A tilt quadrotor unmanned aerial vehicle and unmanned aerial vehicle countermeasure system

Abstract

The application discloses a tilting four-rotor unmanned aerial vehicle and an unmanned aerial vehicle countermeasure system, and relates to the technical field of unmanned aerial vehicles. The system comprises: a flight control module, which is used for controlling the tilting four-rotor unmanned aerial vehicle to fly towards a target unmanned aerial vehicle, and adjusting the flight mode and the flight speed of the tilting four-rotor unmanned aerial vehicle in real time according to the distance and the speed difference during the flight; a target identification module, which is used for collecting video stream information of a target airspace, and detecting whether the target unmanned aerial vehicle exists in the video stream information frame by frame; a target tracking module, which is used for determining the distance and the speed difference between the target unmanned aerial vehicle and the tilting four-rotor unmanned aerial vehicle when the target unmanned aerial vehicle exists in the video stream information; and a net gun capturing module, which is used for calculating the shooting parameters of a capturing net when the distance between the target unmanned aerial vehicle and the tilting four-rotor unmanned aerial vehicle is less than a preset distance, and shooting the capturing net at the target unmanned aerial vehicle according to the shooting parameters. Through the above scheme, the unmanned aerial vehicle can be countermeasured accurately.

Description

Technical Field

[0001] The present invention relates to the technical field of drone countermeasures, and in particular to a tilt-rotating quad-rotor drone and a drone countermeasure system. Background Art

[0002] Traditional drone countermeasures rely primarily on ground-based radar, electronic jamming, or missile interception, which poses challenges such as high cost, significant collateral damage, and insufficient low-altitude target detection. However, a series of studies on drone netting technology have been conducted both domestically and internationally, primarily using drones to counter drones by dropping nets to capture target drones.

[0003] However, the main carriers of the current net capture system are fixed-wing drones or multi-rotor drones. Fixed-wing drones have the ability to search at high speed and over a large area, but they lack flexibility and cannot quickly adjust their course to point to the target drone, making it difficult to accurately track and attack the target drone; multi-rotor drones are maneuverable, can hover, and switch fire quickly, but their posture changes dramatically during flight, and it is difficult for the onboard camera to capture the target, which can easily lead to target loss. In addition, their speed and endurance are far inferior to those of fixed-wing drones, and their efficiency is not high.

[0004] In summary, the existing drone countermeasure technology is unable to accurately detect and track the target drone, resulting in the inability to accurately counter the target drone. Summary of the Invention

[0005] Based on this, it is necessary to provide a tilt-rotor quad-rotor UAV and a UAV countermeasure system to address the technical problem that existing technologies are unable to accurately detect and track target UAVs, resulting in the inability to accurately counter target UAVs.

[0006] The present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a tilt-rotating quad-rotor drone, characterized in that it includes a fuselage 1, a front wing 2 and a rear wing 3 provided on the fuselage 1, a single vertical tail 5 provided at the tail of the fuselage 1, and a tilt-rotating servo 6 provided inside the fuselage 1, wherein the single vertical tail 5 has a rudder 51;

[0008] Ailerons 21 are provided on both sides of the front wing 2, elevators 31 are provided on both sides of the rear wing 3, and tilt pods 4 are provided on both sides of the front wing 2 and the rear wing 3. The tilt pods 4 are tilted by the tilt servos 6.

[0009] The tilt pod 4 is provided with a rotor 41 and a winglet 42 , and the rotor 41 and the winglet 42 are tilted synchronously with the tilt of the tilt pod 4 ;

[0010] The head of the tilt-rotating quad-rotor UAV is provided with a net-catching mechanism 7 and a camera platform 8 , and the fuselage 1 is provided with an onboard device 9 .

[0011] In a second aspect, the present invention provides a UAV countermeasure system based on the tilt-quadrotor UAV, characterized in that the system is mounted on the tilt-quadrotor UAV, comprising:

[0012] A target recognition module is used to collect video stream information of the target airspace in real time through the camera gimbal 8, and detect whether there is a target drone in the video stream information frame by frame using the UAV-Detector algorithm;

[0013] a target tracking module, configured to determine the distance and speed difference between the target drone and the tilt-quadrotor drone when a target drone is present in the video stream information;

[0014] A flight control module is configured to control the tilt-quadrotor UAV to fly toward the target UAV, and dynamically adjust the tilt direction of the tilt pod 4 according to the distance and speed difference between the target UAV and the tilt-quadrotor UAV during flight, so as to change the lift direction of the tilt-quadrotor UAV through the rotor 41 and the winglet 42, thereby controlling the flight mode of the tilt-quadrotor UAV; dynamically adjust the deflection direction of the elevator 31, so as to change the aerodynamic lift of the rear wing 3, thereby controlling the pitch attitude of the tilt-quadrotor UAV; and dynamically adjust the deflection direction of the rudder 51, so as to control the yaw angle of the tilt-quadrotor UAV;

[0015] The net gun capture module is used to calculate the launch parameters of the capture net in real time according to the distance and speed difference when the distance between the target UAV and the tilt-quadrotor UAV is less than a preset distance, and control the capture net mechanism 7 to launch the capture net toward the target UAV according to the launch parameters.

[0016] Furthermore, the detecting whether the target UAV exists in the video stream information frame by frame by using the UAV-Detector algorithm specifically includes:

[0017] Extracting multiple frames of images of the target airspace from the video stream information according to a preset frame interval;

[0018] The UAV-Detector algorithm is used to detect the multiple frames of images frame by frame, and it is determined whether the target drone exists in the video stream information based on the detection results.

[0019] Furthermore, the UAV-Detector algorithm is used to perform frame-by-frame detection on the multiple frames of images, and judging whether the target UAV exists in the video stream information based on the detection results, specifically including:

[0020] Detect all drones in the video stream frame by frame using the UAV-Detector algorithm;

[0021] Each of the drones is compared with the drones in the pre-built target drone library, and whether the target drone exists in the drones is determined according to the comparison result.

[0022] Furthermore, the target recognition module is also used to track the image of the target drone, specifically including:

[0023] Observe the position of the target UAV in the i-th frame image and the i+1-th frame image using the KCF algorithm, where the i-th frame image and the i+1-th frame image belong to the multiple frame images;

[0024] Using an extended Kalman filter to predict the position of the target UAV in the (i+1)th frame image based on the position of the target UAV in the (i)th frame image;

[0025] The position of the target UAV in the i+1th frame image predicted by the extended Kalman filter is optimized to the position of the target UAV in the i+1th frame image observed by the KCF algorithm.

[0026] Furthermore, tracking the image of the target drone also includes:

[0027] When the position of the target UAV in the i+1th frame image is not observed using the KCF algorithm, the search range is expanded based on the position of the target UAV in the i+1th frame image predicted by the extended Kalman filter, and the target UAV is re-detected in the area after the expanded search range using the KCF algorithm.

[0028] Furthermore, the UAV-Detector algorithm is improved by introducing a shared dilated convolutional pyramid module and a hybrid deformable attention module into YOLOv8n.

[0029] Furthermore, the flight modes of the tilt-quadrotor UAV include a vertical take-off mode and a horizontal flight mode.

[0030] At least one technical solution employed by the present invention achieves the following beneficial effects: The present invention utilizes a tilt-rotor drone as the carrier of a drone countermeasure system. By slightly positioning tilt pods 4 on either side of the drone's front wings 2 and rear wings 3, these pods are synchronously tilted by a servo 6, enabling rapid switching of power direction. When the tilt pods 4 are tilted to a horizontal position, the drone switches to level flight mode, enabling high-speed cruising and quickly closing the distance to a target drone, preventing it from escaping from view. When the tilt pods 4 are tilted to a vertical position, the drone achieves vertical takeoff and landing (VTOL) and hovering capabilities, enabling precise tracking or attitude adjustment of target drones at close range, ensuring stable targeting of the target drone by the netting mechanism 7 before launching the net. The drone utilizes a design consisting of tandem wings (front wings 2 and rear wings 3) plus a single vertical tail 5. The tandem wings provide ample lift, which, in conjunction with the ailerons 21 and elevators 31, enables precise adjustment of the drone's pitch angle, reducing attitude fluctuations during flight. The single vertical tail 5 plus rudder 51 enhances longitudinal aerodynamic stability, ensuring heading stability in level flight mode and preventing target loss due to yaw error. This solution integrates the high-speed cruise capabilities of a fixed-wing aircraft with the vertical takeoff and landing and hovering capabilities of a multi-rotor aircraft. This ensures that the camera gimbal 8 located at the nose of the aircraft maintains a direct line of sight to the target in both level flight and vertical takeoff modes, thereby enabling stable acquisition of video stream information from the target drone. This provides the necessary conditions for subsequent accurate identification and stable tracking of the target drone. It also reduces the launch angle deviation of the net mechanism 7, improving the accuracy of countermeasures against target drones. The drone countermeasure system uses the UAV-Detector algorithm to detect the presence of the target drone in the video stream frame by frame, accurately extracting the target drone's features and achieving precise detection. If the target drone is present in the video stream, the drone countermeasure system determines the distance and speed difference between the target drone and the tilt-quadrotor drone, characterizing the target drone's spatial position relative to the tilt-quadrotor drone from both distance and relative speed perspectives. The drone countermeasure system also includes a flight control module, which is used to control the tilt-quadrotor drone to fly toward the target drone, and dynamically adjust the tilt direction of the tilt pod 4 according to the distance and speed difference during the flight, so as to change the lift direction of the tilt-quadrotor drone through the rotor 41 and the winglet 42, and control the flight mode of the tilt-quadrotor drone; and dynamically adjust the deflection direction of the elevator 31 to change the aerodynamic lift received by the rear wing 3 and control the pitch attitude of the tilt-quadrotor drone; and dynamically adjust the deflection direction of the rudder 51 to control the yaw angle of the tilt-quadrotor drone, thereby achieving accurate capture of the target drone. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0032] Figure 1 A schematic diagram of the tilt-rotor quad-rotor UAV provided by the present invention;

[0033] Figure 2 A framework diagram of a UAV countermeasure system based on a tilt-quadrotor UAV provided by the present invention;

[0034] Figure 3 This is an execution flow chart of a UAV countermeasure system based on a tilt-quadrotor UAV provided by the present invention;

[0035] Figure 4 A flowchart of a UAV countermeasure system based on a tilt-quadrotor UAV provided by the present invention for capturing a target UAV;

[0036] Figure numerals: 1. fuselage, 2. wing, 3. wing, 4. tilt pod, 5. single vertical tail, 6. tilt servo, 7. netting mechanism, 8. camera gimbal, 9. airborne equipment, 21. aileron, 31. elevator, 41. rotor, 42. winglet, 51. rudder. DETAILED DESCRIPTION

[0037] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0038] The server mentioned in the present invention can be a server installed on a business platform, or a device such as a desktop computer or laptop computer capable of executing the solution of the present invention. For ease of explanation, the following description will only use the server as the execution entity. The following, combined with the accompanying drawings, details the technical solutions provided by various embodiments of the present invention.

[0039] refer to Figure 1 The present invention provides a tilt-rotating quad-rotor UAV, specifically comprising:

[0040] A fuselage 1 , a front wing 2 and a rear wing 3 arranged on the fuselage 1 , a single vertical tail 5 arranged at the tail of the fuselage 1 , and a tilting servo 6 arranged inside the fuselage 1 . The single vertical tail 5 has a rudder 51 .

[0041] Ailerons 21 are provided on both sides of the front wing 2, elevators 31 are provided on both sides of the rear wing 3, and both side wings of the front wing 2 and the rear wing 3 are provided with tilt pods 4, which are tilted by tilt servos 6.

[0042] The tilt pod 4 is provided with a rotor 41 and a winglet 42 , and the rotor 41 and the winglet 42 tilt synchronously with the tilt pod 4 .

[0043] The head of the tilt-rotating quad-rotor drone is provided with a net-catching mechanism 7 and a camera gimbal 8 , and the fuselage 1 is provided with an onboard device 9 .

[0044] Specifically, the tilt-quadrotor UAV platform of this embodiment adopts a "tandem wing plus four tilt-rotors" layout. A pair of wings 2 and 3 are located at the front and rear of the fuselage 1, respectively. The two pairs of wings have essentially the same structural layout. Ailerons 21 and elevators 31 are located at the rear ends of the front and rear wings 2 and 3, respectively. A tilt pod 4 is located at the wing tips, and rotors 41 and winglets 42 are mounted on the tilt pod 4, both of which tilt integrally with the tilt pod. A single vertical stabilizer 5 with a rudder 51 is located at the rear of the fuselage 1. The tilt pod 4 is tilted by a single, dual-headed tilt servo 6 located within the fuselage 1. This tilt servo 6 can synchronously tilt both pods via a single drive shaft. A netting mechanism 7 and a camera gimbal 8 are located at the nose of the aircraft, ensuring that the net and camera remain facing the target drone during tracking, making it easier to track and capture the drone. A battery and onboard equipment 9 are located within the fuselage 1.

[0045] refer to Figure 2 The present invention provides a UAV countermeasure system based on a tilt-rotor quad-rotor UAV, which is mounted on the tilt-rotor quad-rotor UAV. The system specifically includes:

[0046] The target recognition module is used to collect the video stream information of the target airspace in real time through the camera gimbal 8, and detect whether there is a target drone in the video stream information frame by frame through the UAV-Detector algorithm.

[0047] The target tracking module is used to determine the distance and speed difference between the target drone and the tilt-quadrotor drone when the target drone exists in the video stream information.

[0048] The flight control module is used to control the tilt-quadrotor UAV to fly toward the target UAV, and dynamically adjust the tilt direction of the tilt pod 4 according to the distance and speed difference between the target UAV and the tilt-quadrotor UAV during flight, so as to change the lift direction of the tilt-quadrotor UAV through the rotor 41 and the winglet 42, and control the flight mode of the tilt-quadrotor UAV to be in the vertical take-off mode or the level flight mode; and dynamically adjust the deflection direction of the elevator 31 to change the aerodynamic lift of the rear wing 3 and control the pitch attitude of the tilt-quadrotor UAV; and dynamically adjust the deflection direction of the rudder 51 to control the yaw angle of the tilt-quadrotor UAV.

[0049] The net gun capture module is used to calculate the launch parameters of the capture net in real time according to the distance and speed difference when the distance between the target UAV and the tilt-rotor quadrotor UAV is less than the preset distance, and control the capture mechanism 7 to launch the capture net toward the target UAV according to the launch parameters.

[0050] In this embodiment, the flight control system controls the tilt-quadrotor drone's rotors and tilt mechanism based on the distance and speed difference between the target drone and the tilt-quadrotor drone, thereby adjusting the tilt-quadrotor drone's flight mode and speed in real time. The target drone refers to the drone that needs to be countered, and both the distance and speed difference are vector quantities. The distance comprises the distance between the target drone and the tilt-quadrotor drone in three mutually perpendicular directions in three-dimensional coordinate space (i.e., front-to-back, up-down, and left-to-right), and the speed difference comprises the speed difference between the target drone and the tilt-quadrotor drone in these three directions.

[0051] In this embodiment, dynamically adjusting the tilt direction of the tilt pod 4 according to the distance and speed difference between the target UAV and the tilt quadrotor UAV includes:

[0052] When the distance to the target drone is far (e.g. >25 meters) and the distance needs to be shortened quickly, the tilt-quadrotor drone switches to level flight mode, the tilt pod tilts to a horizontal state, and uses the high-speed cruise characteristics of the fixed wing to increase the flight speed and quickly approach the target; when the tilt-quadrotor drone is close to the target drone (e.g. 10-25 meters) or needs to hover and adjust, it switches to vertical take-off mode, the tilt pod provides lift vertically, achieves hovering or slow movement, and ensures tracking stability.

[0053] In this embodiment, the deflection direction of the elevator 31 is dynamically adjusted to change the aerodynamic lift acting on the rear wing 3 and control the pitch attitude of the tilt-quadrotor drone, including:

[0054] When the target drone is far away and needs to quickly lower its altitude, the flight control system controls the elevator 31 to deflect downward, causing the drone to dive and shorten the vertical distance. Once close to the target, the elevator 31 is controlled to deflect upward, adjusting the pitch angle to less than 5° (attitude stability) to ensure angular accuracy when launching the net.

[0055] If there is a vertical speed difference between the tilt-rotor quadcopter and the target drone (such as ascending or descending), the quadcopter can be raised and lowered synchronously by adjusting the pitch angle in real time to maintain relative height stability.

[0056] In this embodiment, the deflection direction of the rudder 51 is dynamically adjusted to control the yaw angle of the tilt-quadrotor drone, including:

[0057] According to the horizontal distance and speed difference between the target UAV and the tilt-quadrotor UAV, the flight control system controls the UAV's yaw angle by adjusting the rudder 51 so that the nose of the UAV always points to the target UAV.

[0058] When the target drone turns, the rudder deflects synchronously to ensure that the camera gimbal and the net mechanism continue to aim at the target.

[0059] In this embodiment, a flight path is planned to approach the target drone based on its location and motion information. An onboard computer 9 calculates the target drone's relative position, velocity difference, and other parameters in real time. Combined with the net gun's launch characteristics, the optimal launch timing and angle are determined. Once launch conditions are met, the onboard computer sends a launch command to the net gun capture module. Upon receiving the command, the net gun capture system rapidly activates the capture mechanism 7 to launch the capture net toward the target drone. After the capture net is launched, the drone continuously monitors the status of the target and the capture net. If capture is successful, the drone detects a feedback signal confirming capture and immediately initiates a handling procedure, slowing down, adjusting its attitude, and entering hover mode to await instructions from the ground control station. If capture fails, the target is repositioned based on tracking data and the predicted target position, preparing for the next capture attempt.

[0060] Specifically, the UAV-Detector algorithm is used to detect whether the target drone exists in the video stream information frame by frame, including:

[0061] Extract multiple frames of target spatial domain images from the video stream information according to the preset frame interval.

[0062] The UAV-Detector algorithm is used to detect multiple frames of images frame by frame, and the presence of the target drone in the video stream information is determined based on the detection results.

[0063] In this embodiment, the target recognition module first uses the gimbal-mounted camera visual detection equipment to collect real-time video stream information of the target airspace. It then extracts multiple frames of images of the target airspace from the video stream at a preset frame interval. The target airspace refers to the spatial area where the target drone may be located, which is pre-determined by the operator. The preset frame interval is a positive number, pre-set by the operator, at which an image is extracted every certain number of frames from the video stream. For example, if the preset frame interval is 3, an image is extracted every three frames from the video stream.

[0064] Optionally, when a certain UAV appears in 60 consecutive frames of images, the UAV is determined as a target UAV.

[0065] Optionally, in this embodiment, when tracking a target drone, the frame interval can be dynamically adjusted based on the target drone's movement within the image. For example, if the target drone's position in N consecutive frames has not changed significantly, the frame interval can be adaptively increased. This can reduce power consumption while ensuring accurate tracking of the target drone, thereby increasing the operating time of the drone countermeasure system.

[0066] In this embodiment, the UAV-Detector algorithm is improved based on YOLOv8n. This algorithm introduces a shared dilated convolution pyramid module (SDCPM) and a hybrid deformable attention block (HDAB). First, it extracts features from the input image. The shared dilated convolution pyramid module in the backbone network is used to capture multi-scale information. The hybrid deformable attention block enhances attention to complex scenes and small objects, thereby improving the accuracy and speed of drone detection. The improvements are as follows:

[0067] ① Shared Dilated Convolutional Pyramid Module (SDCPM): This module captures multi-scale information through convolutional layers with different dilation rates (1, 3, and 5). Low-dilation convolution (with a dilation rate of 1) captures local details, while high-dilation convolution (with dilation rates of 3 and 5) captures global context. These features are fused through 1x1 convolutional layers to meet the requirements of multi-scale feature extraction.

[0068] ② Hybrid Deformable Attention Module (HDAB): It consists of a Bottleneck module and a Transformer module. The Bottleneck module compresses features and extracts important information through a deformable convolutional layer. The Transformer module calculates the similarity between image positions through a self-attention mechanism, enhancing attention to complex scenes and small objects.

[0069] Specifically, all the UAVs in the video stream information are detected frame by frame by the UAV-Detector algorithm, and whether the target UAV exists in the video stream information is determined according to the detection result, specifically including:

[0070] All the UAVs in the video stream information are detected frame by frame by the UAV-Detector algorithm.

[0071] Each of all the UAVs is compared with the UAVs in the pre-constructed target UAV library, and whether the target UAV exists in all the UAVs is determined according to the comparison result.

[0072] In this embodiment, a plurality of types of UAV images that need to be captured are pre-acquired, a target UAV library is constructed, and the UAV-Detector algorithm is used to learn the images in the target UAV library. After the UAV-Detector algorithm learns a plurality of types of target UAVs, all the UAVs in the video stream information are detected frame by frame, and each of the recognized UAVs is compared with the UAVs in the target UAV library. If the comparison result is that one or more UAVs in the video stream information are consistent with the UAVs in the target UAV library, the UAV is determined as the target UAV that needs to be captured.

[0073] Optionally, after the target UAV is recognized, the information of the acquired target UAV is converted into a topic and published to the information stream based on the Ros operating system. And the target position data is calculated in real time by the laser ranging unit in the target recognition module, the target distance d is calculated, and the distance information is published to the Ros information stream. The UAV flight control system plans the navigation task according to the received target topic, and adjusts the motion state of the UAV in real time.

[0074] Based on Figure 1The illustrated drone countermeasure system utilizes a tilt-rotor drone as the system's carrier. By slightly positioning tilt pods 4 on either side of the front wing 2 and rear wing 3, when the tilt pods 4 are tilted to a horizontal position, the drone switches to level flight mode, enabling high-speed cruising. This allows the drone to quickly reach a target area or shorten the distance to a target drone, enabling rapid tracking with minimal energy consumption and preventing the target drone from escaping. When the tilt pods 4 are tilted to a vertical position, the drone exhibits the vertical take-off and landing (VTOL) and hovering capabilities of a multi-rotor, enabling precise tracking and attitude adjustment at close range. This design is particularly suitable for stable alignment before launching a net. The present invention utilizes a stable design consisting of tandem wings (front wing 2 and rear wing 3) plus a single vertical tail 5. The tandem wings provide ample lift, which, in conjunction with the ailerons 21 and elevators 31, enables precise adjustment of the drone's pitch and roll angles, minimizing attitude fluctuations during flight. The single vertical tail 5 plus a rudder 51 enhance longitudinal aerodynamic stability, ensuring heading stability in level flight mode and preventing target loss due to yaw errors. Four tilting pods 4 (two at each wingtip) are synchronously tilted by a servo 6, enabling rapid switching of power direction. By installing a netting mechanism 7 and a camera gimbal 8 on the drone's nose, the nose camera can be stably pointed at the target drone, improving the accuracy of identifying and tracking the target drone. Furthermore, the netting mechanism 7 can maintain a straight-on orientation in both level flight and vertical takeoff modes, reducing launch angle deviation. Through this approach, the present invention leverages its unique tilt-rotor design, combining the high-speed cruising capabilities of a fixed-wing aircraft with the vertical takeoff and landing and hovering capabilities of a multi-rotor aircraft. This ensures that the camera gimbal 8, located on the nose, is stably pointed at the target drone's heading, enabling stable capture of the target drone's video stream. This provides the necessary conditions for subsequent accurate identification, stable tracking, and precise capture of the target drone, thereby improving the accuracy of countermeasures against the target drone. By integrating a drone countermeasure system on a tilt-rotor quadrotor drone, the maneuverability of a multi-rotor drone and the high speed and stability of a fixed-wing drone can be combined to enable rapid approach to the target drone. This drone countermeasure system collects real-time video stream information from the target airspace and uses the UAV-Detector algorithm to detect the presence of the target drone frame by frame in the video stream. This system accurately extracts the target drone's features, enabling precise detection and laying a solid foundation for subsequent, precise tracking of the target drone. If the target drone is present in the video stream, the drone countermeasure system tracks the target drone and calculates the distance and speed difference between the target drone and the tilt-quadrotor drone. This system characterizes the target drone's spatial position relative to the tilt-quadrotor drone from both the distance and relative speed perspectives. The drone countermeasure system also includes a flight control module, which controls the tilt-quadrotor drone's flight toward the target drone. During flight, the tilt-quadrotor drone's flight mode and speed are adjusted in real time based on the distance and speed difference, ensuring efficient and real-time tracking of the target drone.The drone countermeasure system also includes a net gun capture module. When the distance between the target drone and the tilt-rotor quadrotor drone is less than the preset distance, the launch parameters of the capture net are calculated in real time based on the distance and speed difference, and the capture net is launched at the target drone based on the launch parameters, thereby achieving accurate capture of the target drone.

[0075] In addition, in one or more embodiments of the present invention, tracking the image of the target drone specifically includes:

[0076] The KCF algorithm is used to observe the position of the target UAV in the i-th frame image and the i+1-th frame image. The i-th frame image and the i+1-th frame image belong to multiple frame images.

[0077] The extended Kalman filter is used to predict the position of the target UAV in the i+1th frame image based on the position of the target UAV in the i-th frame image.

[0078] The position of the target UAV in the i+1 frame image predicted by the extended Kalman filter is optimized to the position of the target UAV in the i+1 frame image observed by the KCF algorithm.

[0079] In this embodiment, after confirming the target drone, the drone starts the KCF (Kernelized Correlation Filter) target tracking algorithm combined with the extended Kalman filter to track the target. The principle of the KCF algorithm based on motion prediction is as follows:

[0080] The motion direction of the next frame is predicted based on the motion trends of the current and previous frames, thereby determining the position and reducing tracking errors. A scale-adaptive approach is also implemented, using both large-scale and small-scale detections, selecting the larger size as the new tracking frame size. Furthermore, to prevent target loss, the target is re-detected and its position adjusted at regular intervals, enhancing tracking stability and continuity.

[0081] Specifically, the KCF algorithm saves the identified target drone image and circularly shifts it within the image in the video stream. When the saved target drone image overlaps with the target drone image in the video stream, the target drone's position is determined. During tracking, the extended Kalman filter is used to predict the target's position in the next frame. The KCF algorithm's output is then compared with the extended Kalman filter's prediction, correcting KCF algorithm errors and reducing drift. It can also handle changes in target scale and motion patterns.

[0082] In addition, in one or more embodiments of the present invention, tracking the image of the target drone further includes:

[0083] When the position of the target UAV in the i+1 frame image is not observed using the KCF algorithm, the search range is expanded based on the position of the target UAV in the i+1 frame image predicted by the extended Kalman filter, and the KCF algorithm is used to re-detect the target UAV in the area after the expanded search range.

[0084] In this embodiment, to prevent the target drone from being lost during tracking, the KCF algorithm periodically re-detects. If the target is detected to be lost, the extended Kalman filter algorithm is used to predict the target position, expand the search range and search again. If the target is re-detected, tracking is resumed. Otherwise, the search is carried out in a circling manner according to the preset strategy. The specific process is as follows:

[0085] After the KCF algorithm is initialized, the Kalman filter predicts the next frame's position based on the target's uniform motion model. This prediction is combined with the KCF algorithm's output in a weighted fusion with a 7:3 ratio to correct for tracking frame drift. If the target is lost, the search area radius is expanded by 50% based on the position predicted by the extended Kalman filter, and local re-detection is performed every five frames.

[0086] Through the above solution, even if the video stream of the target drone is temporarily lost, the extended Kalman filter can be used to predict the position of the target drone and continue to counterattack.

[0087] refer to Figure 3 ,The execution process of the drone countermeasure system specifically includes:

[0088] Step S1, Target Detection and Identification: The system starts, and the drone's sensors scan the airspace, capturing images and transmitting them to the onboard computer. The UAV-Detector algorithm determines whether there are threatening targets. If so, the information is uploaded, the target is marked, and tracking begins. If not, scanning continues.

[0089] Specifically, after vertical takeoff, when the drone reaches a preset altitude, the tilt pods located at the tips of wings 2 and 3 begin to tilt, putting the drone into level flight cruise mode. A gimbal located at the nose of the aircraft maintains stability, scanning the airspace at 60 frames per second. The image data is compressed by the onboard computer and transmitted to the ground station. The UAV-Detector detects each frame and marks it as a "threat target" if the confidence level exceeds 0.9 and matches a feature library. The threat level is automatically determined based on target distance, speed, and trajectory prediction, with high-threat targets triggering tracking commands.

[0090] Step S2, Tracking and Locking: After the target is identified, the KCF algorithm initializes and begins tracking, predicting target motion in real time and adaptively adjusting the scale. Regular detection is performed to prevent target loss; if the target is lost, the Kalman filter is used for prediction and a new search is performed.

[0091] Specifically, the drone uses the KCF target tracking algorithm to track the target. After initialization, the KCF algorithm updates the target position each frame, while the Kalman filter predicts the coordinates for the next frame. If the target is detected for 60 consecutive frames, tracking is considered stable. If the target is lost, a spiral search is initiated based on the Kalman filter's predicted position, with the radius increasing by 5% each frame. If the search times out (no relock within 30 seconds), the drone returns to its pre-set patrol route.

[0092] Step S3, Capture Preparation and Launch: When tracking a target, the drone automatically adjusts its flight state to approach the target drone. The onboard computer calculates launch parameters and determines whether launch conditions (such as distance and speed difference) are met, ensuring that the target drone is within the capture range of the net.

[0093] Specifically, during tracking, the drone plans a flight path based on the target's position and motion information, approaching the target in level flight mode. A velocity matching algorithm ensures a relative velocity difference of ≤5 m / s. A laser rangefinder provides real-time feedback on the target's distance, and the onboard computer calculates launch parameters. When the target enters range (10-25 meters) and its attitude is stable (pitch angle <5°), the capture net mechanism 7 located at the head of the fuselage 1 is activated to launch the capture net. 0.5 seconds before launch, the drone fine-tunes its attitude to ensure the capture net covers the target's rotor area.

[0094] Step S4, Capture and Disposition: After launching the capture net, the drone determines whether the target has been successfully captured. If so, the drone slows down, adjusts its attitude, and hovers, awaiting instructions from the ground control station. If capture fails, the drone repositions the target and returns to the tracking and locking process to try again.

[0095] Specifically, refer to Figure 4 After identifying the target (target drone), the tilt-quadrotor drone determines whether it has locked onto the target. If so, it sends the target drone's position and speed instructions to the flight control module. If not, it predicts the target position based on the extended Kalman filter and returns to the target identification step.

[0096] After the capture net is launched, the drone continuously monitors the target and the net's status. A camera monitors the net's deployment status and uses an image segmentation algorithm to determine if capture is successful. If the net is fully deployed and envelops the target, the sensor signals "capture successful" and initiates a handling procedure, slowing down, adjusting its attitude, and entering hover mode to await instructions from the ground control station. If capture fails, such as if the net is not deployed or the target escapes, the system automatically records the failure location and re-enters the tracking process.

[0097] In step S3, the image processing module performs image information processing on the video images captured by the camera. After image processing, the acquired target information is converted into a topic and published to the information stream based on the Ros operating system. The laser ranging module receives target position data in real time, calculates the target distance d, and publishes this distance information to the Ros information stream to support subsequent trajectory calculation. The drone flight module plans navigation missions based on the received target topic and adjusts the drone's motion state in real time to ensure accurate target tracking.

[0098] Throughout the entire process, the drone maintains constant communication with the ground control station. Information such as drone status, target identification, tracking status, target-drone distance, and capture results are uploaded to the ground control station in real time, allowing ground control personnel to issue appropriate instructions based on the actual situation.

[0099] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A UAV countermeasure system for a tilt-rotor quad-rotor UAV, characterized in that: The tilt-rotating quadrotor drone comprises: A fuselage (1), a front wing (2) and a rear wing (3) arranged on the fuselage (1), a single vertical tail (5) arranged at the tail of the fuselage (1), and a tilting servo (6) arranged inside the fuselage (1), wherein the single vertical tail (5) has a rudder (51); Ailerons (21) are provided on both sides of the front wing (2), elevators (31) are provided on both sides of the rear wing (3), and both wing tips of the front wing (2) and the rear wing (3) are provided with tilting pods (4), and the tilting pods (4) are tilted by the tilting servo (6); The tilting pod (4) is provided with a rotor (41) and a winglet (42), and the rotor (41) and the winglet (42) are tilted synchronously with the tilting of the tilting pod (4); The head of the tilt-rotating quad-rotor UAV is provided with a netting mechanism (7) and a camera platform (8), and the fuselage (1) is provided with an onboard device (9); The system is mounted on the tilt-quadrotor UAV and includes: A target recognition module is used to collect video stream information of the target airspace in real time through the camera platform (8), and detect whether there is a target drone in the video stream information frame by frame through the UAV-Detector algorithm; a target tracking module, configured to determine the distance and speed difference between the target drone and the tilt-quadrotor drone when a target drone is present in the video stream information; A flight control module is used to control the tilt-quadrotor UAV to fly toward the target UAV, and dynamically adjust the tilt direction of the tilt pod (4) according to the distance and speed difference between the target UAV and the tilt-quadrotor UAV during the flight, so as to change the lift direction of the tilt-quadrotor UAV through the rotor (41) and the winglet (42) to control the flight mode of the tilt-quadrotor UAV; and dynamically adjust the deflection direction of the elevator (31) to change the aerodynamic lift of the rear wing (3) to control the pitch attitude of the tilt-quadrotor UAV; and dynamically adjust the deflection direction of the rudder (51) to control the yaw angle of the tilt-quadrotor UAV; The net gun capture module is used to calculate the launch parameters of the capture net in real time according to the distance and speed difference when the distance between the target UAV and the tilt-rotor quad-rotor UAV is less than a preset distance, and control the capture net mechanism (7) to launch the capture net toward the target UAV according to the launch parameters.

2. The UAV countermeasure system according to claim 1, characterized in that: The UAV-Detector algorithm is used to detect whether the target UAV exists in the video stream information frame by frame, specifically including: Extracting multiple frames of images of the target airspace from the video stream information according to a preset frame interval; The UAV-Detector algorithm is used to detect the multiple frames of images frame by frame, and it is determined whether the target drone exists in the video stream information based on the detection results.

3. The UAV countermeasure system according to claim 2, characterized in that: Detecting the multiple frames of images frame by frame using the UAV-Detector algorithm, and determining whether the target UAV exists in the video stream information based on the detection results, specifically includes: Detect all drones in the video stream frame by frame using the UAV-Detector algorithm; Each of the drones is compared with the drones in the pre-built target drone library, and whether the target drone exists in the drones is determined according to the comparison result.

4. The UAV countermeasure system according to claim 1, wherein: The target recognition module is also used to track the image of the target drone, specifically including: Observe the position of the target UAV in the i-th frame image and the i+1-th frame image using the KCF algorithm, where the i-th frame image and the i+1-th frame image belong to multiple frame images; Using an extended Kalman filter to predict the position of the target UAV in the (i+1)th frame image based on the position of the target UAV in the (i)th frame image; The position of the target UAV in the i+1th frame image predicted by the extended Kalman filter is optimized to the position of the target UAV in the i+1th frame image observed by the KCF algorithm.

5. The UAV countermeasure system according to claim 4, characterized in that: Tracking the image of the target drone also includes: When the position of the target UAV in the i+1th frame image is not observed using the KCF algorithm, the search range is expanded based on the position of the target UAV in the i+1th frame image predicted by the extended Kalman filter, and the target UAV is re-detected in the area after the expanded search range using the KCF algorithm.

6. The UAV countermeasure system according to claim 1, wherein: The UAV-Detector algorithm is improved by introducing the shared dilated convolutional pyramid module and the hybrid deformable attention module into YOLOv8n.

7. The UAV countermeasure system according to claim 1, wherein: The flight modes of the tilt-quadrotor UAV include a vertical take-off mode and a horizontal flight mode.

Images