal to interfere with UAV navigation, and the radio frequency jamming unit outputs a jamming signal simultaneously to suppress the UAV link, forming a "double insurance" countermeasure mode with a countermeasure success rate of ≥95%, avoiding the evasion of single link countermeasures.

[0044] It also includes a parameter adaptive adjustment module 6, which is electrically connected to the dual-domain cooperative countermeasure module 5. The parameter adaptive adjustment module 6 is used to collect the position coordinates, motion trajectory and link type of the target UAV, and dynamically adjust the interference parameters.

[0045] The parameter adaptive adjustment module 6 collects the UAV's position coordinates, flight speed, trajectory, and link type (remote control link / navigation link) information in real time. Combined with the environmental electromagnetic interference intensity, it dynamically adjusts the interference parameters of the dual-domain collaborative countermeasure module 5 through a fuzzy control algorithm. The adjustment logic is as follows: when the UAV is close to the countermeasure system (≤1km), the interference power is reduced (≤10dBm) to avoid interfering with nearby legitimate wireless equipment; when the UAV is far from the countermeasure system (>1km), the interference power is increased (≥20dBm) to ensure the countermeasure effect; when the UAV is in anti-interference mode, the interference power and signal strength are increased, and the interference frequency is adjusted to adapt to the UAV's frequency hopping link to ensure the effectiveness of the countermeasure; the adjustment response time is ≤300ms to achieve a balance between the countermeasure effect and environmental compatibility.

[0046] It also includes a background management module 7, which is electrically connected to the multi-target scheduling module 4 and the parameter adaptive adjustment module 6, respectively, and is used to configure system parameters, monitor the working status of each module, and store countermeasure data.

[0047] The backend management module 7 is used for system parameter configuration, status monitoring, data storage and log query, and supports manual intervention in the countermeasure process. It can configure the parameters of the space-time joint perception module 2 and the dual-domain collaborative countermeasure module 5, monitor the working status of each module, and store data such as multi-target locking information, countermeasure records, and interference parameter adjustment records. It also supports data backup and export. At the same time, it can display information such as the drone's location, movement trajectory, and countermeasure status, so that staff can keep track of the countermeasure situation in real time and adapt to the control needs of high-end security scenarios.

[0048] In summary, the present invention provides a dual-domain countermeasure method and system for UAVs based on joint space-time perception. First, by simultaneously acquiring GNSS navigation and radio frequency signals from multiple target UAVs, it ensures from the source that subsequent countermeasures can cover all communication and navigation channels of the UAVs, avoiding signal omissions caused by single-link acquisition. Second, it extracts features from the two types of signals and fuses them into a unified space-time feature matrix, solving the problem of temporal and spatial misalignment between GNSS positioning data and radio frequency features, providing a precise spatiotemporal reference for simultaneous dual-link strikes. Then, it uses a Transformer model to perform joint space-time perception calculations on the space-time feature matrix, capturing the correlation features between the dual-link signals through a multi-head attention mechanism, thereby enabling countermeasures against multiple targets. In complex environments with strong interference, this method accurately outputs the link type and real-time status information of each UAV, solving the problem that traditional sensing methods struggle to simultaneously track the dual-link status of multiple targets. Next, it dynamically allocates countermeasure resources for navigation and remote control links based on target status information and generates countermeasure commands. This solves the resource waste and signal interference problems caused by simply transmitting the same dual signals to all targets in multi-aircraft environments, ensuring efficient and orderly allocation of countermeasure resources. Finally, it synchronously generates GNSS navigation decoy signals and radio frequency jamming signals, simultaneously applying them to the target UAV's navigation and remote control links. This prevents high-end UAVs from evading via inertial navigation or backup communication links, and also prevents them from manually correcting their trajectories after a single link is interfered with, achieving synchronous and coordinated dual-link strikes at the physical level. These steps are sequentially linked, from signal acquisition to feature fusion, from intelligent sensing to resource scheduling, and then to dual-domain collaborative execution, forming a complete dual-link synchronous and coordinated countermeasure method that does not rely on a single link for countermeasures. This effectively addresses the problem of high-end UAVs using link redundancy to evade, improving the success rate of countermeasures and ensuring the reliability and effectiveness of low-altitude security control.

[0049] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for dual-domain countermeasures against unmanned aerial vehicles (UAVs) based on spatiotemporal joint sensing, characterized in that, Includes the following steps: S1. Collect GNSS navigation signals and radio frequency signals from multiple target UAVs; S2. Perform signal feature extraction on the GNSS navigation signal and radio frequency signal to obtain a space-time feature matrix; S3. Perform joint spatial-temporal sensing calculations on the spatio-temporal feature matrix using the Transformer model to obtain target state information of multiple target UAVs; S4. Based on the target status information, dynamically allocate countermeasure resources for the target UAV's navigation link and remote control link, and generate countermeasure commands; S5. Based on the countermeasure command, generate GNSS navigation decoy signal and radio frequency interference signal, and apply the GNSS navigation decoy signal and radio frequency interference signal to the navigation link and remote control link of the target UAV.

2. The UAV dual-domain countermeasure method based on spatiotemporal joint sensing according to claim 1, characterized in that, Step S2 specifically includes the following steps: The GNSS navigation signal is parsed to obtain the position coordinates of the target UAV. Parallel acquisition of radio frequency signals is performed to extract the amplitude, phase, and frequency characteristics of the radio frequency signals; A clock synchronization algorithm is used to synchronize the position coordinates, amplitude, phase and frequency features to generate a space-time feature matrix.

3. The UAV dual-domain countermeasure method based on spatiotemporal joint sensing according to claim 1, characterized in that, Step S4 specifically includes the following steps: Based on the flight status in the target status information, multiple target drones are classified into different threat levels; An improved genetic algorithm is used to dynamically allocate countermeasure resources for the target UAV's navigation and remote control links. The fitness function of the improved genetic algorithm is: ； in, , and All are weighting coefficients, and , This is the scheduling delay value. The current number of counter-drones, To maximize the number of counter-drones, To disrupt resource utilization, To disrupt the maximum utilization of resources; Based on the threat level, the countermeasure resources are preferentially allocated to high-threat target drones, and countermeasure commands are generated.

4. The UAV dual-domain countermeasure method based on spatiotemporal joint sensing according to claim 1, characterized in that, The Transformer model in step S3 includes an input layer, a spatiotemporal feature extraction layer, a multi-head attention layer, a fully connected layer, and an output layer connected in sequence. The input layer receives the spatiotemporal feature matrix, converts it into a feature vector, and outputs it to the spatiotemporal feature extraction layer; The spatiotemporal feature extraction layer contains multiple neurons, which perform spatiotemporal local feature extraction on the feature vector and output it to the multi-head attention layer; The multi-head attention layer employs the Scaled Dot-Product Attention mechanism, and its attention calculation formula is as follows: ； ； in, For querying the matrix, The key matrix, For value matrices, To query the dimensions of the key matrix and the key matrix, To focus on the number of heads, To output the weight matrix; The multi-head attention layer outputs global features to the fully connected layer; The fully connected layer contains multiple layers of neurons, and the output of these multiple neurons is fused and dimensionality reduced before being sent to the output layer. The output layer outputs target status information, which includes the target UAV's position coordinates, movement trajectory, threat level, and link type.

5. The UAV dual-domain countermeasure method based on spatiotemporal joint sensing according to claim 4, characterized in that, The position coordinates of the target UAV were obtained using the BeiDou differential positioning algorithm, and the BeiDou differential positioning formula is as follows: ； ； in, This is the differential correction amount. To observe pseudorange, Given the pseudorange of the base station, This is the error correction amount. This represents the true pseudo-distance.

6. The UAV dual-domain countermeasure method based on spatiotemporal joint sensing according to claim 1, characterized in that, It also includes the following steps: The location coordinates, motion trajectory, and link type of the target UAV are collected, and the interference parameters are dynamically adjusted.

7. The UAV dual-domain countermeasure method based on spatiotemporal joint sensing according to claim 6, characterized in that, The interference parameters are dynamically adjusted using a fuzzy control algorithm. The adaptive adjustment formula for the parameters of the fuzzy control algorithm is as follows: ； in, The adjusted interference power, As the reference interference power, For adjustment coefficients, The distance between the drone and the countermeasures system, For the minimum countermeasure distance, This represents the maximum countermeasure distance.

8. A space-time joint sensing UAV dual-domain countermeasure system for executing the space-time joint sensing UAV dual-domain countermeasure method of any one of claims 1 to 7, characterized in that, include: Antenna module, used to collect GNSS navigation signals and radio frequency signals from multiple target UAVs; The space-time joint sensing module is electrically connected to the antenna module and is used to perform signal feature extraction operations on the GNSS navigation signal and radio frequency signal to obtain a space-time feature matrix. The perception computing module is electrically connected to the spatiotemporal joint perception module. It uses the Transformer model to perform spatiotemporal joint perception computing on the spatiotemporal feature matrix to obtain target state information of multiple target UAVs. The multi-target scheduling module, electrically connected to the perception computing module, is used to dynamically allocate countermeasure resources for the target UAV navigation link and remote control link according to the target status information, and generate countermeasure commands. The dual-domain collaborative countermeasure module is electrically connected to the antenna module and the multi-target scheduling module, respectively. It is used to generate GNSS navigation decoy signals and radio frequency interference signals according to the countermeasure command, and apply the GNSS navigation decoy signals and radio frequency interference signals to the navigation link and remote control link of the target UAV through the antenna module.

9. The UAV dual-domain countermeasure system based on spatiotemporal joint sensing according to claim 8, characterized in that, It also includes a parameter adaptive adjustment module, which is electrically connected to the dual-domain collaborative countermeasure module. This module is used to collect the position coordinates, motion trajectory, and link type of the target UAV and dynamically adjust the interference parameters.

10. The UAV dual-domain countermeasure system based on spatiotemporal joint sensing according to claim 8, characterized in that, It also includes a background management module, which is electrically connected to the multi-target scheduling module and the parameter adaptive adjustment module, respectively, and is used to configure system parameters, monitor the working status of each module, and store countermeasure data.

Images