Anti-drone detection system and method

Abstract

The application discloses an anti-UAV detection system and method, and belongs to the technical field of anti-UAVs, which comprises: a fixed detection base station for wide-area monitoring and target detection of airspace; a mobile countermeasure unit for flying to a target area to execute accurate detection and directional countermeasure tasks according to instructions; an edge intelligent node for receiving and fusing multi-source detection data, running a lightweight AI model to perform target identification and risk assessment, and generating a preliminary countermeasure strategy; a cloud digital twin platform for running a high-precision simulation model, verifying, optimizing and deducing the reported countermeasure strategy, and generating an optimal countermeasure scheme; and a unified command and control terminal for providing a man-machine interface and instruction issuing for an operator. The application solves the problem of target misidentification of a traditional anti-UAV system.

Description

Technical Field

[0001] This invention relates to the field of anti-drone technology, and more specifically to an anti-drone detection system and method. Background Technology

[0002] Currently, with the rapid popularization of drone technology, the number of consumer and industrial drones has surged, and incidents of unauthorized drones entering no-fly zones are frequent, posing a serious threat to public safety, military security, and critical infrastructure security. Traditional anti-drone systems mostly rely on a single detection method or a single countermeasure.

[0003] However, traditional anti-drone systems are susceptible to environmental interference, leading to misidentification or missed identification of targets; and the data from multiple sensors are not effectively fused, failing to fully leverage the advantages of each device.

[0004] Therefore, how to provide a counter-drone detection system and method is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides an anti-drone detection system and method to solve the technical problems existing in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: An anti-drone detection system includes: Fixed detection base stations are used for wide-area airspace surveillance and target detection; The mobile countermeasure unit receives instructions and flies to the target area to perform precision detection and targeted countermeasure missions. The edge intelligent node is communicatively connected to the fixed detection base station and the mobile countermeasure unit. It is used to receive and fuse multi-source detection data, run a lightweight AI model to perform target identification and risk assessment, and generate a preliminary countermeasure strategy. The cloud-based digital twin platform communicates with the edge intelligent nodes and mobile countermeasures units to run high-precision simulation models, verify, optimize, and deduce the reported countermeasures strategies, generate the optimal countermeasures scheme, and send the control commands of the optimal countermeasures scheme to the mobile countermeasures units. A unified command and control terminal provides operators with a human-machine interface and enables the issuance of commands.

[0007] Preferably, the fixed detection base station includes a radar device, an optoelectronic camera, and a radio detection device.

[0008] Preferably, the mobile countermeasure unit includes an unmanned aerial vehicle (UAV) platform equipped with a jammer and a detection payload.

[0009] Preferably, the edge intelligent node includes: Front-end inference unit: Runs lightweight AI models to achieve sensor-level real-time target detection; Regional Fusion Unit: Aggregates the results of multiple front-end inference units to perform trajectory correlation and threat fusion assessment; Strategy generation unit: Generates preliminary countermeasure strategies based on the fusion evaluation results.

[0010] Preferably, the lightweight AI model includes: The feature extraction branch is used to process sensor data of different modalities; An adaptive gating fusion module is used to dynamically calculate the contribution weight of each modality data in the current environment, and to perform weighted fusion of features based on the weights. The classification and regression header is used to output information about the target's identity, location, and threat level.

[0011] Preferably, the adaptive gating fusion module generates a set of gating coefficients based on the current fusion feature context using a differentiable neural network, and determines the proportion of each sensor mode in the final decision based on the gating coefficients.

[0012] Preferably, the cloud-based digital twin platform includes: A simulation engine used to simulate drone dynamics, sensor detection processes, and electromagnetic interference effects; The strategy sandbox is used to perform parallel simulations and performance evaluations of multiple schemes for the initial strategies reported by edge nodes. The model training module is used to aggregate anonymized data from each edge node based on a federated learning framework to continuously train and update the lightweight AI model.

[0013] Preferably, the unified command and control terminal includes: Virtual-Real Fusion Display Unit: Used to overlay virtual drone flight paths, threat levels, and countermeasures onto the real scene; Gesture interaction unit: used for recognizing target selection, plan confirmation, and command issuance; Voice control unit: used for natural language command input and voice feedback; Haptic feedback: Provides tactile alerts when high-threat targets are detected or when countermeasures are successfully taken.

[0014] A method for counter-drone detection system, comprising: Wide-area surveillance and target detection of the airspace are carried out through fixed detection base stations; The mobile countermeasure unit receives instructions and flies to the target area to perform precise detection and targeted countermeasure missions. Edge intelligent nodes receive and fuse multi-source detection data, run lightweight AI models to identify targets and assess risks, and generate preliminary countermeasure strategies. The cloud-based digital twin platform communicates with the edge intelligent nodes and mobile countermeasures units to run high-precision simulation models, verify, optimize, and deduce the reported countermeasures strategies, generate the optimal countermeasures scheme, and send the control commands of the optimal countermeasures scheme to the mobile countermeasures units. The unified command and control terminal provides operators with a human-machine interface and enables the issuance of commands. The mobile countermeasure unit receives instructions and executes a directional jamming mission.

[0015] As can be seen from the above technical solutions, compared with the prior art, the present invention discloses an anti-drone detection system and method, which can effectively improve the decision response speed and is more suitable for high-speed drone maneuvering scenarios; through the adaptive gating fusion module, the weight of each modality data is dynamically adjusted to maintain the target recognition accuracy in scenarios such as electromagnetic interference and severe weather; based on the target identity, location and threat level output by the model, the cloud can generate targeted countermeasures to avoid accidental damage. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the system structure provided by the present invention; Figure 2 This is a diagram of the lightweight AI model framework provided by the present invention. Detailed Implementation

[0018] 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 embodiments of the present invention, and not all embodiments. 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.

[0019] See Figure 1 This invention discloses an anti-drone detection system, comprising: Fixed detection base stations are used for wide-area airspace surveillance and target detection; The mobile countermeasure unit receives instructions and flies to the target area to perform precision detection and targeted countermeasure missions. The edge intelligent node is communicatively connected to the fixed detection base station and the mobile countermeasure unit. It is used to receive and fuse multi-source detection data, run a lightweight AI model to perform target identification and risk assessment, and generate a preliminary countermeasure strategy. The cloud-based digital twin platform communicates with the edge intelligent nodes and mobile countermeasures units to run high-precision simulation models, verify, optimize, and deduce the reported countermeasures strategies, generate the optimal countermeasures scheme, and send the control commands of the optimal countermeasures scheme to the mobile countermeasures units. A unified command and control terminal provides operators with a human-machine interface and enables the issuance of commands.

[0020] Specifically, fixed detection base stations include radar devices, photoelectric cameras, and radio detection equipment.

[0021] Specifically, mobile countermeasures units include drone platforms equipped with jammers and detection payloads.

[0022] Specifically, the jammer is a multi-band switchable jammer that supports directional jamming in the 2.4GHz, 5.8GHz and 1.5GHz bands. The selection of its operating frequency band is dynamically specified by the cloud-based digital twin platform based on the simulation analysis results of the target UAV signal.

[0023] Specifically, edge intelligent nodes include: Front-end inference unit: Runs lightweight AI models to achieve sensor-level real-time target detection; Regional Fusion Unit: Aggregates the results of multiple front-end inference units to perform trajectory correlation and threat fusion assessment; Strategy generation unit: Generates preliminary countermeasure strategies based on the fusion evaluation results.

[0024] See Figure 2 Specifically, lightweight AI models include: The feature extraction branch is used to process sensor data of different modalities; An adaptive gating fusion module is used to dynamically calculate the contribution weight of each modality data in the current environment, and to perform weighted fusion of features based on the weights. The classification and regression header is used to output information about the target's identity, location, and threat level.

[0025] The feature extraction branch employs a lightweight network architecture and designs dedicated processing paths for the characteristics of sensor data from different modalities, preserving key feature information while reducing computational load. Specifically, the first branch can be used for radar data processing, which includes: employing a lightweight convolutional neural network, with the input being a 32×32×3 matrix of three-dimensional data collected by the radar device, representing range, azimuth, and velocity. Through three layers of 3×3 depthwise separable convolution, batch normalization (BN), and ReLU activation functions, the motion features of the target are extracted. In this embodiment, the velocity change rate and trajectory curvature are extracted, and the output is a 64×8×8 feature map F1.

[0026] Specifically, the second branch can be used for photoelectric data processing, including: employing a lightweight vision Transformer, with the input being a 256×256×3 RGB image captured by a photoelectric camera. Through four attention block-feedforward network units, the image is segmented into 16×16 image blocks, and the visual features of the target are extracted. In this embodiment, the extracted features are the fuselage shape, the number of wings, and the paint color, with an output feature map F2 of 64×16×16 dimensions.

[0027] In one specific embodiment, a third branch may be included for radio data processing, specifically comprising: employing a multilayer perceptron, with the input being a feature vector of signal, frequency, bandwidth, and modulation scheme collected by a radio detection device, having a dimension of 1×24. Radio features of the target are extracted through three fully connected layers and a LeakyReLU activation function, wherein the number of neurons in the three fully connected layers are 64, 32, and 64, respectively. The output is a feature map F3 with a dimension of 64×1×1.

[0028] Specifically, the first branch, the second branch, and the third branch mentioned above are implemented in parallel.

[0029] Specifically, the adaptive gating fusion module generates a set of gating coefficients based on the current fusion feature context using a differentiable neural network. These gating coefficients determine the proportion of each sensor mode in the final decision, specifically including: Input the feature map output from the feature extraction branch, and the current environment parameter vector, which has a dimension of 1×8, including electromagnetic interference intensity, visibility, and target distance.

[0030] It incorporates a differentiable neural network, which includes two fully connected layers and a sigmoid activation function, outputting a corresponding number of gating coefficients, each corresponding to the contribution weight of the feature.

[0031] The features are fused using a weighted summation method, and the output is a fused feature map with dimensions of 64×16×16.

[0032] Specifically, the classification and regression head module outputs target information based on fused features, and is divided into a classification sub-module and a regression sub-module, with the following structure: The classification submodule flattens the fused features into a 16384-dimensional vector and inputs it into two fully connected layers with 1024 and 64 neurons respectively. Finally, the target identity probability distribution is output through the Softmax activation function, and the category with the highest probability is the target identity.

[0033] The regression submodule shares the first two fully connected layers with the classification submodule, and outputs the corresponding regression parameters through a linear activation function. Location coordinates: The target's (X, Y, Z) coordinates in the geodetic coordinate system; Threat Level: Based on target type, flight speed, and distance from the no-fly zone, a level of 0-5 is output. In this embodiment, level 0 is no threat; level 5 is a high-threat military drone.

[0034] Specifically, cloud-based digital twin platforms include: A simulation engine used to simulate drone dynamics, sensor detection processes, and electromagnetic interference effects; The strategy sandbox is used to perform parallel simulations and performance evaluations of multiple schemes for the initial strategies reported by edge nodes. The model training module is used to aggregate anonymized data from each edge node based on a federated learning framework to continuously train and update the lightweight AI model.

[0035] Specifically, the unified command and control terminal includes: Virtual-Real Fusion Display Unit: Used to overlay virtual drone flight paths, threat levels, and countermeasures onto the real scene; Gesture interaction unit: used for recognizing target selection, plan confirmation, and command issuance; Voice control unit: used for natural language command input and voice feedback; Haptic feedback: Provides tactile alerts when high-threat targets are detected or when countermeasures are successfully taken.

[0036] In one specific embodiment, to improve the adaptability of lightweight AI models in complex scenarios, this invention utilizes the model training module of a cloud-based digital twin platform to achieve model iteration based on a federated learning framework. The specific process is as follows: After completing the target recognition task, each edge intelligent node anonymizes the preprocessed data, model output results, and actual countermeasure effect labels. Each edge node uses local anonymized data to fine-tune the feature extraction branches and gating modules of the lightweight AI model, generating local model parameters; Receive local parameters uploaded by each edge node, aggregate them using a weighted average method, and generate global model parameters; The global parameters are distributed to all edge intelligent nodes to replace the original model parameters, thus completing the iteration.

[0037] Through the above iterations, the target recognition accuracy of lightweight AI models can be improved, the threat level determination error can be reduced, and the ability to adapt to complex scenarios such as electromagnetic interference and severe weather can be significantly enhanced.

[0038] A method for counter-drone detection system, comprising: Wide-area surveillance and target detection of the airspace are carried out through fixed detection base stations; The mobile countermeasure unit receives instructions and flies to the target area to perform precise detection and targeted countermeasure missions. Edge intelligent nodes receive and fuse multi-source detection data, run lightweight AI models to identify targets and assess risks, and generate preliminary countermeasure strategies. The cloud-based digital twin platform communicates with the edge intelligent nodes and mobile countermeasures units to run high-precision simulation models, verify, optimize, and deduce the reported countermeasures strategies, generate the optimal countermeasures scheme, and send the control commands of the optimal countermeasures scheme to the mobile countermeasures units. The unified command and control terminal provides operators with a human-machine interface and enables the issuance of commands. The mobile countermeasure unit receives instructions and executes a directional jamming mission.

[0039] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0040] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A counter-drone detection system, characterized in that, include: Fixed detection base stations are used for wide-area airspace surveillance and target detection; The mobile countermeasure unit receives instructions and flies to the target area to perform precision detection and targeted countermeasure missions. The edge intelligent node is communicatively connected to the fixed detection base station and the mobile countermeasure unit. It is used to receive and fuse multi-source detection data, run a lightweight AI model to perform target identification and risk assessment, and generate a preliminary countermeasure strategy. The cloud-based digital twin platform communicates with the edge intelligent nodes and mobile countermeasures units to run high-precision simulation models, verify, optimize, and deduce the reported countermeasures strategies, generate the optimal countermeasures scheme, and send the control commands of the optimal countermeasures scheme to the mobile countermeasures units. A unified command and control terminal provides operators with a human-machine interface and enables the issuance of commands.

2. The anti-drone detection system according to claim 1, characterized in that, The fixed detection base station includes radar devices, photoelectric cameras, and radio detection equipment.

3. The anti-drone detection system according to claim 1, characterized in that, The mobile countermeasures unit includes an unmanned aerial vehicle (UAV) platform equipped with jammers and detection payloads.

4. The anti-drone detection system according to claim 1, characterized in that, The edge intelligent nodes include: Front-end inference unit: Runs lightweight AI models to achieve sensor-level real-time target detection; Regional Fusion Unit: Aggregates the results of multiple front-end inference units to perform trajectory correlation and threat fusion assessment; Strategy generation unit: Generates preliminary countermeasure strategies based on the fusion evaluation results.

5. The anti-drone detection system according to claim 1, characterized in that, The lightweight AI model includes: The feature extraction branch is used to process sensor data of different modalities; An adaptive gating fusion module is used to dynamically calculate the contribution weight of each modality data in the current environment, and to perform weighted fusion of features based on the weights. The classification and regression header is used to output information about the target's identity, location, and threat level.

6. The anti-drone detection system according to claim 5, characterized in that, The adaptive gating fusion module generates a set of gating coefficients based on the current fusion feature context using a differentiable neural network, and determines the proportion of each sensor mode in the final decision based on the gating coefficients.

7. The anti-drone detection system according to claim 1, characterized in that, The cloud-based digital twin platform includes: A simulation engine used to simulate drone dynamics, sensor detection processes, and electromagnetic interference effects; The strategy sandbox is used to perform parallel simulations and performance evaluations of multiple schemes for the initial strategies reported by edge nodes. The model training module is used to aggregate anonymized data from each edge node based on a federated learning framework to continuously train and update the lightweight AI model.

8. The anti-drone detection system according to claim 1, characterized in that, The unified command and control terminal includes: Virtual-Real Fusion Display Unit: Used to overlay virtual drone flight paths, threat levels, and countermeasures onto the real scene; Gesture interaction unit: used for recognizing target selection, plan confirmation, and command issuance; Voice control unit: used for natural language command input and voice feedback; Haptic feedback: Provides tactile alerts when high-threat targets are detected or when countermeasures are successfully taken.

9. A detection method applied to the anti-drone detection system according to any one of claims 1-8, characterized in that, include: Wide-area surveillance and target detection of the airspace are carried out through fixed detection base stations; The mobile countermeasure unit receives instructions and flies to the target area to perform precise detection and targeted countermeasure missions. Edge intelligent nodes receive and fuse multi-source detection data, run lightweight AI models to identify targets and assess risks, and generate preliminary countermeasure strategies. The cloud-based digital twin platform communicates with the edge intelligent nodes and mobile countermeasures units to run high-precision simulation models, verify, optimize, and deduce the reported countermeasures strategies, generate the optimal countermeasures scheme, and send the control commands of the optimal countermeasures scheme to the mobile countermeasures units. The unified command and control terminal provides operators with a human-machine interface and enables the issuance of commands. The mobile countermeasure unit receives instructions and executes a directional jamming mission.

Images