Remote mother-daughter unmanned aerial vehicle combat system and control method

Abstract

This invention discloses a long-range mother-daughter unmanned aerial vehicle (UAV) combat system and control method. The system includes a mother aircraft, several loitering munitions, and a distributed AI decision-making system. It is a UAV swarm combat system with ultra-long range, strong anti-jamming capability, highly intelligent autonomous collaboration, and controllable cost. It breaks through the technical bottlenecks of existing military UAV combat systems, such as "difficulty in balancing long-range strike and anti-jamming, reliance on external communication for swarm collaboration, and prominent contradiction between high performance and high cost". Through the collaborative innovation of three-mode fusion navigation, self-organizing network, distributed AI decision-making, and modular design, it achieves the organic unity of 3,000-kilometer-level long-range combat, high-reliability navigation in GPS-denied environments, autonomous collaborative strike in communication-free environments, and low-cost large-scale application.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) technology, specifically relating to a remote mother-daughter UAV combat system and control method. Background Technology

[0002] In recent years, unmanned aerial vehicles (UAVs) and loitering munitions have played an increasingly important role in modern military reconnaissance and strike operations. Their flexibility, low cost, and ability to avoid personnel casualties make them ideal platforms for high-risk missions. However, existing military UAVs and loitering munitions, especially when attempting to achieve long-range, intelligent, and swarm-based operations, still face technical bottlenecks in practical applications, specifically in the following aspects: 1. Limited operational radius: The range of existing loitering munition carriers is generally limited to within 1,000 kilometers, making it difficult to effectively strike strategic targets deep behind enemy lines. To achieve operational effectiveness, they need to rely on forward base deployment and resupply, which not only increases the complexity and time cost of operational preparation, but also easily exposes one's own operational intentions.

[0003] 2. Weak anti-jamming capability: Existing systems heavily rely on global satellite navigation systems (such as GPS and BeiDou) for positioning and navigation. In environments where the enemy implements strong electromagnetic interference or signal jamming to deny GPS access, the navigation signal loss rate exceeds 60%, leading to a sharp decline in navigation accuracy or even complete failure. This prevents the UAV platform and its released loitering munitions from accurately reaching the target area, causing mission interruptions, strike deviations, or munition loss, severely limiting its operational effectiveness in complex electromagnetic environments.

[0004] 3. Poor Cluster Coordination: Traditional cluster systems mostly employ a centralized control architecture, heavily relying on continuous and stable ground control stations or satellite communication links to issue commands and conduct coordinated control. This communication mode has inherent drawbacks such as high latency, limited bandwidth, and fragile links. Once the communication link is intercepted, interfered with, or interrupted by the enemy, the entire cluster system will fall into a state of disorder, unable to perform autonomous coordinated reconnaissance, target allocation, and tactical coordination tasks.

[0005] 4. Excessive Cost: Currently, similar combat systems with long-range strike and swarm capabilities are extremely complex in technology and expensive to manufacture, with unit prices typically exceeding 20 million RMB. This high cost significantly limits the scale of military procurement and the quantity of equipment, making it difficult to form effective large-scale, saturation combat capabilities and failing to meet the operational needs of future high-intensity, attrition-based conflicts.

[0006] Therefore, achieving long-range strike capability at a range of 3,000 kilometers, enhancing anti-jamming capabilities in GPS-denied environments, ensuring autonomous collaborative decision-making in communication-free environments, and reducing system costs have become the core technical challenges that existing military unmanned aerial vehicle (UAV) combat systems urgently need to address. Summary of the Invention

[0007] To address the aforementioned problems, the main objective of this invention is to design a long-range mother-daughter unmanned aerial vehicle (UAV) combat system and control method. This system employs a mother aircraft, several loitering munitions, and a distributed AI decision-making system. By optimizing the mother aircraft's power system and onboard satellite relay signals, as well as employing anti-jamming multi-mode navigation methods, self-organizing network communication methods for the loitering munitions, and swarm decision-making and coordinated strike methods for the loitering munitions, the system solves four core challenges: long-range delivery capability of UAVs, reliable navigation in complex electromagnetic environments, autonomous swarm coordination without external communication support, and system cost control.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: A long-range mother-daughter unmanned aerial vehicle (UAV) combat system, wherein the mother aircraft includes: The main body integrates the power system, fuel storage unit, AI decision-making center and navigation module installation compartment; Fixed wings with an internal honeycomb composite structure; The ventral bomb bay has an internal pylon equipped with a mechanical locking mechanism for mounting and securing loitering munitions. Satellite communication antenna array, used for remote data exchange with the rear command center; The navigation module, located inside the navigation module installation compartment, is used to collect terrain images, target outlines and distance data to provide the mother machine with positioning, navigation and attitude references; The three-light pod integrates infrared, visible light, and laser rangefinders to acquire real-time thermal distribution, visual images, distance data, and terrain reconnaissance data of ground targets. The loitering munition includes: The folding wings are folded when mounted on the mothership and automatically unfold when released. Self-organizing network communication unit is used for real-time data interaction between multiple loitering munitions to build a loitering munition cluster communication network; The warhead contains 5 kg of high-energy explosive and a precision fuse with a trigger mode that can be set according to the target type; The distributed AI decision-making system: Deployed inside the mothership, it receives navigation and reconnaissance data transmitted back by the navigation module, the three-light pod module, and the loitering munitions, performs real-time analysis and decision-making, and sends instructions to the loitering munition cluster.

[0009] As a further description of the present invention, the navigation module of the mother machine is configured as an anti-interference multi-mode navigation system, including: a satellite positioning module, a fiber optic inertial navigation module, a visual SLAM camera group and an AI navigation decision center; The satellite positioning module, fiber optic inertial navigation module, and visual SLAM camera group are all connected to the AI ​​navigation decision center via a data bus, transmitting positioning data and status information to the AI ​​navigation decision center in real time.

[0010] As a further description of the present invention, the self-organizing network communication unit of the loitering munition is configured to adopt a simplified OSI model protocol stack, wherein: The application layer is used to encapsulate target allocation instructions; At the network layer, the AODVv2 self-organizing network routing protocol is used for dynamic route selection; At the data link layer, TDMA or OFDMA time slot allocation methods are used to allocate independent communication time slots to each node of different loitering munitions. The physical layer is equipped with a frequency hopping spread spectrum module to avoid signal interference.

[0011] A remote mother-daughter unmanned aerial vehicle (UAV) control method, based on the aforementioned combat system and employing an anti-jamming multi-mode navigation method, includes the following steps: Step S11: The AI ​​navigation decision center monitors the signal-to-noise ratio of satellite signals in real time; Step S12: Switch the navigation mode according to the range of signal-to-noise ratio values: When the signal-to-noise ratio is greater than the first threshold, the satellite-dominated mode is adopted; When the second threshold < signal-to-noise ratio ≤ the first threshold, the inertial navigation-satellite fusion mode is adopted; When the signal-to-noise ratio is less than or equal to the second threshold or the signal is lost, the visual-inertial navigation emergency mode is activated. Step S13: Based on the judgment result, select and activate the corresponding navigation mode; Step S14: In any navigation mode, the AI ​​navigation decision center continuously compares the actual flight trajectory with the planned path. If the deviation exceeds the threshold, the path correction algorithm is automatically activated to adjust the flight parameters.

[0012] As a further description of the present invention, the first threshold is 15dB, the second threshold is 10dB, and the deviation threshold is ±50m. The visual-inertial navigation emergency mode is based on odometry calculation using visual SLAM and inertial navigation data fusion, supplemented by laser rangefinder for altitude correction and visual terrain matching.

[0013] As a further description of the present invention, the self-organizing network communication method of the loitering munition includes: Step S21: At the application layer, the instructions are packaged into standard data frames; Step S22: At the network layer, select the communication path according to the AODVv2 protocol; Step S23: At the data link layer, allocate independent communication time slots to each node; Step S24: At the physical layer, frequency hopping spread spectrum is performed to switch communication frequencies.

[0014] As a further description of the present invention, the loitering munitions swarm decision-making and coordinated strike includes the following steps: Step S31: Reconnaissance phase. After the loitering munition cluster is released, it automatically forms a mesh topology, collects ground target data through the three-light pod, shares reconnaissance data in real time, and dynamically allocates reconnaissance areas based on the improved auction algorithm. Step S32: Target recognition stage. A deep learning-based multi-source data fusion target recognition model is used to fuse infrared and visible light images and point cloud data transmitted back by the three-light pod and loitering munition, and output target category and location information. Step S33: Strike phase. The cluster is divided into multiple sub-clusters according to the target distribution, a cluster head node is elected, and a coordinated strike strategy is formulated. Strike commands are sent to each loitering munition through the ad hoc network communication unit.

[0015] As a further description of the present invention, the target recognition model adopts a lightweight YOLOv7 architecture as its target detection head. The architecture includes: a feature alignment module, an attention weight allocator, and a dual-branch feature extraction network based on ResNet34 and PointNet++. The YOLOv7 detection head outputs the recognition results, including target category and location information. The confidence threshold for the target recognition model was set to 0.89.

[0016] As a further description of the present invention, the cluster-coordinated attack timing includes: The mothership releases a cluster of loitering munitions; Some loitering munitions are used as decoys and jammers. Some loitering munitions perform ultra-low-altitude penetration and strike missions; Some loitering munitions perform target identification, tracking, and guidance correction tasks; To carry out the final strike and transmit the strike effectiveness assessment data back.

[0017] More specifically: The mothership releases loitering munitions, including reconnaissance munitions, decoy munitions, and strike munitions; The decoy flares began their dive and prepared to release jamming. The strike missile then initiated an ultra-low-altitude penetration maneuver; After the reconnaissance missile identifies and confirms the target, it shares the information with the cluster. The distributed AI decision-making system plans coordinated paths for reconnaissance missiles, decoy missiles, and strike missiles based on target information; Decoy flares are used for interference; The reconnaissance missile performs terminal tracking and provides guidance correction for the strike missile; The strike missiles will carry out the final strike; The reconnaissance missile transmits data on the effectiveness of the strike.

[0018] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described control method.

[0019] Compared with the prior art, the technical advantages of the present invention are as follows: This invention provides a long-range mother-daughter unmanned aerial vehicle (UAV) combat system and control method. The system includes a mother aircraft, several loitering munitions, and a distributed AI decision-making system. It is a UAV swarm combat system with ultra-long range, strong anti-jamming capability, highly intelligent autonomous collaboration, and controllable cost. It breaks through the technical bottlenecks of existing military UAV combat systems, such as "difficulty in balancing long-range strike and anti-jamming, reliance on external communication for swarm collaboration, and prominent contradiction between high performance and high cost". Through the collaborative innovation of three-mode fusion navigation, self-organizing network, distributed AI decision-making, and modular design, it achieves the organic unity of 3,000-kilometer-level long-range combat, high-reliability navigation in GPS-denied environments, autonomous collaborative strike in communication-free environments, and low-cost large-scale application. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall combat system of the present invention; Figure 2 This is a front structural view of the mother machine of the present invention; Figure 3 This is a rear structural view of the mother machine of the present invention; Figure 4 This is a structural view of the loitering munition of the present invention; Figure 5 This is a schematic diagram of the anti-interference navigation workflow of the present invention; Figure 6 This is a diagram of the self-organizing network communication topology of the loitering munition of the present invention; Figure 7 This is a schematic diagram of the cluster decision-making and collaborative attack process of the present invention; Figure 8 This is a diagram of the AI ​​target recognition model architecture of the present invention; Figure 9 This is a timing diagram of the cluster-coordinated attack of the present invention.

[0021] In the diagram, 1. Main fuselage, 2. Power system, 3. Fuel storage unit, 4. AI decision-making center, 5. Navigation module installation bay, 6. Fixed wing, 7. Underbelly bomb bay, 8. pylon, 9. Mechanical locking mechanism, 10. Satellite communication antenna array, 11. Navigation module, 12. Three-light pod, 13. Folding wing, 14. Self-organizing network communication unit, 15. Warhead, A. Mother aircraft, B. Loitering munition. Detailed Implementation

[0022] The present invention will now be described in detail with reference to the accompanying drawings: In one embodiment of the present invention, a long-range mother-daughter unmanned aerial vehicle (UAV) combat system is disclosed, with reference to... Figure 1-4 As shown, this embodiment aims to provide a long-range mother-daughter unmanned aerial vehicle (UAV) combat system with anti-jamming swarm decision-making capability. The system is based on the coordinated operation of the mother UAV and the loitering munition swarm, and specifically includes a mother UAV A, several loitering munitions B, and a distributed AI decision-making system.

[0023] In this embodiment, the aforementioned mother machine A is as follows: Figure 2-3 As shown, it includes: Main fuselage 1: Wingspan 22m, constructed from high-strength, lightweight materials, internally integrating the power system 2, fuel storage unit 3, AI decision-making center 4, and navigation module installation bay 5; specifically, the power system 2 is located at the rear center of the main fuselage 1, employing a high-thrust turbofan engine, coupled with efficient fuel utilization technology, supporting a range of over 3000km. The fuel storage unit 3 is located on both sides of the fixed wing 6, and the AI ​​decision-making center 4 and navigation module installation bay 5 are located at the nose of the main fuselage 1.

[0024] Fixed Wing 6: Built-in honeycomb composite structure (honeycomb support unit) provides sufficient lift during flight. The honeycomb composite structure can improve the wing's impact resistance and enhance its structural stability.

[0025] Abdominal bomb bay 7: Built-in pylon 8, the pylon 8 is equipped with a mechanical locking mechanism 9 for mounting and fixing the loitering munitions; in the open state, it can display 30 loitering munition pylons. The pylon 8 is equipped with a mechanical locking mechanism 9 to ensure that the loitering munitions are stably fixed during the flight of the mother aircraft.

[0026] Satellite communication antenna array 10: used for remote data interaction with the rear command center; it is installed at the tail of the main fuselage 1, adopts multi-band reception, and can receive satellite relay signals to realize remote data interaction between the mother aircraft and the rear command center, ensuring the transmission of key commands and mission status feedback within a 3000km flight range.

[0027] Navigation Module 11: Installed inside the navigation module installation compartment 5, it provides positioning, navigation and attitude reference for the mother machine. It consists of a multi-lens assembly, including a high-resolution visible light camera, an infrared camera and a laser rangefinder, used to collect terrain images, target outlines and distance data, providing data support for visual navigation and target recognition.

[0028] The Tri-Light Pod 12 integrates infrared, visible light, and laser rangefinders, located at the head of the main body 1. It is used to acquire real-time thermal distribution, high-definition visual images and distance data, and terrain reconnaissance data of ground targets for target reconnaissance, identification and tracking, providing accurate target information for cluster strikes.

[0029] In this embodiment, the aforementioned loitering munition B is as follows: Figure 4 As shown, it includes: Loitering munition B: Each munition is 1.6m long and is mounted on the mothership via folding wing 13.

[0030] Folding Wing 13: When mounted on the mother aircraft, it is in a folded state and automatically unfolds after release; in the folded state, it can be tightly mounted on the pylon 8 in the bomb bay 7 under the mother aircraft. After release, the folding wing automatically unfolds to provide lift for flight.

[0031] Self-organizing network communication unit 14: used for real-time data interaction between multiple loitering munitions to build a loitering munition cluster communication network; located at the head of the loitering munition, it supports real-time data interaction between multiple munitions. It is not only used to transmit commands, but also to build a highly reliable and interference-resistant data backhaul backbone network, which is the key to cluster collaboration.

[0032] The warhead 15 is located at the rear of the self-organizing network communication unit 14. It contains 5kg of high-energy explosive and a precision fuse. The fuse can be set to trigger mode according to the target type (such as ground contact trigger or airburst trigger) to ensure strike power and accuracy.

[0033] In addition, the loitering munition B also includes, but is not limited to, a power unit, a seeker head, and control surfaces, and its power system is a single-motor tail thruster.

[0034] In this embodiment, the aforementioned distributed AI decision-making system: Deployed inside the mothership, it consists of high-performance computing chips and dedicated algorithm modules, including improved auction algorithms, target allocation algorithms, path planning algorithms, anti-jamming decision-making algorithms, and cooperative strike planning algorithms. It receives navigation and reconnaissance data transmitted from the navigation module, the three-light pod module, and the loitering munitions, performs real-time analysis and decision-making, and sends commands to the loitering munition cluster, including target allocation commands and strike path planning results.

[0035] As the "brain" of cluster operations, the distributed AI decision-making system integrates an improved auction algorithm, a lightweight YOLOv7 recognition model, and a collaborative strike planning algorithm. It enables dynamic allocation of reconnaissance areas, high-accuracy target identification, and the formulation of collaborative strike strategies without communication, making it the core of improving the efficiency of cluster operations.

[0036] In this embodiment, the navigation module of the mother machine is configured as an anti-interference multi-mode navigation system, including: a satellite positioning module, a fiber optic inertial navigation module, a visual SLAM camera group, and an AI navigation decision center; the satellite positioning module, the fiber optic inertial navigation module, and the visual SLAM camera group are all connected to the AI ​​navigation decision center through a data bus. The satellite positioning module, the fiber optic inertial navigation module, and the visual SLAM camera group send their own positioning data and status information to the AI ​​navigation decision center in real time through the data bus. The AI ​​navigation decision center receives and processes the received data.

[0037] The modules (satellite + inertial navigation + vision) of the aforementioned mother machine anti-interference multi-mode navigation system form the hardware foundation for anti-interference navigation. Specifically, it is an anti-interference three-mode navigation system. The satellite positioning module ensures accurate remote positioning, the fiber optic inertial navigation module provides continuous positioning when satellite signals are lost, and the visual SLAM camera group corrects inertial navigation drift through terrain matching. The three work together to ensure navigation reliability in complex electromagnetic environments.

[0038] It should be noted that the aforementioned AI navigation decision center is responsible for the mother machine's own navigation filtering, multi-source information fusion, and path planning; the AI ​​decision hub is an onboard computer responsible for managing the mother machine platform resources such as navigation, communication, and payload; and the distributed AI decision system is responsible for the macro-management and decision-making of cluster collaborative tasks.

[0039] In this embodiment, the loitering munition's self-organizing network communication unit is configured to use an OSI model protocol stack, where the functions of each layer are as follows: The application layer is used to encapsulate target allocation instructions; it packages target information, attack priorities, and other instructions generated by the AI ​​decision-making system into standard data frames.

[0040] At the network layer, the AODVv2 self-organizing network routing protocol is adopted to dynamically select the communication path between loitering missile nodes. When the communication of a node is interrupted, it automatically switches to the backup path to ensure the continuity of data transmission.

[0041] At the data link layer, TDMA or OFDMA time slot allocation methods are used to allocate independent communication time slots to each node of different loitering munitions, avoiding signal conflicts and improving communication efficiency.

[0042] The physical layer is equipped with a frequency hopping spread spectrum module to avoid signal interference; by quickly switching communication frequencies, it avoids enemy signal interference and reduces the probability of data interception.

[0043] The aforementioned loitering munition's self-organizing network communication unit supports autonomous coordination of the cluster without external communication. Its frequency hopping spread spectrum technology and AODVv2 routing protocol ensure communication anti-interference capability and data transmission continuity, making it a key communication carrier for cluster decision-making and coordinated strikes.

[0044] In another embodiment of the present invention, a remote mother-daughter unmanned aerial vehicle (UAV) control method is disclosed, such as... Figure 5-9 As shown, this control method is implemented based on the aforementioned combat system.

[0045] Specifically, such as Figure 5 As shown in this embodiment, the anti-interference multi-mode navigation method, which collects data through a satellite positioning module and continuously evaluates the signal-to-noise ratio (SNR) based on an AI navigation decision center, includes the following steps: Step S11: The AI ​​navigation decision center monitors the signal-to-noise ratio of satellite signals in real time; Step S12: Switch between the three navigation modes according to the range of signal-to-noise ratio values: When the signal-to-noise ratio is greater than the first threshold, the satellite-dominated mode is adopted, with the data from the satellite positioning module as the main navigation basis and the data from the fiber optic inertial navigation module as auxiliary correction to ensure navigation accuracy. When the second threshold < signal-to-noise ratio ≤ the first threshold, the inertial navigation-satellite fusion mode is adopted to fuse the data of the fiber optic inertial navigation module with the data of the satellite positioning module to reduce the impact of single signal fluctuations on navigation accuracy. When the signal-to-noise ratio is less than or equal to the second threshold or the signal is lost, the visual-inertial navigation emergency mode is activated to perform anti-interference navigation. Step S13: Based on the judgment result, select and activate the corresponding navigation mode; Step S14: In any navigation mode, the AI ​​navigation decision center continuously receives the actual flight trajectory fed back by the flight control system and compares it with the planned path. If the deviation exceeds the threshold, the path correction algorithm is automatically activated to adjust the flight parameters and ensure that the UAV flies along the optimal path.

[0046] It should be noted that in this embodiment, the first threshold is 15dB, the second threshold is 10dB, and the deviation threshold is ±50m.

[0047] The visual-inertial navigation (VIS) emergency mode is based on odometry calculation using data fusion from a visual SLAM camera group and a fiber optic inertial navigation module, supplemented by altitude correction and visual terrain matching using a laser rangefinder. Specifically, in VIS emergency mode, the fiber optic inertial navigation module is first activated with an accuracy compensation factor of 0.98 for signal compensation. Next, data from the visual SLAM camera group is used for map matching to correct inertial navigation drift and assist in positioning. Then, data from the laser rangefinder is used to improve odometry accuracy. A GNN neural network model is then used to generate a path. Finally, a risk assessment is performed on the generated path (risk assessment threshold = 0.2), ensuring the path risk value is below the threshold of 0.2 to guarantee navigation safety. If successful, emergency navigation continues; otherwise, landing / return to home is initiated.

[0048] Specifically, in this embodiment, the visual-inertial navigation emergency mode is activated, and the anti-interference navigation code is executed as follows: def anti_jamming_navigation(): if satellite_signal_lost(): activate_ins(precision compensation factor=θ.98) # Inertial navigation compensation match_terrain_map() # Visual terrain matching `generate_path_by_ai(model="GNN")` # AI generates an interference-resistant path return validate_path(risk_threshold=0.2) # Risk assessment In this embodiment, multi-mode navigation switching and path correction solve the navigation problem in GPS-denied environments by dynamically switching navigation modes through real-time assessment of satellite signal quality, combined with inertial navigation accuracy compensation, visual terrain matching and GNN neural network model path generation, thus ensuring the arrival rate of long-range operations.

[0049] like Figure 6 As shown in this embodiment, the self-organizing network communication method of the loitering munition includes: Step S21: At the application layer, the instructions are packaged into standard data frames; Step S22: At the network layer, select the communication path according to the AODVv2 protocol; Step S23: At the data link layer, allocate independent communication time slots to each node; Step S24: At the physical layer, frequency hopping spread spectrum is performed to switch communication frequencies.

[0050] like Figure 7-8 As shown in this embodiment, the coordinated attack decision-making of a loitering munition swarm includes the following steps: Step S31: Reconnaissance Phase (Network Topology): After the loitering munition cluster is released, the self-organizing network communication unit is automatically activated to form a mesh topology. Each loitering munition collects ground target data through the three-light pod and shares the reconnaissance data to the entire cluster in real time. The distributed AI decision-making system, based on the improved auction algorithm, dynamically allocates reconnaissance areas according to the location of each loitering munition, the reconnaissance range, and the target priority, to avoid duplicate reconnaissance or reconnaissance blind spots. Step S32: Target recognition stage: A deep learning-based multi-source data fusion target recognition model is used to fuse infrared and visible light images and point cloud data transmitted back by the three-light pod and loitering munition, and output target category and location information. Step S33: Strike Phase (Cluster Coordination): Based on the number and distribution of targets, the cluster is divided into multiple clusters, cluster head nodes are elected, and a coordinated strike strategy is formulated. Strike commands are sent to each loitering munition through the ad hoc network communication unit.

[0051] Specifically, in the target recognition stage of step S32 above, a multi-source data fusion AI target recognition model is adopted. This model uses a lightweight YOLOv7 architecture as its target detection head. Its overall architecture includes: a feature alignment module, an attention weight allocation module, and a dual-branch feature extraction network based on ResNet34 and PointNet++. Finally, the YOLOv7 detection head outputs the recognition result. The AI ​​target recognition model is input to the three-light pod data stream (640×480 resolution infrared thermal image, visible light RGB image, and laser point cloud data). After processing by the feature alignment module (achieving spatial synchronization of multi-source data based on spatial transformation matrix) and the attention weight allocation module (0.6 weight for infrared thermal image, prioritizing target thermal features), 1024-dimensional fused features are extracted by the dual-branch CNN of ResNet34 (visible light feature extraction) and PointNet++ (point cloud feature extraction). Finally, the YOLOv7 target detection head outputs the target category and location information. The AI ​​target recognition model is optimized with an adversarial training dataset, and the recognition confidence threshold is set to 0.89 to ensure recognition accuracy in complex environments.

[0052] In the strike phase of step S33 above, the cluster is automatically divided into 3 clusters based on the number and distribution of targets. Each cluster elects a cluster head node (responsible for coordinating instructions and summarizing data within the cluster). The distributed AI decision-making system formulates a coordinated strike strategy based on the defense strength, positional relationship, and remaining energy of each target and loitering munition. For example, some munitions are responsible for attracting air defense fire, some munitions carry out ultra-low-altitude penetration, and some munitions perform vertical top attack. Strike commands are issued to each loitering munition through the self-organizing network communication unit, realizing autonomous coordinated strike without external communication.

[0053] In this embodiment, the cluster-coordinated attack sequence includes: The mothership releases a cluster of loitering munitions; Some loitering munitions are used as decoys and jammers. Some loitering munitions perform ultra-low-altitude penetration and strike missions; Some loitering munitions perform target identification, tracking, and guidance correction tasks; To carry out the final strike and transmit the strike effectiveness assessment data back.

[0054] Specifically, in this embodiment, the aforementioned loitering munition cluster is limited to reconnaissance munitions, decoy munitions, and strike munitions, and the cluster's coordinated strike sequence includes: The mothership releases loitering munitions, including reconnaissance munitions, decoy munitions, and strike munitions; The decoy flares began their dive and prepared to release jamming. The strike missile then initiated an ultra-low-altitude penetration maneuver; After the reconnaissance missile identifies and confirms the target, it shares the information with the cluster. The distributed AI decision-making system plans coordinated paths for reconnaissance missiles, decoy missiles, and strike missiles based on target information; Decoy flares are used for interference; The reconnaissance missile performs terminal tracking and provides guidance correction for the strike missile; The strike missiles will carry out the final strike; The reconnaissance missile transmits data on the effectiveness of the strike.

[0055] like Figure 9 As shown in the figure, this embodiment takes the attack on the rear command post as a typical scenario, and the coordinated attack sequence is as follows (X-axis is time 0-120 seconds, Y-axis is distance from the target 30km-0km): t=0s: The mothership releases a loitering munition at a distance of 30km from the target and an altitude of 8000m. The initial positions of the reconnaissance munitions are 30km and 2000m, the initial positions of the decoy munitions are 30km and 2000m, and the initial positions of the strike munitions are 30km and 2000m. t=30s: The decoy missile begins its dive, descending from 25km to an altitude of 800m, preparing to release chaff; the reconnaissance missile continues to fly towards the target, activating its three-light pod for scanning; t=50s: The strike missile activates the ultra-low altitude penetration mode, descends to an altitude of 50m, and flies covertly along the terrain to avoid detection by enemy radar; t=60s (Event point E1: Target confirmation): The reconnaissance missile arrives at a distance of 10km and an altitude of 500m, confirms the enemy command post through the AI ​​target recognition model (recognition confidence > 0.89), and shares the target coordinates and feature information with the cluster; t=65s (Event point E2: Cooperative path planning): Based on the target information, the distributed AI decision-making system plans cooperative paths for the reconnaissance missile, decoy missile, and strike missile. The reconnaissance missile continues to track the target, the decoy missile prepares to release jamming, and the strike missile adjusts its penetration route. t=70s: The decoy flares release chaff at a distance of 8km, creating an electromagnetic interference zone to cover the penetration of the attack flares; t=90s (Event point E3: Terminal guidance activated): The reconnaissance missile reaches a distance of 5km and an altitude of 1000m, activates terminal tracking mode, and transmits target position correction data to the strike missile in real time; the strike missile activates terminal guidance, adjusts its flight attitude, and prepares for attack. t=100s: The strike missile reaches a distance of 3km and uses a 90° vertical top attack method to carry out a precision strike on the command post; at the same time, decoy missiles continue to maintain interference, and reconnaissance missiles monitor the strike effect; t=120s: The strike mission is completed. The reconnaissance missiles transmit strike effect assessment data back. The cluster decides whether to carry out a second strike or return to base based on the instructions.

[0056] In this embodiment, a distributed cluster decision-making and collaborative strike based on a self-organizing network is achieved by using an improved auction algorithm to allocate reconnaissance areas, a lightweight YOLOv7 model to identify targets, and a cluster collaborative strategy to carry out strikes. This enables autonomous cluster collaboration without external communication, significantly improving the anti-interference capability and efficiency of cluster operations.

[0057] The above content discloses the combat system and control method of the present invention. Compared with the prior art, the present invention achieves significant breakthroughs in combat radius, anti-jamming capability, cluster coordination efficiency, and cost control, and has the following advantages: 1. Significantly enhanced long-range combat capability: The mothership has a maximum range of ≥3000km. With the support of satellite relay communication, it can achieve long-range deep strikes. Its combat radius is more than 200% greater than that of existing loitering munition motherships (<1000km). It can launch attacks outside the enemy's defense zone without the need for frequent transits or deployments close to the combat area, reducing the risk of the mothership being exposed and expanding the combat range and tactical flexibility. 2. Significantly enhanced anti-jamming navigation reliability: In GPS-denied environments, the anti-jamming multi-mode navigation achieves a 98.7% arrival rate through inertial navigation compensation, visual terrain matching, and AI path self-correction, which is 146% higher than traditional systems (≤40%). This reduces the disconnection rate from over 60% in existing technologies to an extremely low level, ensuring stable flight and mission execution of UAVs in complex electromagnetic environments. It completely solves the problem of single satellite navigation being susceptible to interference, ensuring the stability of mission execution. 3. Optimization of target recognition and cluster decision-making efficiency: After adversarial training, the lightweight YOLOv7 target recognition model achieves a recognition accuracy of 92.4%, which is 23% higher than the traditional system (75%), and can accurately identify targets in complex environments; the cluster decision-making latency is ≤1.2s, which is 76% lower than the traditional system (≥5s), enabling rapid response to battlefield changes and improving the timeliness and coordination of cluster operations; 4. Autonomous decision-making in a communication-free environment: Construct a self-organizing network communication unit for loitering munitions and a distributed AI decision-making system, enabling swarm drones to autonomously complete target identification, area allocation, and coordinated strike path planning without external communication commands. This enhances the independence and anti-interference capabilities of swarm operations and avoids operational failures caused by communication interruptions or command interception. 5. Significantly reduced operational costs: The modular design keeps the total system cost below 3.5 million yuan, with a single target strike cost of only 138,000 yuan, a 72% reduction compared to existing similar systems (single target strike cost of 500,000+ yuan). This significantly improves the cost-effectiveness of the equipment, facilitates large-scale deployment and practical application, and enables a stronger operational scale effect within a limited budget.

[0058] In another embodiment of the present invention, a computer-readable storage medium is disclosed, on which a computer program is stored, which, when executed by a processor, implements the steps of the above-described control method.

[0059] It should be clarified that the present invention is not limited to the methods and systems disclosed above, but also includes various changes, modifications and additions made by those skilled in the art based on the ideas of the present invention, or changes in the order of steps.

[0060] When implemented in hardware, this invention can be electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the required tasks. These programs or code segments can be stored on a machine-readable medium or transmitted via a data signal carried on a carrier wave through a transmission medium or communication link. "Machine-readable medium" can include any medium capable of storing or transmitting information, such as electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, optical disks, hard disks, fiber optic media, radio frequency links, etc. The code segments can be downloaded via computer networks such as the Internet or intranets.

[0061] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Any other modifications or equivalent substitutions made by those skilled in the art to the technical solutions of the present invention, as long as they do not depart from the spirit and scope of the technical solutions of the present invention, should be covered within the scope of the claims of the present invention.

Claims

1. A long-range mother-daughter unmanned aerial vehicle (UAV) combat system, characterized in that, This system is a combat system based on the collaboration between a mothership and a loitering munition cluster, including: a mothership, several loitering munitions, and a distributed AI decision-making system; The mother machine includes: The main body integrates the power system, fuel storage unit, AI decision-making center and navigation module installation compartment; Fixed wings with an internal honeycomb composite structure; The ventral bomb bay has an internal pylon equipped with a mechanical locking mechanism for mounting and securing loitering munitions. Satellite communication antenna array, used for remote data exchange with the rear command center; The navigation module, located inside the navigation module installation compartment, is used to collect terrain images, target outlines, and distance data to provide positioning, navigation, and attitude references for the mother machine. The navigation module is configured as an anti-interference multi-mode navigation system, including: a satellite positioning module, a fiber optic inertial navigation module, a visual SLAM camera group, and an AI navigation decision center. The satellite positioning module, fiber optic inertial navigation module, and visual SLAM camera group are all connected to the AI ​​navigation decision center via a data bus, transmitting positioning data and status information to the AI ​​navigation decision center in real time. The three-light pod integrates infrared, visible light, and laser rangefinders to acquire real-time thermal distribution, visual images, distance data, and terrain reconnaissance data of ground targets. The loitering munition includes: The folding wings are folded when mounted on the mothership and automatically unfold when released. Self-organizing network communication unit is used for real-time data interaction between multiple loitering munitions to build a loitering munition cluster communication network; The warhead contains 5 kg of high-energy explosive and a precision fuse with a trigger mode that can be set according to the target type; The distributed AI decision-making system: Deployed inside the mothership, it receives navigation and reconnaissance data transmitted back by the navigation module, the three-light pod module, and the loitering munitions, performs real-time analysis and decision-making, and sends instructions to the loitering munition cluster.

2. The long-range mother-daughter unmanned aerial vehicle combat system according to claim 1, characterized in that: The loitering munition's self-organizing network communication unit adopts the OSI model protocol stack, wherein: The application layer is used to encapsulate target allocation instructions; At the network layer, the AODVv2 self-organizing network routing protocol is used for dynamic route selection; At the data link layer, TDMA or OFDMA time slot allocation methods are used to allocate independent communication time slots to each node of different loitering munitions. The physical layer is equipped with a frequency hopping spread spectrum module to avoid signal interference.

3. A remote mother-daughter unmanned aerial vehicle (UAV) control method, characterized in that: This control method is based on the combat system described in claim 1 or 2, wherein the step of navigating the mother aircraft based on an anti-jamming multi-mode navigation system includes the following steps: Step S11: The AI ​​navigation decision center monitors the signal-to-noise ratio of satellite signals in real time; Step S12: Switch the navigation mode according to the range of signal-to-noise ratio values: When the signal-to-noise ratio is greater than the first threshold, the satellite-dominated mode is adopted; When the second threshold < signal-to-noise ratio ≤ the first threshold, the inertial navigation-satellite fusion mode is adopted; When the signal-to-noise ratio is less than or equal to the second threshold or the signal is lost, the visual-inertial navigation emergency mode is activated. Step S13: Based on the judgment result, select and activate the corresponding navigation mode; Step S14: In any navigation mode, the AI ​​navigation decision center continuously compares the actual flight trajectory with the planned path. If the deviation exceeds the threshold, the path correction algorithm is automatically activated to adjust the flight parameters.

4. The remote mother-daughter unmanned aerial vehicle (UAV) control method according to claim 3, characterized in that: The first threshold is 15dB, the second threshold is 10dB, and the deviation threshold is ±50m; The visual-inertial navigation emergency mode is based on odometry calculation using visual SLAM and inertial navigation data fusion, supplemented by laser rangefinder for altitude correction and visual terrain matching.

5. A remote mother-daughter unmanned aerial vehicle (UAV) control method according to claim 3, characterized in that: The self-organizing network communication methods of loitering munitions include: Step S21: At the application layer, the instructions are packaged into standard data frames; Step S22: At the network layer, select the communication path according to the AODVv2 protocol; Step S23: At the data link layer, allocate independent communication time slots to each node; Step S24: At the physical layer, frequency hopping spread spectrum is performed to switch communication frequencies.

6. The remote mother-daughter unmanned aerial vehicle (UAV) control method according to claim 3, characterized in that: The coordinated strike decision-making of loitering munition swarms includes the following steps: Step S31: Reconnaissance phase. After the loitering munition cluster is released, it automatically forms a mesh topology, collects ground target data through the three-light pod, shares reconnaissance data in real time, and dynamically allocates reconnaissance areas based on the improved auction algorithm. Step S32: Target recognition stage. A deep learning-based multi-source data fusion target recognition model is used to fuse infrared and visible light images and point cloud data transmitted back by the three-light pod and loitering munition, and output target category and location information. Step S33: Strike phase. The cluster is divided into multiple sub-clusters according to the target distribution, a cluster head node is elected, and a coordinated strike strategy is formulated. Strike commands are sent to each loitering munition through the ad hoc network communication unit.

7. A remote mother-daughter unmanned aerial vehicle (UAV) control method according to claim 6, characterized in that: The target recognition model uses a lightweight YOLOv7 architecture as its target detection head. Its architecture includes: a feature alignment module, an attention weight allocator, and a two-branch feature extraction network based on ResNet34 and PointNet++. The YOLOv7 detection head outputs the recognition results, including target category and location information. The confidence threshold for the target recognition model was set to 0.

89.

8. A remote mother-daughter unmanned aerial vehicle (UAV) control method according to claim 6, characterized in that: The timing sequence of clustered coordinated attacks includes: The mothership releases a cluster of loitering munitions; Some loitering munitions are used as decoys and jammers. Some loitering munitions perform ultra-low-altitude penetration and strike missions; Some loitering munitions perform target identification, tracking, and guidance correction tasks; To carry out the final strike and transmit the strike effectiveness assessment data back.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the remote mother-daughter unmanned aerial vehicle control method as described in any one of claims 3-8.

Images