Multi-agent cooperative unmanned aerial vehicle cluster simulation training method, system, device and storage medium

Abstract

This application relates to the field of multi-agent cooperative technology, providing a method, system, device, and storage medium for multi-agent cooperative UAV swarm simulation training, addressing the problems of low penetration success rate and poor formation coordination efficiency of UAV swarms. The method includes: collecting biological motion data of pigeon flocks, swarm trajectory data of UAV swarms performing historical penetration missions, and magnetic field gradient data of the simulated environment; constructing a three-dimensional path library and calculating the geomagnetic heading angle deviation; generating multi-target cooperative paths based on the geomagnetic heading angle deviation and the three-dimensional path library, combined with a swarm intelligence optimization model; generating flight control commands based on the multi-target cooperative paths and the real-time status data of the UAV swarm; and loading the flight control commands into the simulation model of the UAV swarm to drive the UAV swarm to conduct cooperative penetration training in the simulated environment. This application improves the penetration success rate and formation coordination efficiency of UAV swarms in complex electromagnetic environments.

Description

Technical Field

[0001] This application relates to the field of multi-agent cooperative technology, and in particular to a multi-agent cooperative unmanned aerial vehicle (UAV) swarm simulation training method and system. Background Technology

[0002] In drone swarm penetration training under complex electromagnetic interference environments, it is necessary to address issues such as heading deviation caused by strong magnetic field disturbances, dynamic obstacle avoidance, and multi-drone collaborative path planning. Traditional simulation training methods are ill-suited to the real-time path optimization requirements of complex terrains such as mountains and urban canyons, necessitating an efficient training scheme that integrates environmental perception, biomimetic path optimization, and intelligent decision-making.

[0003] Existing solutions employ a multi-UAV cooperative path planning method based on reinforcement learning. This method generates swarm flight trajectories through a pre-trained neural network model and dynamically adjusts the path using real-time sensor data. This approach optimizes model parameters using historical mission data and verifies the swarm cooperation effect in a simulation environment.

[0004] The scheme relies on a large amount of pre-training data, and the model's generalization ability is limited by the coverage of training samples; the path planning process lacks an active compensation mechanism for geomagnetic interference, resulting in insufficient heading stability during actual flight; in addition, the biomimetic behavioral features are not deeply integrated into the decision-making model, making it difficult to adapt to the rapid response requirements of sudden terrain changes. Summary of the Invention

[0005] This application provides a multi-agent collaborative UAV swarm simulation training method and system to solve the problems of low penetration success rate and poor formation coordination efficiency of UAV swarms in complex electromagnetic environments in the prior art.

[0006] Firstly, this application provides a method for simulating and training multi-agent cooperative unmanned aerial vehicle (UAV) swarms, including:

[0007] Collect data on the biological movement of pigeon flocks, the swarm trajectory data of drone swarms during historical infiltration missions, and the magnetic field gradient data of the simulated environment;

[0008] Based on the cluster trajectory data and the pigeon flock biological movement data, a three-dimensional path library is constructed.

[0009] Calculate the geomagnetic heading angle deviation based on the magnetic field gradient data;

[0010] Based on the geomagnetic heading angle deviation and the three-dimensional path library, a multi-objective collaborative path is generated by combining a swarm intelligence optimization model.

[0011] Based on the multi-target cooperative path and combined with the real-time status data of the UAV swarm, flight control commands are generated.

[0012] The flight control commands are loaded into the simulation model of the UAV swarm, driving the UAV swarm to conduct collaborative penetration training in a simulated environment.

[0013] Optionally, the step of generating a multi-objective cooperative path based on the geomagnetic heading angle deviation and the three-dimensional path library, combined with a swarm intelligence optimization model, includes:

[0014] Retrieve the set of path points that match the current terrain from the 3D path library;

[0015] Based on the geomagnetic heading angle deviation, correct the heading coordinates of each path point in the path point set;

[0016] Identify terrain feature change points in the corrected path point set, and determine the avoidance type identifier based on the terrain feature change points;

[0017] The avoidance type identifier and the modified path point set are input into the swarm intelligence optimization model, and the terrain avoidance constraint path segment corresponding to the avoidance type identifier is output.

[0018] All the aforementioned terrain avoidance constraint path segments are spliced ​​together to form a multi-objective collaborative path.

[0019] Optionally, the step of inputting the avoidance type identifier and the modified path point set into the swarm intelligence optimization model, and outputting the terrain avoidance constrained path segment corresponding to the avoidance type identifier, includes:

[0020] The avoidance type identifier and the corrected path point set are input into the swarm intelligence optimization model. Based on the avoidance type identifier, the corresponding terrain spatial constraint relationship is called from the preset constraint rule base. The terrain spatial constraint relationship includes obstacle spacing threshold and height safety margin.

[0021] The modified path point set is used to perform path feasibility verification according to the terrain spatial constraint relationship, and conflicting path points that are less than the obstacle spacing threshold or less than the height safety margin are eliminated.

[0022] Based on the degree of association between the avoidance type identifier and the task, a corresponding first priority value is assigned to the avoidance type identifier;

[0023] Based on the first priority value, the non-conflict path points in the verified path point set are reassembled into a sequence to generate a terrain avoidance constraint path segment.

[0024] Optionally, the step of reorganizing the non-conflicting path points in the verified path point set based on the first priority value to generate terrain avoidance constraint path segments includes:

[0025] Based on a preset mapping table between the first priority value range and the urgency level, the urgency level identifier corresponding to the first priority value is determined.

[0026] Based on the first priority value and the terrain adaptability score, a corresponding second priority value is assigned to each non-conflict path point.

[0027] According to the sorting rules corresponding to the urgency indicators, the second priority values ​​of all non-conflicting path points are sorted to generate an ordered path point sequence.

[0028] The adjacent non-conflicting path points in the ordered path point sequence are smoothed by interpolation to form a continuous flight trajectory.

[0029] Apply terrain constraints corresponding to the avoidance type identifier to the continuous flight trajectory to generate terrain avoidance constraint path segments.

[0030] Optionally, the step of constructing a three-dimensional path library based on the cluster trajectory data and the pigeon flock biological movement data includes:

[0031] The cluster trajectory data is classified according to terrain type into canyon crossing trajectory, climbing trajectory, and diving trajectory.

[0032] Extract flight turning features and altitude change features from the pigeon flock's biological movement data;

[0033] The flight turning features are mapped to the canyon crossing trajectory, and the altitude change features are mapped to the climb trajectory and the dive trajectory;

[0034] Based on the mapping results, corresponding terrain-adaptive path units are generated, and all the terrain-adaptive path units are aggregated to form a three-dimensional path library.

[0035] Optionally, calculating the geomagnetic heading angle deviation based on the magnetic field gradient data includes:

[0036] Extract the magnetic field strength component of the current measurement point from the magnetic field gradient data;

[0037] Based on the preset geomagnetic reference field model, the theoretical magnetic field direction of the current measurement point is determined;

[0038] Calculate the directional deviation between the magnetic field intensity component and the theoretical magnetic field direction;

[0039] The directional deviation is converted into an angular offset in the heading coordinate system, and the angular offset is used as the geomagnetic heading angle deviation.

[0040] Optionally, the step of generating flight control commands based on the multi-target cooperative path and combined with the real-time status data of the UAV swarm includes:

[0041] The flight attitude of each UAV and the relative position of the enemy aircraft are extracted from the real-time status data;

[0042] The coordinates and arrival time of each UAV at the next navigation point are analyzed from the multi-target cooperative path;

[0043] For each UAV, the roll angle control parameters and angle of attack control parameters are calculated by integrating the coordinates, arrival time limit, flight attitude, and relative position of the enemy aircraft.

[0044] All the roll angle control parameters and angle of attack control parameters of the UAV are encapsulated into flight control commands.

[0045] Secondly, this application provides a multi-agent collaborative unmanned aerial vehicle (UAV) swarm simulation training system, comprising:

[0046] The data acquisition module is used to collect data on the biological movement of pigeon flocks, the swarm trajectory data of drone clusters when performing historical infiltration missions, and the magnetic field gradient data of the simulated environment.

[0047] The construction module is used to construct a three-dimensional path library based on the cluster trajectory data and the pigeon flock biological movement data;

[0048] The calculation module is used to calculate the geomagnetic heading angle deviation based on the magnetic field gradient data;

[0049] The first generation module is used to generate a multi-objective collaborative path based on the geomagnetic heading angle deviation and the three-dimensional path library, combined with a swarm intelligence optimization model.

[0050] The second generation module is used to generate flight control commands based on the multi-target cooperative path and combined with the real-time status data of the UAV cluster.

[0051] The loading module is used to load the flight control commands into the simulation model of the UAV cluster, driving the UAV cluster to conduct collaborative penetration training in a simulated environment.

[0052] Thirdly, this application provides a computing device including a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute a multi-agent cooperative drone swarm simulation training method as described in any of the first aspects.

[0053] Fourthly, this application provides a computer storage medium storing computer program instructions thereon, which, when executed by a processor, implement the multi-agent collaborative unmanned aerial vehicle swarm simulation training method described in any one of the first aspects.

[0054] This application provides a multi-agent collaborative unmanned aerial vehicle (UAV) swarm simulation training method. The method includes: collecting pigeon flock biological movement data, UAV swarm trajectory data during historical penetration missions, and magnetic field gradient data of the simulated environment; constructing a three-dimensional path library based on the swarm trajectory data and the pigeon flock biological movement data; calculating the geomagnetic heading angle deviation based on the magnetic field gradient data; generating a multi-target collaborative path based on the geomagnetic heading angle deviation and the three-dimensional path library, combined with a swarm intelligence optimization model; generating flight control commands based on the multi-target collaborative path and the real-time status data of the UAV swarm; and loading the flight control commands into the UAV swarm simulation model to drive the UAV swarm to conduct collaborative penetration training in the simulated environment.

[0055] The technical solution provided in this application has the following beneficial effects:

[0056] This application provides a biomimetic basis, historical experience data, and environmental perception foundation for the training system. It establishes a standardized path reference system integrating biomimetic features and historical experience. It enables precise quantification of heading errors in complex electromagnetic environments. It outputs optimized path schemes that take into account terrain avoidance, formation coordination, and electromagnetic compensation. It transforms the planned path into executable, high-precision flight control parameters. It achieves closed-loop verification throughout the entire process and enhances swarm coordination capabilities.

[0057] Furthermore, this application also ensures that the path planning matches the real-time environment by calling the terrain matching path point set and correcting the heading coordinates; it also achieves targeted path optimization by identifying terrain feature change points to determine avoidance type identifiers; and finally, it splices the optimized path segments to form a complete collaborative path.

[0058] Furthermore, this process enables dynamic adaptability of UAV swarm path planning and precision of tactical evasion in complex terrain, enhancing the realism and effectiveness of penetration training.

[0059] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0061] Figure 1 A flowchart illustrating a multi-agent collaborative unmanned aerial vehicle (UAV) swarm simulation training method provided in this application embodiment;

[0062] Figure 2 This is a schematic diagram of the structure of a multi-agent collaborative unmanned aerial vehicle swarm simulation training system provided in an embodiment of this application;

[0063] Figure 3 This is a schematic diagram of the structure of a computing device provided in an embodiment of this application. Detailed Implementation

[0064] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0065] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0066] In the field of UAV swarm collaborative training technology, existing reinforcement learning-based solutions suffer from three key problems: First, the training process requires massive amounts of data, but the availability of high-quality training samples is limited, leading to unstable model performance in unfamiliar environments. Second, the system lacks a dedicated mechanism to handle geomagnetic interference, making it prone to heading deviations in complex electromagnetic environments. Finally, path planning lacks consideration of biological swarm intelligence, resulting in insufficient flexibility in response to sudden terrain changes. These issues make existing training systems unable to meet the demands of realistic combat training.

[0067] To address these issues, this application proposes a multi-agent collaborative UAV swarm simulation training method. This method innovatively incorporates pigeon flock biomovement data to optimize path planning, while automatically correcting heading deviations through real-time geomagnetic field measurements. Specifically, the system analyzes the flight characteristics of pigeon flocks to build a more intelligent path library and dynamically adjusts flight routes based on real-time geomagnetic data. This design reduces dependence on the amount of training data, effectively counteracts electromagnetic interference, and endows the UAV swarm with environmental adaptability similar to that of biological communities. Simulation verification demonstrates that this method improves the collaborative combat effectiveness of UAV swarms in complex environments and solves the problem of insufficient adaptability in existing technologies.

[0068] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0069] Figure 1 A flowchart of a multi-agent cooperative UAV swarm simulation training method provided in this application embodiment is shown below. Figure 1 As shown, the method includes:

[0070] Step 101: Collect data on the biological movement of pigeon flocks, the cluster trajectory data of drone swarms when performing historical infiltration missions, and the magnetic field gradient data of the simulated environment.

[0071] In step 101, the pigeon flock biomovement data represents the recorded motion characteristics of the pigeon flock during flight, including changes in flight direction and altitude adjustment. The swarm trajectory data represents the position and attitude changes recorded by the UAV swarm during mission execution. The magnetic field gradient data represents information reflecting changes in the intensity of the spatial magnetic field obtained through sensors.

[0072] In this embodiment, flight data of the pigeon flock under natural conditions is first collected using observation equipment, with a focus on recording the patterns of change in their turning angle and altitude. Simultaneously, flight trajectory information recorded by the UAV swarm during previous missions is retrieved, including position coordinates and flight attitude. A magnetic field sensor network is deployed in a simulated training environment to monitor changes in magnetic field strength at various points in space in real time. These three types of data will serve as the basis for subsequent path planning and heading correction.

[0073] For example, during the preparation phase of a simulation training exercise, researchers first observed and recorded the flight trajectories of 100 homing pigeons in the mountains, focusing on extracting their average turning angle and altitude adjustment rate when crossing canyons. They also retrieved data from 20 penetration missions conducted by a drone swarm in similar terrain last year. Additionally, 50 magnetic field sensors were deployed at the training site to continuously monitor the geomagnetic field changes in the training area. This data was then processed and stored in the training system's database.

[0074] Step 102: Construct a three-dimensional path library based on the cluster trajectory data and the pigeon flock biological movement data.

[0075] In step 102, the three-dimensional path library represents a path reference database containing spatial location information and terrain features.

[0076] In this embodiment of the application, the collected historical trajectories of UAVs are classified and organized according to terrain features, including level flight segments, climb segments, and dive segments; at the same time, the motion features in the pigeon flock data are analyzed, and their feature parameters are matched and optimized with the corresponding types of historical trajectories; finally, the optimized path segments are integrated into a standardized three-dimensional path library.

[0077] For example, the system categorizes historical trajectory data into three types: those between 2000-2200 meters above sea level are classified as canyon crossing sections, those with a greater rate of ascent are classified as climbing sections, and those with a greater rate of descent are classified as diving sections. Then, the system optimizes the turning angles of the canyon sections based on the turning characteristics of the pigeon flock data, and optimizes the altitude adjustment rates of the climbing and diving sections based on the altitude change characteristics of the pigeon flock, ultimately forming a three-dimensional path library containing 30 optimized paths.

[0078] Step 103: Calculate the geomagnetic heading angle deviation based on the magnetic field gradient data.

[0079] In step 103, the geomagnetic heading angle deviation represents the angular difference between the actual heading and the theoretical heading.

[0080] In this embodiment, the magnetic field strength component of the current measurement point is obtained from the magnetic field sensor; the theoretical magnetic field direction of the point is calculated according to the preset geomagnetic model; the directional difference between the measured value and the theoretical value is obtained through vector operation; and finally, the difference is converted into an angular offset in the heading coordinate system.

[0081] For example, in a certain training exercise, the system measured the magnetic field strength components at the current point as 45 units in the X direction, -12 units in the Y direction, and 38 units in the Z direction; according to the theoretical values ​​of the model, they should be X 48, Y -8, and Z 40; the angle deviation was calculated to be 8.2 degrees using the formula θ=arccos[(A·B) / (|A||B|)], where A is the measured vector and B is the theoretical vector; finally, this deviation value was converted into heading correction parameters.

[0082] Step 104: Based on the geomagnetic heading angle deviation and the three-dimensional path library, generate a multi-objective collaborative path using a swarm intelligence optimization model.

[0083] In step 104, the swarm intelligence optimization model refers to a computational model that simulates the cooperative behavior of biological groups. Its physical meaning lies in coordinating the flight paths of multiple UAVs through a distributed decision-making mechanism, enabling the swarm to autonomously adapt to environmental changes like a flock of birds, dynamically adjusting its flight path to avoid collisions and maintain formation, while simultaneously optimizing overall penetration efficiency. This model does not rely on centralized control but achieves global optimization through local interaction. The multi-objective cooperative path refers to the optimized flight path generated by a UAV swarm in complex terrain environments to simultaneously meet multiple requirements such as heading deviation compensation, terrain avoidance, and formation coordination. This path is composed of multiple path segments with different tactical objectives, each corresponding to a specific flight strategy, such as altitude-maintaining segments and sharp-turn segments, ultimately forming a complete flight trajectory that balances safety and mission efficiency.

[0084] In this embodiment, a set of path points matching the current terrain is selected from the path library; the heading coordinates of each point are corrected according to the geomagnetic deviation; key terrain change points in the path are identified and the avoidance strategy type is determined; the strategy type and the corrected path points are input into the optimization model; path segments that meet the requirements of each strategy are output; and finally, the path segments are spliced ​​together to form a complete collaborative path.

[0085] For example, the system selects a set of path points near an altitude of 2150 meters from the path library and corrects the heading by an 8.2-degree deviation; it identifies three key points: altitude change point, turn start point, and turn end point; these correspond to three strategies: altitude maintenance, turn preparation, and turn execution, respectively; after optimization, it generates three path segments: level flight segment, turn segment, and recovery segment; and finally splices them together to form a complete path.

[0086] Step 105: Based on the multi-target cooperative path and combined with the real-time status data of the UAV swarm, generate flight control commands.

[0087] In step 105, the real-time status data comes from real-time sensor acquisition and historical mission records in the simulated training environment. Specifically, it includes wind speed data (acquired in real-time by meteorological sensors) and enemy aircraft position information (generated through the red and blue coordinate data of the simulated combat system). This data is used to describe the spatiotemporal state changes of the UAV swarm in the dynamic combat environment. Flight control commands are specific control parameters that convert the planned path into executable actions of the UAV, including adjustments to roll angle and angle of attack. Physically, they control the flight direction and altitude by changing the UAV's attitude angles (left and right tilt angles and up and down pitch angles), ensuring that the UAV accurately tracks the planned path. The command content is dynamically generated based on real-time calculated waypoint positions, time limits, and current flight status, and is finally sent to the UAV's actuators in the form of digital signals.

[0088] In this embodiment, the next waypoint information in the cooperative path is parsed; the current flight status and threat information of the UAV are obtained; the required roll angle and angle of attack adjustment are calculated; and the calculation results are converted into control command format.

[0089] For example, the system resolves the coordinates of the next waypoint as X degrees east longitude, Y degrees north latitude, and 2153 meters above sea level; detects that the current UAV attitude is 3 degrees horizontal tilt and 2 degrees pitch; calculates that a 15-degree left roll and a 3-degree upward roll are required; and generates the corresponding binary control commands.

[0090] Step 106: Load the flight control commands into the simulation model of the UAV swarm and drive the UAV swarm to conduct collaborative penetration training in the simulated environment.

[0091] In step 106, the simulation model represents a mathematical model that simulates the physical characteristics and motion laws of the UAV. Cooperative penetration training refers to mission drills performed by multiple UAVs working together.

[0092] In this embodiment, control commands are input into the simulation system; the virtual drone swarm is driven to fly along the planned path; the training effect is monitored in real time; and the training data is recorded for subsequent optimization.

[0093] For example, the system sends control commands to the simulation platform, and the virtual drone swarm begins to execute the mission. It maintains formation flight in the simulated canyon environment, successfully avoids simulated threat targets, maintains a stable course throughout the flight, and completes the predetermined penetration mission.

[0094] This method integrates biological motion characteristics and geomagnetic environment data to construct an intelligent UAV swarm training system. The system can automatically generate flight paths adapted to complex terrain, accurately compensate for geomagnetic interference, and achieve collaborative control of multiple UAVs. In simulated training, it demonstrates good environmental adaptability and task completion capabilities, thus improving training effectiveness.

[0095] To address the collaborative path planning problem of UAV swarms in complex terrain, in some embodiments, step 104: generating a multi-objective collaborative path based on the geomagnetic heading angle deviation and the three-dimensional path library, combined with a swarm intelligence optimization model, includes:

[0096] Step 201: Retrieve the set of path points that match the current terrain from the 3D path library.

[0097] In step 201, the current terrain refers to the actual terrain environment characteristics of the UAV cluster in real time, specifically determined as follows: a mountain terrain digital elevation model is preset in the simulation training system, and the corresponding terrain type data is matched according to the real-time positioning coordinates of the UAV cluster, including altitude, slope, and obstacle distribution characteristics. Areas with an altitude of 2000-2500 meters and a slope greater than 30 degrees are defined as complex mountain terrain that needs to be processed. These parameters are derived from the 3D terrain data of the training area provided by the geographic information system. The path point set refers to the set of path points selected from the 3D path library that match the terrain characteristics of the training area. These point sets contain location coordinates, altitude values, and terrain type labels, serving as a basic reference for subsequent path planning.

[0098] In this embodiment, the system first determines the terrain type based on the current location of the UAV, and then retrieves historical path points stored under the same terrain category from the path library. These point sets are stored in categories such as canyons, climbs, and dives to ensure that the retrieved path points match the current environmental characteristics.

[0099] Step 202: Correct the heading coordinates of each path point in the path point set according to the geomagnetic heading angle deviation.

[0100] In step 202, the correction of the heading coordinates is a process of adjusting the direction angle of the path point according to the geomagnetic deviation. The corrected coordinates can offset the navigation error caused by geomagnetic interference.

[0101] In this embodiment, the system reads the current geomagnetic deviation angle and superimposes the angle value onto the original direction angle of each path point, so that the heading of all path points is uniformly compensated for the influence of geomagnetic interference, ensuring that the subsequent planned path maintains the correct direction during actual flight.

[0102] Step 203: Identify terrain feature change points in the corrected path point set, and determine the avoidance type identifier based on the terrain feature change points.

[0103] In step 203, terrain feature change points refer to special points in the path point set that identify abrupt terrain changes. Avoidance type identifiers are classification labels for response strategies defined based on the characteristics of the abrupt changes.

[0104] In this embodiment of the application, the system scans the corrected path point set, automatically identifies key location points such as sudden changes in altitude and the start of turns, and assigns a type label to each feature point, such as "altitude maintenance" or "sharp turn". These labels will guide the selection of subsequent path optimization strategies.

[0105] Step 204: Input the avoidance type identifier and the corrected path point set into the swarm intelligence optimization model, and output the terrain avoidance constraint path segment corresponding to the avoidance type identifier.

[0106] In step 204, the terrain avoidance constraint path segment refers to an optimized path fragment that meets specific avoidance requirements, and its geometric features strictly correspond to the avoidance type identifier. There are multiple terrain avoidance constraint path segments, and each avoidance type identifier corresponds to an independent path segment. Finally, a complete multi-objective cooperative path is formed by splicing all the generated path segments. Specifically, an avoidance type identifier corresponds to a set of continuous path points (rather than a single path point), and each set of path points constitutes an independent path segment. Each path segment has different geometric features and constraints according to its corresponding sub-objective type (such as high ground capture, obstacle avoidance, etc.).

[0107] In this embodiment, the system inputs the type identifier and path point set into the optimization model. The model first matches the corresponding constraints for each identifier, and then adjusts the position of the path points to meet the constraints. For example, for sharp turn path segments, the turning radius must be ensured to meet the standard. Finally, it outputs several optimized paths that meet the requirements.

[0108] Step 205: Connect all the terrain avoidance constraint path segments to form a multi-objective collaborative path.

[0109] In this embodiment, the system connects the horizontal flight segment, the turning segment, and the recovery segment end to end in order from low altitude to high altitude to form a continuous flight path. This path simultaneously meets the requirements of heading stability, terrain avoidance, and mission timing.

[0110] Here is a specific example:

[0111] In a drone swarm mountain penetration training mission, the system first retrieved a set of path points matching the current canyon terrain from an established 3D path library. This set included 30 path points near an altitude of 2150 meters, derived from previously optimized historical trajectory data. Next, the system applied a previously calculated 8.2-degree geomagnetic heading angle deviation to correct these path points. Specifically, the heading angle of each point was increased by 8.2 degrees to compensate for geomagnetic interference. After correction, the system automatically scanned the path point set and identified three key terrain feature changes: the first was the starting point where the altitude suddenly increased from 2150 meters to 2155 meters, marked as an altitude-maintaining marker; the second was the starting point of a sharp turn, marked as a turn preparation marker; and the third was the ending point of the turn, marked as a turn execution marker. After inputting these markers and the corrected waypoints into the swarm intelligence optimization model, the model first matches the altitude-maintaining markers with a constraint that vertical fluctuations do not exceed 5 meters, and matches the two turning markers with a turning radius greater than 50 meters. Then, through iterative calculation, the model filters out the waypoints that meet the conditions and reorders them. The turning radius of the waypoints in the turning segments is calculated using the formula R=v² / (g×tanφ) to ensure compliance, where R is the turning radius, v is the preset flight speed of 25 m / s, g is the gravitational acceleration, and φ is the roll angle. The final model outputs three optimized paths: the first segment maintains a horizontal flight at an altitude of 2150-2155 meters for 60 meters; the second segment completes a left turn with a radius of 52 meters; and the third segment returns to an altitude of 2150 meters to continue flight. The system then stitches these three paths together in the order of the tasks into a complete multi-objective cooperative path. This path compensates for geomagnetic deviations and meets terrain avoidance requirements, providing a precise route basis for the subsequent generation of flight control commands.

[0112] In this embodiment of the application, the method intelligently integrates geomagnetic correction and terrain feature recognition, so that the generated cooperative path not only ensures heading accuracy but also has good terrain adaptability, effectively improving the penetration capability and cooperative efficiency of UAV swarms in complex environments, while reducing the need for manual intervention.

[0113] To address the path optimization problem of drone swarms in complex terrain, in some embodiments, step 204: inputting the avoidance type identifier and the corrected path point set into a swarm intelligence optimization model, and outputting the terrain avoidance constrained path segment corresponding to the avoidance type identifier, includes:

[0114] Step 301: Input the avoidance type identifier and the corrected path point set into the swarm intelligence optimization model. Based on the avoidance type identifier, call the corresponding terrain spatial constraint relationship from the preset constraint rule base. The terrain spatial constraint relationship includes obstacle spacing threshold and height safety margin.

[0115] In step 301, the terrain spatial constraint relationship refers to the spatial restriction conditions set for different avoidance types. Among them, the obstacle spacing threshold specifies the minimum safe distance between the UAV and the obstacle, and the altitude safety margin limits the adjustable range of the flight altitude.

[0116] In this embodiment, the system first reads the avoidance type identifier and retrieves the corresponding constraint conditions from the rule base according to the identifier type. For example, the sharp turn type corresponds to a larger spacing threshold, and the height maintenance type corresponds to a strict height margin, so as to ensure that the subsequent path planning meets the special requirements of various terrains.

[0117] Step 302: Perform path feasibility verification on the modified path point set according to the terrain spatial constraint relationship, and remove conflicting path points that are less than the obstacle spacing threshold or less than the height safety margin.

[0118] In step 302, conflict path points refer to path points that violate terrain spatial constraints. Path feasibility verification is a process of determining whether each path point meets the constraints through geometric calculations.

[0119] In this embodiment, the system checks each corrected path point, calculates the distance and height deviation between each point and the nearest obstacle, marks points that do not meet the spacing or height requirements as conflict points and removes them, and retains the set of points that meet the conditions for subsequent optimization.

[0120] Step 303: Based on the degree of association between the avoidance type identifier and the task, assign a corresponding first priority value to the avoidance type identifier.

[0121] In step 303, the correlation degree is determined by analyzing the importance relationship between the training task objective and the terrain avoidance requirements. It refers to the weight of different avoidance types on task completion, specifically preset by the commander based on task characteristics. For example, in a penetration mission, maintaining altitude is crucial for concealment, resulting in a high correlation degree, while conventional turns have a relatively small impact on the mission, resulting in a low correlation degree. The system quantifies these correlation degrees into priority values ​​for path optimization. The first priority value is a quantitative indicator reflecting the importance of the avoidance type, and its magnitude depends on the criticality of that type for task completion.

[0122] In this embodiment, the system assigns a priority value to each avoidance type according to task requirements. For example, altitude maintenance types that are related to flight safety are given a higher priority, while regular turning types are given a relatively lower priority. These values ​​will guide the subsequent sorting and reorganization of waypoints.

[0123] Step 304: Based on the first priority value, the non-conflict path points in the verified path point set are reassembled into a sequence to generate a terrain avoidance constraint path segment.

[0124] In step 304, sequence recombination refers to the process of reordering path points according to priority.

[0125] In this embodiment, the system sorts the retained path points according to their priority values, with high-priority points placed in key locations. Then, the ordered points are connected into smooth path segments using a curve fitting method, ensuring that each path segment conforms to both terrain constraints and task priority requirements.

[0126] Here is a specific example:

[0127] During a mountain penetration training mission, the system performed in-depth optimization on the three identified evasion type markers. First, based on mission safety requirements, the altitude-maintaining marker was assigned a priority value of 0.9, as it directly relates to the effectiveness of stealth penetration. The turning preparation and turning execution markers were assigned priority values ​​of 0.7 and 0.6 respectively. These values ​​were determined by the commander based on the fact that altitude concealment was more important than maneuverability in this penetration mission. Next, the system retrieved the corresponding terrain spatial constraints from the constraint rule base. The altitude-maintaining marker was matched with a vertical fluctuation limit of no more than 5 meters, derived from the minimum safe flight altitude standard for UAVs. The turning marker was matched with an obstacle spacing threshold of 50 meters and a lower limit for the turning radius of 52 meters. Finally, the system verified the corrected 30 path points using a distance formula. Calculate the distance between each point and the simulated mountain, where x, y, and z are coordinate values. Remove three dangerous points with a distance less than 50 meters. Simultaneously, check the altitude values ​​and remove two points exceeding the 2150-2155 meter range. The remaining 25 non-conflict path points are reordered according to priority, with altitude-critical points given priority, followed by turning points. Three optimized paths are generated using curve fitting: the first segment involves a 40-meter straight flight at an altitude of 2150-2152 meters; the second segment involves a turn with a radius of 53 meters. The calculation ensures safety, where v is taken as 25 m / s and g is... φ is the system-recommended roll angle of 25 degrees. The third segment descends smoothly to an altitude of 2150 meters and continues flying. The resulting terrain avoidance constraint path segment not only meets all safety constraints but also reflects the mission priority, providing a reliable route for subsequent coordinated penetration.

[0128] In this embodiment of the application, the method uses intelligent constraint verification and priority sorting to ensure that the generated path segments not only guarantee flight safety but also highlight key points of the mission, effectively improving the adaptability of UAV swarms in complex terrain and the quality of mission completion.

[0129] To further improve the accuracy and adaptability of UAV swarm path planning, in some embodiments, step 304: based on the first priority value, the non-conflicting path points in the verified path point set are reassembled into a sequence to generate terrain avoidance constraint path segments, including:

[0130] Step 401: Based on the mapping table between the preset first priority value range and urgency, determine the urgency identifier corresponding to the first priority value.

[0131] In step 401, the urgency indicator refers to the level label divided according to the urgency of the task. The priority value is converted into three urgency levels: high, medium, and low through a pre-set numerical range mapping table.

[0132] In this embodiment of the application, the system reads the first priority value of each avoidance type, looks up the corresponding urgency identifier in the table, such as priority 0.9 is mapped to high urgency, 0.7 is mapped to medium urgency, and 0.6 is mapped to low urgency. These identifiers will guide the sorting strategy of subsequent path points.

[0133] Step 402: Based on the first priority value and the terrain adaptability score, assign a corresponding second priority value to each non-conflict path point.

[0134] In step 402, the terrain adaptability score is derived by quantitatively analyzing the degree of matching between the waypoint and terrain features. Specifically, it considers factors such as the distance between the waypoint and the nearest obstacle, altitude stability, and deviation from the ideal flight path. A weighted calculation method is used to integrate these indicators into a single score value. The waypoints that are farther from obstacles, have more stable altitudes, and have smaller deviations score higher. The second priority value is the final priority calculated by combining the first priority and the terrain adaptability score.

[0135] In this embodiment, the system calculates a terrain adaptability score for each non-conflict path point, taking into account factors such as the distance between the point and obstacles and height stability. Then, the score is weighted and summed with the first priority to obtain the second priority value, ensuring that important path points simultaneously meet the terrain adaptability requirements.

[0136] Step 403: Sort the second priority values ​​of all non-conflicting path points according to the sorting rules corresponding to the urgency indicators to generate an ordered path point sequence.

[0137] In step 403, the sorting rule refers to the arrangement of pathpoints corresponding to different urgency levels. High urgency points are sorted in descending order, while medium and low urgency points are sorted in ascending order. In the sorting rule corresponding to urgency indicators, the sorting of different urgency indicators is performed independently. The pathpoint sequence [c, b, a] corresponding to high urgency indicators and the sequence [d, e] corresponding to low urgency indicators are not mixed. Instead, they are first sorted by urgency level (within the high urgency group, sorted in descending order of [c, b, a], and within the low urgency group, sorted in ascending order of [d, e]). Then, the complete sorting result of the high urgency group is output with priority over the low urgency group, ultimately forming the overall sequence [c, b, a, d, e]. The ordered pathpoint sequence refers to the set of pathpoints rearranged according to urgency indicators and priority rules, where high-urgency, high-priority points are placed first, and low-urgency, low-priority points are placed last, forming a pathpoint queue arranged from most important to least important points, conforming to task priority requirements.

[0138] In this embodiment, the system groups waypoints according to their urgency level. The high urgency group is arranged from high to low according to the second priority, and the medium and low urgency groups are arranged from low to high. Finally, the high urgency group is placed at the beginning of the sequence to form an ordered sequence that takes into account both urgency and terrain adaptability.

[0139] Step 404: Perform smooth interpolation on adjacent non-conflicting path points in the ordered path point sequence to form a continuous flight trajectory.

[0140] In step 404, adjacent non-conflicting path points refer to two non-conflicting path points whose order in the ordered path point sequence is determined by their arrangement in the sequence and is independent of the original spatial distribution. Smoothing interpolation is the process of adding intermediate points between adjacent path points to make the trajectory continuous. A continuous flight trajectory refers to a coherent flight path generated after smoothing interpolation of the ordered path point sequence. By adding calculation points between the original measurement points, abrupt changes and transitions in the path are eliminated, ensuring that the UAV can fly smoothly and steadily along the trajectory, avoiding dangerous maneuvers such as sharp turns or sudden changes in altitude.

[0141] In this embodiment, the system analyzes the sorted path point sequence, inserts several calculation points between every two actual measurement points, and uses a curve fitting algorithm to ensure that the generated trajectory has smooth transitions and continuous flight, avoiding sharp turns or sudden changes in altitude.

[0142] Step 405: Apply terrain constraints corresponding to the avoidance type identifier to the continuous flight trajectory to generate terrain avoidance constraint path segments.

[0143] In step 405, the terrain spatial constraint relationship is a general constraint rule (such as a spacing threshold) called from the constraint rule library, while the terrain constraint condition is an instance of applying the relationship to the currently generated continuous flight trajectory. The two are related as rule and implementation.

[0144] In this embodiment, the system checks whether the interpolated continuous trajectory fully meets the constraints of various avoidance markers, and makes fine adjustments to local non-compliant areas, such as adjusting the turning radius or the rate of change of height, until a fully compliant terrain avoidance constraint path segment is generated.

[0145] Here is a specific example:

[0146] In a drone swarm mountain penetration training mission, the system performed in-depth optimization on 25 verified non-conflict path points. First, based on a pre-defined mapping table, the altitude hold indicator (0.9 priority) was assigned to high urgency, turn preparation (0.7 priority) to medium urgency, and turn execution (0.6 priority) to low urgency. Next, a terrain adaptability score was calculated for each path point, using the following formula: Where d represents the actual distance between the point and the nearest obstacle, D is the safety threshold of 50 meters, h represents the height fluctuation value, and H is the allowable fluctuation range of 5 meters. and With weighting coefficients set to 0.6 and 0.4 respectively, the calculated score for each point ranges from 0.5 to 0.9 using this formula. Then, combining the first priority score with the terrain adaptability score, a weighted formula is applied. Calculate the second priority value for each point, where Assuming the highest priority and S the terrain adaptability score, pathpoints in the high-urgency group are sorted in descending order according to the urgency indicator, while those in the medium-low urgency group are sorted in ascending order, resulting in an ordered sequence. When smoothing this sequence, two intermediate points are inserted between every two adjacent points, and a cubic spline interpolation algorithm is used to ensure trajectory continuity. Pathpoints in key turning segments are determined using a formula... Verify the turning radius, where v remains at 25 m / s and g is... φ is set to 25 degrees to ensure that the radius is not less than 52 meters. Finally, the entire trajectory is constrained and checked, and the positions of the three points are adjusted to keep the altitude within the range of 2150-2155 meters throughout the entire journey. Three complete paths are output: the first segment flies straight for 45 meters at an altitude of 2151-2153 meters, the second segment smoothly turns with a radius of 53 meters, and the third segment descends steadily to 2150 meters, forming a terrain avoidance constraint path segment that meets all the constraints.

[0147] In this embodiment of the application, the method uses multi-level priority division and terrain adaptability optimization to make the generated path segments highlight the key points of the task while fully considering the terrain features, thus achieving the best balance between safety and task completion and improving the combat effectiveness of UAV swarms in complex environments.

[0148] To construct a UAV path planning database adapted to complex terrain, in some embodiments, step 102: constructing a three-dimensional path library based on the cluster trajectory data and the pigeon flock biological movement data includes:

[0149] Step 501: Classify the cluster trajectory data according to terrain type into canyon crossing trajectory, climbing trajectory, and diving trajectory.

[0150] In step 501, terrain type classification refers to the process of dividing the flight path into different segments based on the rate of change of altitude and spatial characteristics. The canyon crossing segment refers to the flight portion that maintains a relatively stable altitude within a narrow space. The climb segment refers to the flight portion that continuously increases altitude. The dive segment refers to the flight portion that continuously decreases altitude.

[0151] In this embodiment of the application, the system analyzes the altitude change curve and spatial location distribution in historical trajectory data, and automatically identifies three types of flight segments with obvious characteristics. Among them, the area with gentle altitude change and obstacles on both sides is classified as the canyon crossing segment, the segment with continuous altitude increase is classified as the climb segment, and the segment with continuous altitude decrease is classified as the dive segment, thus establishing a basic classification framework for subsequent feature mapping.

[0152] Step 502: Extract flight turning features and altitude change features from the pigeon flock's biological movement data.

[0153] In step 502, flight turning characteristics refer to the motion characteristics of a flock of pigeons when changing direction in complex terrain, including turning angle and turning rate. Altitude change characteristics refer to the motion characteristics of a flock of pigeons when adjusting their flight altitude, including climb rate and dive rate.

[0154] In this embodiment, the system processes pigeon flock observation data, extracts the typical turning angle range and rate of change of the flock when crossing obstacles as turning features, and statistically analyzes the average rate of change of the flock when adjusting altitude as altitude features. These biological motion features provide a natural reference for UAV path optimization.

[0155] Step 503: Map the flight turning features to the canyon crossing trajectory, and map the altitude change features to the climb trajectory and the dive trajectory.

[0156] In step 503, feature mapping refers to the process of matching and optimizing biological motion characteristic parameters with UAV trajectory segments, by adjusting the UAV trajectory parameters to approximate the motion characteristics of the biological population. The specific implementation process of mapping the altitude change characteristics to the climbing trajectory and the diving trajectory is as follows: First, extract two characteristic values ​​from the pigeon flock data: a typical altitude change rate of 2 meters per second ascent and 1.8 meters per second descent. Then, apply a forward mapping to the climbing trajectory, uniformly adjusting all climbing rates in the original trajectory to 2 meters per second according to the formula "adjusted rate = original rate × 2 / average pigeon flock ascent rate 1.5". At the same time, apply a reverse mapping to the diving trajectory, uniformly adjusting the original descent rate to 1.8 meters per second according to the formula "adjusted rate = original rate × 1.8 / average pigeon flock descent rate 2.2". For example, the rate of a certain segment of the original climbing trajectory is adjusted from 1.2 meters per second to 1.6 meters per second, and the rate of the original diving trajectory is adjusted from 2.5 meters per second to 2.05 meters per second. Finally, the two types of trajectories are matched with different altitude change characteristics of the pigeon flock.

[0157] In this embodiment, the system applies the typical turning angle of a flock of pigeons to the turning portion of the canyon crossing trajectory, making the turning action of the UAV closer to the smooth characteristics of a flock of pigeons; at the same time, it applies the climb and dive rate parameters of the flock of pigeons to the corresponding UAV trajectory segment, optimizing the altitude adjustment process, so that the altitude change of the UAV is more in line with the laws of natural flight.

[0158] Step 504: Based on the mapping results, generate corresponding terrain adaptation path units, and aggregate all the terrain adaptation path units to form a three-dimensional path library.

[0159] In step 504, the mapping result refers to the output after matching the pigeon flock flight characteristics with the UAV historical trajectory classification segments. It includes: matching parameters (such as turning angle values ​​and altitude change rate) between each terrain classification segment (canyon crossing segment, climb segment, and dive segment) and the corresponding pigeon flock characteristics; specifically, parameterized key-value pairs, for example, {"canyon crossing segment": turning angle ±30°}, {"climb segment": altitude change rate 2m / s}; quantitative relationships: each terrain classification segment generates an independent mapping result (3 in total), and each result corresponds to a path unit (such as canyon path unit and climb path unit). Finally, all units are aggregated to form a three-dimensional path library. The terrain-adapted path unit refers to a standardized flight path segment optimized by biometric features.

[0160] In this embodiment, the system encapsulates the optimized trajectory segments into independent units. Each unit contains information such as location coordinates, motion parameters, and applicable terrain. Then, all units are organized into a structured database according to terrain type and spatial location, providing a modular reference for subsequent path planning.

[0161] Here is a specific example:

[0162] During the construction of a drone swarm training system, the system first analyzed and processed the collected flight data of pigeons. From the flight trajectories of 100 pigeons, the average turning angle when crossing a canyon was extracted as 35 degrees, the average rate of ascent during the climb phase was 2 meters per second, and the average rate of descent during the dive phase was 1.8 meters per second. These values ​​were obtained by taking the arithmetic mean of the statistical data of the pigeon flock. Simultaneously, the system retrieved data from 20 historical missions performed by the drone swarm in similar terrain. Of these, 120 trajectories located at an altitude of 2000-2200 meters with altitude fluctuations of less than 3 meters were classified as canyon crossing sections; 80 trajectories with an ascent rate greater than 1.5 meters per second were classified as climb sections; and 60 trajectories with a descent rate greater than 2 meters per second were classified as dive sections. Next, the system mapped the 35-degree turning characteristic of the pigeon flock onto the canyon crossing section trajectory, using the angle adjustment formula θ_new=θ_o. The algorithm ld+(35-θ_avg) optimizes each turning point, where θ_old is the original turning angle and θ_avg is the average turning angle of the historical trajectory (30 degrees), ensuring that the turning angles of the optimized canyon sections are all close to 35 degrees. Simultaneously, the ascent and descent rate characteristics of the pigeon flock are mapped to corresponding trajectory segments. The rate of change of altitude in each segment is optimized using the rate adjustment formula v_new=v_old×(v_pigeon / v_avg), where v_pigeon is the pigeon flock rate and v_avg is the average rate of the historical trajectory, making the rate of ascent segments approach 2 meters per second and the rate of descent segments approach 1.8 meters per second. Finally, the system generates 150 optimized path units, including 60 canyon crossing units, 50 ascent units, and 40 descent units. Each unit is labeled with its applicable altitude range and motion parameters, forming a three-dimensional path library containing 30 typical paths.

[0163] In this embodiment, the method integrates biological swarm intelligence and historical experience data of UAVs to construct a three-dimensional path library that retains the effective experience of actual UAV missions and incorporates the optimization characteristics of natural flight, thereby improving the natural adaptability and environmental matching of path planning and providing reliable path support for collaborative UAV swarm missions in complex terrain.

[0164] To further improve the heading accuracy of UAVs in complex geomagnetic environments, in some embodiments, step 103: calculating the geomagnetic heading angle deviation based on the magnetic field gradient data includes:

[0165] Step 601: Extract the magnetic field strength component of the current measurement point from the magnetic field gradient data.

[0166] In step 601, the current measurement point refers to the location of the UAV cluster during real-time flight in the simulated training environment. The magnetic field data of this location point is measured in real time by the geomagnetic induction module installed on the UAV. The magnetic field intensity component refers to the measured value of the magnetic field intensity at a certain point in space in three mutually perpendicular directions, reflecting the spatial distribution characteristics of the magnetic field at that point.

[0167] In this embodiment of the application, the system reads the raw data collected by the geomagnetic sensor at the current measurement location and extracts the magnetic field strength values ​​in three directions. These data reflect the actual magnetic field conditions at the location affected by the terrain and environment.

[0168] Step 602: Determine the theoretical magnetic field direction of the current measurement point according to the preset geomagnetic reference field model.

[0169] In step 602, the theoretical magnetic field direction refers to the magnetic field direction vector that should exist at this geographical location, calculated according to the standard geomagnetic model, and is used as a heading reference.

[0170] In this embodiment of the application, the system queries a preset geomagnetic reference field model database based on the latitude and longitude coordinates of the current measurement point to obtain the theoretical magnetic field direction data of the location, which is then used for comparative analysis with the actual measurement values.

[0171] Step 603: Calculate the directional deviation between the magnetic field intensity component and the theoretical magnetic field direction.

[0172] In step 603, the directional deviation refers to the spatial angular difference between the measured magnetic field direction and the theoretical magnetic field direction.

[0173] In this embodiment of the application, the system combines the measured three directional magnetic field components into a spatial vector, and performs spatial angle calculation with the theoretical magnetic field direction vector to obtain the degree of directional deviation between the two.

[0174] Step 604: Convert the directional deviation into an angular offset in the heading coordinate system, and use the angular offset as the geomagnetic heading angle deviation.

[0175] In step 604, the heading coordinate system refers to a local coordinate system established based on the current heading direction of the UAV. Its source is a three-dimensional Cartesian coordinate system preset in the UAV navigation system, with the longitudinal axis of the UAV as the X-axis, the horizontal direction perpendicular to the longitudinal axis as the Y-axis, and the vertically downward direction as the Z-axis. This coordinate system is dynamically adjusted as the UAV's heading changes, and is used to convert geomagnetic direction deviation into angular offsets directly related to the UAV's flight control. The angular offset is the heading correction value converted from spatial direction deviation in this coordinate system.

[0176] In this embodiment of the application, the system converts the calculated directional deviation to the body coordinate system based on the current attitude of the UAV, and outputs an angle correction amount that can be directly used for heading control.

[0177] Here is a specific example:

[0178] During a drone swarm mountain penetration training mission, when the swarm flew to a canyon area at an altitude of 2150 meters, the system initiated a geomagnetic heading deviation calculation process. First, the onboard geomagnetic sensor acquired the magnetic field strength components at the current location, measuring 45 units in the X direction, -12 units in the Y direction, and 38 units in the Z direction. These values ​​were obtained directly from the sensor. The system then looked up the theoretical magnetic field direction from a pre-set geomagnetic reference field model based on the geographic coordinates of the location, obtaining the standard values ​​of 48 units in the X direction, -8 units in the Y direction, and 40 units in the Z direction. Next, the system calculated the directional deviation between the measured and theoretical values ​​using the vector angle formula θ=arccos[(A·B) / (|A||B|)], where A represents the measured vector (45, -12, 38) and B represents the theoretical vector (48, -8, 40). The calculated spatial directional deviation was 8.2 degrees. Then, the system combines the current attitude data of the UAV's horizontal tilt angle of 3 degrees and pitch angle of 2 degrees, transforms the spatial deviation into the body coordinate system, and finally outputs a clockwise heading correction value of 8.2 degrees.

[0179] In this embodiment of the application, the method achieves real-time and accurate correction of the UAV's heading in complex electromagnetic environments by accurately measuring and calculating geomagnetic deviation, effectively overcoming navigation errors caused by terrain interference, and providing a reliable heading reference for the coordinated flight of UAV swarms in areas with strong interference.

[0180] To further improve the collaborative control accuracy of UAV swarms, in some embodiments, step 105: generating flight control commands based on the multi-target collaborative path and combined with the real-time status data of the UAV swarm includes:

[0181] Step 701: Extract the flight attitude of each UAV and the relative position of the enemy aircraft from the real-time status data.

[0182] In step 701, flight attitude refers to the current pitch, roll, and yaw angles of the UAV. The relative position of the enemy aircraft refers to the azimuth and distance of the enemy target relative to the UAV.

[0183] In this embodiment, the system acquires real-time flight attitude data through the attitude sensors of each UAV, and at the same time detects the azimuth and distance information of surrounding enemy aircraft through a simulated combat system, providing an environmental awareness basis for the generation of subsequent control commands.

[0184] Step 702: Analyze the coordinates and arrival time of each UAV at the next navigation point from the multi-target cooperative path.

[0185] In step 702, the next navigation point refers to the next critical path point that the UAV swarm is about to fly to according to the multi-target collaborative path, including its coordinate position and arrival time requirement; while the "current measurement point" refers to the location of the UAV when the geomagnetic induction module collects magnetic field data in real time; the relationship between the two is that the current measurement point is the reference position for calculating the geomagnetic heading angle deviation, while the next navigation point is the expected arrival position planned based on the path after the deviation is corrected. The next navigation point is generated only after the current measurement point is determined, and the two constitute a continuous navigation position sequence. The arrival time limit refers to the time requirement for completing this segment of flight.

[0186] In this embodiment of the application, the system parses the cooperative path data and extracts the target point location and the specified arrival time that each UAV needs to reach in the next stage. These parameters are derived from the previous path planning results to ensure the synchronization and coordination of the swarm flight.

[0187] Step 703: For each UAV, integrate the coordinates, arrival time limit, flight attitude, and relative position of the enemy aircraft to calculate the roll angle control parameters and angle of attack control parameters.

[0188] In step 703, the roll angle control parameter refers to the angle value that controls the left and right tilt of the UAV. The angle of attack control parameter refers to the angle value that controls the up and down pitch of the UAV.

[0189] In this embodiment of the application, the system comprehensively considers factors such as the target location, remaining time, current attitude, and enemy threat for each UAV, and calculates the optimal roll and pitch adjustment through a kinematic model to ensure that the UAV can arrive at the target point on time and safely.

[0190] Step 704: Encapsulate the roll angle control parameters and angle of attack control parameters of all UAVs into flight control commands.

[0191] In this embodiment, the system encodes and encapsulates the calculated control parameters of each UAV according to a predetermined protocol format to generate a digital instruction set that can be received and executed by the UAV, thereby realizing the coordinated control of the cluster.

[0192] Here is a specific example:

[0193] In a drone swarm penetration training exercise in a canyon, when the swarm flew along the planned path to a key area, the system began generating flight control commands. First, the real-time status of the three drones was acquired: Drone 1 had a horizontal tilt angle of 3 degrees and a pitch angle of 2 degrees, and detected an enemy aircraft 300 meters to its northeast; Drone 2 had a horizontal tilt angle of 1 degree and a pitch angle of 1 degree, and the enemy aircraft was 400 meters away; Drone 3 had a horizontal tilt angle of 0 degrees and a pitch angle of 3 degrees, and the enemy aircraft was 350 meters away. The system analyzed the previously generated multi-target cooperative path and determined that the next waypoint for each drone was X degrees east longitude, Y degrees north latitude, and an altitude of 2153 meters, with a uniform arrival time limit of 30 seconds. For Drone 1, the system comprehensively considered its current attitude, enemy distance, and arrival time limit, and used a formula... The calculation shows that a roll to the left of 18 degrees is required, where v is the current speed of 25 m / s and g is the acceleration due to gravity. R represents the turning radius of the path, 53 meters; K is the adjustment coefficient, 0.6; θ is the enemy threat weight angle, 10 degrees; the angle of attack is adjusted to 4 degrees. Aircraft No. 2, due to the greater distance of the enemy aircraft, only requires basic adjustments, calculated to have a roll of 12 degrees and an angle of attack of 3 degrees; aircraft No. 3 has a roll of 15 degrees and an angle of attack of 5 degrees. The system converts these parameters into binary commands. Aircraft No. 1 is instructed to focus on enhancing evasive maneuvers, aircraft No. 2 to maintain standard flight, and aircraft No. 3 to make moderate adjustments.

[0194] In this embodiment, the method generates precise cluster control commands by fusing path planning and environmental perception data in real time, enabling coordinated maneuvering and precise control of UAV swarms in complex environments, effectively improving the success rate and safety of penetration missions.

[0195] Figure 2 This is a schematic diagram of the structure of a multi-agent cooperative UAV swarm simulation training system provided in an embodiment of this application, as shown below. Figure 2 As shown, the system includes:

[0196] The data acquisition module 21 is used to collect data on the biological movement of pigeon flocks, the cluster trajectory data of drone swarms when performing historical penetration missions, and the magnetic field gradient data of the simulated environment.

[0197] Module 22 is used to construct a three-dimensional path library based on the cluster trajectory data and the pigeon flock biological movement data.

[0198] The calculation module 23 is used to calculate the geomagnetic heading angle deviation based on the magnetic field gradient data.

[0199] The first generation module 24 is used to generate a multi-objective collaborative path based on the geomagnetic heading angle deviation and the three-dimensional path library, combined with a swarm intelligence optimization model.

[0200] The second generation module 25 is used to generate flight control commands based on the multi-target cooperative path and combined with the real-time status data of the UAV cluster.

[0201] The loading module 26 is used to load the flight control commands into the simulation model of the UAV cluster, driving the UAV cluster to conduct collaborative penetration training in a simulated environment.

[0202] Figure 2 The aforementioned multi-agent collaborative UAV swarm simulation training system can execute... Figure 1 The implementation principle and technical effects of the multi-agent collaborative UAV swarm simulation training method described in the illustrated embodiment will not be repeated here. The specific methods by which each module and unit performs operations in the multi-agent collaborative UAV swarm simulation training system described in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0203] In one possible design, Figure 2 The multi-agent collaborative UAV swarm simulation training system of the embodiment shown can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0204] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0205] The processing component 32 is used to perform the above. Figure 1 The embodiment describes a multi-agent collaborative drone swarm simulation training method.

[0206] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0207] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented from any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0208] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0209] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0210] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0211] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0212] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The embodiment shown is a multi-agent collaborative drone swarm simulation training method.

[0213] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0214] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0215] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0216] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for simulating and training multi-agent collaborative unmanned aerial vehicle (UAV) swarms, characterized in that, include: Collect data on the biological movement of pigeon flocks, the swarm trajectory data of drone swarms during historical infiltration missions, and the magnetic field gradient data of the simulated environment; Based on the cluster trajectory data and the pigeon flock biological movement data, a three-dimensional path library is constructed. Calculate the geomagnetic heading angle deviation based on the magnetic field gradient data; Based on the geomagnetic heading angle deviation and the three-dimensional path library, a multi-objective collaborative path is generated by combining a swarm intelligence optimization model. Based on the multi-target cooperative path and combined with the real-time status data of the UAV swarm, flight control commands are generated. The flight control commands are loaded into the simulation model of the UAV swarm, driving the UAV swarm to conduct collaborative penetration training in a simulated environment; The generation of a multi-objective cooperative path based on the geomagnetic heading angle deviation and the three-dimensional path library, combined with a swarm intelligence optimization model, includes: Retrieve the set of path points that match the current terrain from the 3D path library; Based on the geomagnetic heading angle deviation, correct the heading coordinates of each path point in the path point set; Identify terrain feature change points in the corrected path point set, and determine the avoidance type identifier based on the terrain feature change points; The avoidance type identifier and the modified path point set are input into the swarm intelligence optimization model, and the terrain avoidance constraint path segment corresponding to the avoidance type identifier is output. All the aforementioned terrain avoidance constraint path segments are spliced ​​together to form a multi-objective collaborative path; The step of inputting the avoidance type identifier and the corrected path point set into the swarm intelligence optimization model, and outputting the terrain avoidance constraint path segment corresponding to the avoidance type identifier, includes: The avoidance type identifier and the corrected path point set are input into the swarm intelligence optimization model. Based on the avoidance type identifier, the corresponding terrain spatial constraint relationship is called from the preset constraint rule base. The terrain spatial constraint relationship includes obstacle spacing threshold and height safety margin. The modified path point set is used to perform path feasibility verification according to the terrain spatial constraint relationship, and conflicting path points that are less than the obstacle spacing threshold or less than the height safety margin are eliminated. Based on the degree of association between the avoidance type identifier and the task, a corresponding first priority value is assigned to the avoidance type identifier; Based on the first priority value, the non-conflict path points in the verified path point set are reassembled into a sequence to generate a terrain avoidance constraint path segment.

2. The multi-agent collaborative UAV swarm simulation training method according to claim 1, characterized in that, The step of reorganizing the non-conflicting path points in the verified path point set based on the first priority value to generate terrain avoidance constraint path segments includes: Based on a preset mapping table between the first priority value range and the urgency level, the urgency level identifier corresponding to the first priority value is determined. Based on the first priority value and the terrain adaptability score, a corresponding second priority value is assigned to each non-conflict path point. According to the sorting rules corresponding to the urgency indicators, the second priority values ​​of all non-conflicting path points are sorted to generate an ordered path point sequence. The adjacent non-conflicting path points in the ordered path point sequence are smoothed by interpolation to form a continuous flight trajectory. Apply terrain constraints corresponding to the avoidance type identifier to the continuous flight trajectory to generate terrain avoidance constraint path segments.

3. The multi-agent collaborative UAV swarm simulation training method according to claim 1, characterized in that, The construction of a three-dimensional path library based on the cluster trajectory data and the pigeon flock biological movement data includes: The cluster trajectory data is classified according to terrain type into canyon crossing trajectory, climbing trajectory, and diving trajectory. Extract flight turning features and altitude change features from the pigeon flock's biological movement data; The flight turning features are mapped to the canyon crossing trajectory, and the altitude change features are mapped to the climb trajectory and the dive trajectory; Based on the mapping results, corresponding terrain-adaptive path units are generated, and all the terrain-adaptive path units are aggregated to form a three-dimensional path library.

4. The multi-agent cooperative UAV swarm simulation training method according to claim 1, characterized in that, The step of calculating the geomagnetic heading angle deviation based on the magnetic field gradient data includes: Extract the magnetic field strength component of the current measurement point from the magnetic field gradient data; Based on the preset geomagnetic reference field model, the theoretical magnetic field direction of the current measurement point is determined; Calculate the directional deviation between the magnetic field intensity component and the theoretical magnetic field direction; The directional deviation is converted into an angular offset in the heading coordinate system, and the angular offset is used as the geomagnetic heading angle deviation.

5. The multi-agent collaborative UAV swarm simulation training method according to claim 1, characterized in that, The step of generating flight control commands based on the multi-target cooperative path and combined with the real-time status data of the UAV swarm includes: The flight attitude of each UAV and the relative position of the enemy aircraft are extracted from the real-time status data; The coordinates and arrival time of each UAV at the next navigation point are analyzed from the multi-target cooperative path; For each UAV, the roll angle control parameters and angle of attack control parameters are calculated by integrating the coordinates, arrival time limit, flight attitude, and relative position of the enemy aircraft. All the roll angle control parameters and angle of attack control parameters of the UAV are encapsulated into flight control commands.

6. A multi-agent collaborative unmanned aerial vehicle (UAV) swarm simulation training system, characterized in that, include: The data acquisition module is used to collect data on the biological movement of pigeon flocks, the swarm trajectory data of drone clusters when performing historical infiltration missions, and the magnetic field gradient data of the simulated environment. The construction module is used to construct a three-dimensional path library based on the cluster trajectory data and the pigeon flock biological movement data; The calculation module is used to calculate the geomagnetic heading angle deviation based on the magnetic field gradient data; The first generation module is used to generate a multi-objective collaborative path based on the geomagnetic heading angle deviation and the three-dimensional path library, combined with a swarm intelligence optimization model. The second generation module is used to generate flight control commands based on the multi-target cooperative path and combined with the real-time status data of the UAV cluster. The loading module is used to load the flight control commands into the simulation model of the UAV cluster, driving the UAV cluster to conduct collaborative penetration training in a simulated environment; The generation of a multi-objective cooperative path based on the geomagnetic heading angle deviation and the three-dimensional path library, combined with a swarm intelligence optimization model, includes: Retrieve the set of path points that match the current terrain from the 3D path library; Based on the geomagnetic heading angle deviation, correct the heading coordinates of each path point in the path point set; Identify terrain feature change points in the corrected path point set, and determine the avoidance type identifier based on the terrain feature change points; The avoidance type identifier and the modified path point set are input into the swarm intelligence optimization model, and the terrain avoidance constraint path segment corresponding to the avoidance type identifier is output. All the aforementioned terrain avoidance constraint path segments are spliced ​​together to form a multi-objective collaborative path; The step of inputting the avoidance type identifier and the corrected path point set into the swarm intelligence optimization model, and outputting the terrain avoidance constraint path segment corresponding to the avoidance type identifier, includes: The avoidance type identifier and the corrected path point set are input into the swarm intelligence optimization model. Based on the avoidance type identifier, the corresponding terrain spatial constraint relationship is called from the preset constraint rule base. The terrain spatial constraint relationship includes obstacle spacing threshold and height safety margin. The modified path point set is used to perform path feasibility verification according to the terrain spatial constraint relationship, and conflicting path points that are less than the obstacle spacing threshold or less than the height safety margin are eliminated. Based on the degree of association between the avoidance type identifier and the task, a corresponding first priority value is assigned to the avoidance type identifier; Based on the first priority value, the non-conflict path points in the verified path point set are reassembled into a sequence to generate a terrain avoidance constraint path segment.

7. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement a multi-agent collaborative UAV swarm simulation training method as described in any one of claims 1-5.

8. A computer storage medium, characterized in that, The system contains a computer program that, when executed by a computer, implements a multi-agent collaborative unmanned aerial vehicle (UAV) swarm simulation training method as described in any one of claims 1-5.

Images