Method and system for optimizing drone swarm formation cooperation and task allocation

Abstract

The application provides a UAV group formation cooperation and task allocation optimization method and system, relates to the technical field of UAV control, and comprises the following steps: acquiring UAV flight states and task area information, dividing sub-tasks and setting initial allocation parameters, constructing an information interaction mechanism based on communication weights, minimizing total task execution cost to obtain an optimized allocation scheme, calculating an expected route to generate initial formation positions, collecting environmental parameters to calculate optimal distances, dynamically adjusting relative positions and compensating for aerodynamic interference, and monitoring obstacles in real time and updating formation schemes.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control technology, and in particular to a method and system for optimizing UAV swarm formation coordination and task allocation. Background Technology

[0002] Unmanned aerial vehicle (UAV) swarms can efficiently complete complex reconnaissance, surveillance, search and rescue missions through multi-platform collaboration, offering advantages that are unmatched by single-unit UAVs. UAV swarm formation collaboration technology, as the key to achieving efficient collaboration among multiple UAV systems, has become a current research hotspot. UAV swarm collaborative task allocation involves how to rationally decompose complex tasks and assign them to different UAVs for execution, while formation flight focuses on how to maintain the spatial positional relationship between UAVs to ensure maximum overall efficiency.

[0003] In traditional UAV swarm collaboration technology, task allocation usually adopts a centralized control method, with the ground control station uniformly planning and allocating tasks. Each UAV executes according to a preset trajectory. Formation control is mostly achieved using methods such as leader-follower models or virtual structure methods, maintaining the formation through preset relative positional relationships.

[0004] However, existing UAV formation collaborative control technologies still have problems such as a lack of sufficient consideration for the quality of communication between UAVs, which can easily lead to a decrease in mission collaboration efficiency or even mission failure; ignoring the impact of aerodynamic interference between UAVs, which can lead to formation instability or even safety accidents; and lacking the ability to dynamically adjust formation structure and task allocation based on real-time environmental parameters, making it difficult to cope with complex and ever-changing mission environments.

[0005] Therefore, a solution is urgently needed to address the problems existing in the current technology. Summary of the Invention

[0006] This invention provides a method and system for optimizing unmanned aerial vehicle (UAV) swarm formation coordination and task allocation, which can at least solve some of the problems existing in the prior art.

[0007] A first aspect of the present invention provides a method for optimizing unmanned aerial vehicle (UAV) swarm formation coordination and task allocation, comprising:

[0008] The flight status information and mission area information of the drones in the drone swarm are obtained, and the initial spatial distribution of the drone swarm is determined based on the flight status information.

[0009] The task area information is divided into multiple sub-tasks and initial task allocation parameters are set, communication weights of each UAV are determined according to communication quality and relative distance between UAVs, an information interaction mechanism between UAVs is constructed based on the communication weights, execution costs of each UAV for executing sub-tasks are calculated based on the initial task allocation parameters, and an optimized task allocation scheme is obtained by minimizing total task execution costs of all UAVs;

[0010] Expected flight routes of each UAV for executing allocated tasks are calculated according to the optimized task allocation scheme, and initial formation position information is generated;

[0011] Environment parameters in a flight process of the UAV group are collected based on the initial formation position information, optimal distances between UAVs are calculated based on the environment parameters, relative positions of the UAVs are dynamically adjusted according to the optimal distances, aerodynamic interference effects are predicted based on relative motions between the UAVs and are compensated in real time, and an optimized formation scheme is obtained;

[0012] The optimized formation scheme is sent to the UAV group for execution, obstacle information in a flight process is monitored in real time, if an obstacle is monitored, an obstacle avoidance strategy is calculated and the optimized formation scheme is updated until a task is completed.

[0013] In an optional implementation,

[0014] Flight state information and task area information of UAVs in the UAV group are acquired, and initial spatial distribution of the UAV group is determined based on the flight state information, including:

[0015] Flight state information of each UAV in the UAV group is acquired, and the flight state information includes current position coordinates, flight speed and remaining power of each UAV;

[0016] Task area information is acquired, and the task area information includes spatial range of a task execution area and position coordinates of a task target point;

[0017] Spatial distances between each UAV and adjacent UAVs are calculated, when the spatial distance is less than a preset safety threshold, a safety distance between adjacent UAVs is maintained by adjusting flight height or horizontal position of the UAV, and initial spatial distribution of the UAV group is obtained.

[0018] In an optional implementation,

[0019] The task area information is divided into multiple sub-tasks and initial task allocation parameters are set, communication weights of each unmanned aerial vehicle are determined according to communication quality and relative distance between unmanned aerial vehicles, an information interaction mechanism between unmanned aerial vehicles is constructed based on the communication weights, execution costs of unmanned aerial vehicles for each sub-task are calculated based on the initial task allocation parameters, and an optimized task allocation scheme is obtained by minimizing total task execution costs of all unmanned aerial vehicles, including:

[0020] The task area information is divided into multiple sub-tasks, and initial task allocation parameters are set. The task area is divided into multiple sub-task areas by an adaptive grid division method, and a task feature vector is determined for each sub-task area. The task feature vector includes position coordinates of the sub-task, a task difficulty coefficient, and a task weight.

[0021] Communication quality and relative distance between unmanned aerial vehicles are obtained, and communication weights of each unmanned aerial vehicle are determined based on the communication quality and the relative distance. The communication weights are calculated by an exponential decay function and a product of the relative distance and the communication quality, and a communication weight matrix is constructed.

[0022] An information interaction mechanism between unmanned aerial vehicles is constructed based on the communication weights, execution costs of unmanned aerial vehicles for each sub-task are calculated based on the initial task allocation parameters, and the execution costs are calculated by a weighted sum of energy consumption, task execution time, and task load degree.

[0023] A benefit matrix is constructed based on the communication weight matrix, the information interaction mechanism, and the execution costs. An optimized task allocation scheme is obtained by iterative solution of the Hungarian algorithm under the conditions of meeting each sub-task being allocated to only one unmanned aerial vehicle, a task quantity upper limit of a single unmanned aerial vehicle, and a remaining energy constraint of the unmanned aerial vehicle.

[0024] In an optional embodiment,

[0025] A benefit matrix is constructed based on the communication weight matrix, the information interaction mechanism, and the execution costs. An optimized task allocation scheme is obtained by iterative solution of the Hungarian algorithm under the conditions of meeting each sub-task being allocated to only one unmanned aerial vehicle, a task quantity upper limit of a single unmanned aerial vehicle, and a remaining energy constraint of the unmanned aerial vehicle, including:

[0026] Communication weight coefficients and task execution cost coefficients are set according to flight tasks of the unmanned aerial vehicle group, a communication influence coefficient is generated based on an information interaction mechanism, elements in the communication weight matrix are multiplied by the communication weight coefficients and the communication influence coefficient, the execution costs are multiplied by the task execution cost coefficients, and a benefit matrix is obtained by addition.

[0027] The unique constraint data of determining that each subtask can be assigned to only one unmanned aerial vehicle, the maximum task carrying quantity constraint data of each unmanned aerial vehicle, the residual energy constraint data, the communication quality evaluation data obtained based on the information interaction mechanism between unmanned aerial vehicles, the weight increment data of adjacent unmanned aerial vehicles calculated according to the communication weight change rate of adjacent unmanned aerial vehicles, the state information update data obtained by multiplying the communication quality evaluation data and the weight increment data, and accumulating them;

[0028] Subtracting the minimum value of each row in the benefit matrix from the current row, subtracting the minimum value of each column in the matrix after subtracting the row minimum value from the current column, determining independent zero elements in the modified benefit matrix and covering all zero elements with the minimum number of cover lines, when the number of cover lines is less than the order of the matrix, subtracting the minimum value of the uncovered elements from the uncovered elements, adding the minimum value of the uncovered elements to the elements covered twice, and repeating the execution to obtain an optimization task allocation scheme that meets the constraint data.

[0029] In an optional implementation,

[0030] According to the optimization task allocation scheme, the expected flight route of each unmanned aerial vehicle for executing the allocated task is calculated, and initial formation position information is generated, including:

[0031] The task execution sequence of each unmanned aerial vehicle in the optimization task allocation scheme is obtained, and the shortest flight path of each unmanned aerial vehicle is calculated based on the task execution sequence using an ant colony algorithm, wherein the shortest flight path meets the task timing constraint and the unmanned aerial vehicle turning radius constraint;

[0032] According to the shortest flight path, the flight time of each unmanned aerial vehicle to reach the first task execution point is calculated, the unmanned aerial vehicle with the longest flight time is determined as a key unmanned aerial vehicle, and the delay takeoff time of other unmanned aerial vehicles is calculated based on the flight time of the key unmanned aerial vehicle;

[0033] Based on the delay takeoff time, the initial position coordinates of each unmanned aerial vehicle are determined, and initial formation position information that meets the minimum spacing constraint of unmanned aerial vehicles is constructed.

[0034] In an optional implementation,

[0035] Based on the initial formation position information, environmental parameters in the flight process of the unmanned aerial vehicle group are collected, and the optimal spacing between unmanned aerial vehicles is calculated based on the environmental parameters, including:

[0036] Based on the initial formation position information, a topological space of the unmanned aerial vehicle group is constructed, a space connection structure is constructed according to the relative distance relationship between unmanned aerial vehicles, and a topological invariant is calculated to obtain the environmental parameters;

[0037] The stability of the environment parameters is analyzed by using topological persistence map analysis, and a persistence measure value is calculated, a topological complexity evaluation index is constructed, and a local topological complexity corresponding to each unmanned aerial vehicle is determined, aerodynamic energy consumption generated by airflow interference between unmanned aerial vehicles is calculated by flight dynamics method, and a communication cost is calculated based on unmanned aerial vehicle communication link quality and information transmission demand, the topological complexity evaluation index, the aerodynamic energy consumption and the communication cost are taken as optimization objectives, and an optimal distance between unmanned aerial vehicles is solved.

[0038] In an optional implementation,

[0039] The relative positions of the unmanned aerial vehicles are dynamically adjusted according to the optimal distance, aerodynamic interference is predicted based on relative motion between the unmanned aerial vehicles and is compensated in real time, and an optimized formation scheme is obtained.

[0040] The distance error between adjacent unmanned aerial vehicles is calculated based on the optimal distance, the motion speed and direction of the unmanned aerial vehicles are determined according to the distance error, the motion speed and direction are corrected based on the change trend of the local topological complexity, the position adjustment instruction of the unmanned aerial vehicles is obtained and executed;

[0041] During execution of the position adjustment instruction, the evolution process of the formation configuration is analyzed by topological persistence analysis of the environment parameters, stable topological structure features are extracted, topological invariant feature values and persistence measure feature values are calculated according to the stable topological structure features, the motion trend between adjacent unmanned aerial vehicles is analyzed based on the feature values, and aerodynamic interference between the unmanned aerial vehicles is predicted.

[0042] The aerodynamic compensation control amount is calculated according to the aerodynamic interference, the tracking compensation control amount is calculated according to the tracking error between the actual position and the target position of the unmanned aerial vehicles, the aerodynamic compensation control amount and the tracking compensation control amount are taken as the control input of the unmanned aerial vehicles, and an optimized formation scheme is obtained.

[0043] In an optional implementation,

[0044] The optimized formation scheme is sent to the unmanned aerial vehicle group for execution, obstacle information in the flight process is monitored in real time, if an obstacle is monitored, an obstacle avoidance strategy is calculated and the optimized formation scheme is updated, and the task is completed.

[0045] The optimized formation scheme is sent to the unmanned aerial vehicle group, and the unmanned aerial vehicle group flies in formation according to the optimized formation scheme;

[0046] Obstacle information is monitored in real time during the formation flight, obstacle position coordinates are obtained according to the obstacle information, if an obstacle is detected, a safe flight path of each unmanned aerial vehicle is calculated based on the obstacle position coordinates, and the minimum distance between the safe flight path and the obstacle is greater than a preset safety threshold.

[0047] According to the safe flight path, an updated optimal formation scheme is generated, the updated optimal formation scheme including a flight trajectory of the UAV flying around the obstacle and a reconstruction formation strategy after obstacle avoidance, and the updated optimal formation scheme is sent to the UAV group for execution until the task is completed.

[0048] In a second aspect, the embodiment of the present application provides a UAV group formation coordination and task allocation optimization system, comprising:

[0049] A first unit is configured to acquire flight state information of UAVs in a UAV group and task area information, determine an initial spatial distribution of the UAV group based on the flight state information;

[0050] A second unit is configured to divide the task area information into multiple subtasks and set initial task allocation parameters, determine communication weights of the UAVs according to communication quality and relative distances between the UAVs, construct an information interaction mechanism between the UAVs based on the communication weights, calculate execution costs of the UAVs for executing the subtasks based on the initial task allocation parameters, and obtain an optimal task allocation scheme by minimizing total task execution costs of all the UAVs;

[0051] A third unit is configured to calculate expected flight paths of each UAV for executing allocated tasks according to the optimal task allocation scheme, and generate initial formation position information;

[0052] A fourth unit is configured to acquire environmental parameters in a flight process of the UAV group based on the initial formation position information, calculate optimal distances between the UAVs based on the environmental parameters, dynamically adjust relative positions of the UAVs according to the optimal distances, predict aerodynamic interference effects based on relative motions between the UAVs and perform real-time compensation, and obtain an optimal formation scheme;

[0053] A fifth unit is configured to send the optimal formation scheme to the UAV group for execution, monitor obstacle information in the flight process in real time, calculate an obstacle avoidance strategy and update the optimal formation scheme if the obstacle information is monitored, and execute until the task is completed.

[0054] In a third aspect, the embodiment of the present application provides a computer readable storage medium having computer program instructions stored thereon, the computer program instructions being executed by a processor to implement the method.

[0055] In the present application, by constructing an information interaction mechanism based on the flight state information of the unmanned aerial vehicle and the communication weight, the accurate optimization of task allocation is realized, the overall task execution cost is reduced, the resource utilization efficiency is improved, the optimal distance between unmanned aerial vehicles is dynamically adjusted according to the environmental parameters and the aerodynamic interference is compensated in real time, the formation flight is more stable and reliable, the energy loss in the formation process is effectively reduced, the endurance time of the unmanned aerial vehicle is prolonged, the real-time monitoring and dynamic obstacle avoidance strategy are combined, the adaptability and task completion rate of the unmanned aerial vehicle group in the complex environment are significantly improved, and the flight safety is ensured, which provides efficient and reliable technical support for the cooperative task execution of multiple unmanned aerial vehicles. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 A flowchart of the unmanned aerial vehicle group formation cooperation and task allocation optimization method of the embodiment of the present application is shown in

[0057] Figure 2 A flowchart of the unmanned aerial vehicle task allocation optimization of the unmanned aerial vehicle group formation cooperation and task allocation optimization method of the embodiment of the present application is shown in

[0058] Figure 3 A comparison chart of the unmanned aerial vehicle formation control efficiency of the unmanned aerial vehicle group formation cooperation and task allocation optimization method of the embodiment of the present application is shown in DETAILED DESCRIPTION

[0059] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be described clearly and completely below in combination with the drawings of the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0060] The technical scheme of the present application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in some embodiments.

[0061] Figure 1 A flowchart of the unmanned aerial vehicle group formation cooperation and task allocation optimization method of the embodiment of the present application is shown in Figure 1 As shown in the figure, the method comprises:

[0062] Obtaining the flight state information and task area information of the unmanned aerial vehicles in the unmanned aerial vehicle group, determining the initial spatial distribution of the unmanned aerial vehicle group based on the flight state information;

[0063] The task area information is divided into multiple sub-tasks and initial task allocation parameters are set, communication weights of each UAV are determined according to communication quality and relative distance between UAVs, an information interaction mechanism between UAVs is constructed based on the communication weights, execution costs of each UAV for executing sub-tasks are calculated based on the initial task allocation parameters, and an optimized task allocation scheme is obtained by minimizing total task execution costs of all UAVs;

[0064] According to the optimized task allocation scheme, an expected flight route of each UAV for executing allocated tasks is calculated, and initial formation position information is generated;

[0065] Based on the initial formation position information, environmental parameters in a flight process of the UAV group are collected, optimal distances between UAVs are calculated based on the environmental parameters, relative positions of the UAVs are dynamically adjusted according to the optimal distances, aerodynamic interference effects are predicted based on relative motions between the UAVs and are compensated in real time, and an optimized formation scheme is obtained;

[0066] The optimized formation scheme is sent to the UAV group for execution, obstacle information in a flight process is monitored in real time, if an obstacle is monitored, an obstacle avoidance strategy is calculated and the optimized formation scheme is updated until a task is completed.

[0067] In an optional implementation,

[0068] Flight state information and task area information of UAVs in a UAV group are acquired, and initial spatial distribution of the UAV group is determined based on the flight state information, including:

[0069] Flight state information of each UAV in the UAV group is acquired, and the flight state information includes current position coordinates, flight speed and remaining power of each UAV;

[0070] Task area information is acquired, and the task area information includes spatial range of a task execution area and position coordinates of a task target point;

[0071] Spatial distances between each UAV and adjacent UAVs are calculated, when the spatial distance is less than a preset safety threshold, a safety distance between adjacent UAVs is maintained by adjusting flight height or horizontal position of the UAV, and initial spatial distribution of the UAV group is obtained.

[0072] The flight state information of each UAV in the UAV fleet is obtained, and real-time data is received from each UAV through a wireless communication network, including the current three-dimensional position coordinates (x, y, z) of the UAV, where x and y represent horizontal coordinates and z represents height; flight speed, including horizontal and vertical speed components; and remaining battery percentage. For example, for a UAV numbered UAV001, its position coordinates can be (120.5 meters, 85.3 meters, 50 meters), flight speed (5 meters / second, 3 meters / second, 0 meters / second), and remaining battery percentage 75%, updated at a frequency of 10 times per second, ensuring real-time grasp of the dynamic state of the UAV.

[0073] Task area information is obtained. The task area information is usually set by the ground control station in advance and sent to all UAVs, including the spatial range of the task execution area, usually represented as a three-dimensional cube or polygonal region, such as region boundary coordinates {(0, 0, 0), (1000, 0, 0), (1000, 1000, 0), (0, 1000, 0)}, with a maximum height limit of 120 meters; and the position coordinates of the task target points, such as {(250, 300, 30), (550, 600, 50), (750, 400, 60)} and a series of specific positions that the UAV needs to reach or monitor, the task area information may also include no-fly zones such as the positions of buildings, high-voltage lines or other obstacles to ensure the safe flight of the UAV.

[0074] After obtaining all the necessary information, the spatial distance between each UAV and adjacent UAVs is calculated. For each pair of UAVs i and j in the UAV fleet, the three-dimensional Euclidean distance between them is calculated. Specifically, assuming the position of UAV i is (xi, yi, zi) and the position of UAV j is (xj, yj, zj), the spatial distance between them can be obtained by calculating the straight-line distance between the two points. For example, if the position of UAV UAV001 is (120.5 meters, 85.3 meters, 50 meters) and the position of UAV UAV002 is (125.5 meters, 90.3 meters, 52 meters), the spatial distance between them is approximately 7.4 meters.

[0075] The calculated spatial distance is compared with the preset safety threshold, and the safety threshold is set to 10 meters as an example. The safety threshold can be adjusted according to the size of the UAV, flight speed and task requirements. When the distance between any two UAVs is less than the safety threshold, the adjustment program is started.

[0076] The adjustment procedure evaluates the feasibility of adjusting the flight height, checks the current height of the UAVs, the maximum allowed flight height, and the remaining power, and determines whether there are sufficient resources for vertical adjustment. For example, for UAVs with low remaining power (e.g., less than 30%), horizontal adjustment is preferred to save energy. If height adjustment is possible, different flight heights are assigned to adjacent UAVs to achieve a vertical separation of at least 5 meters. For example, the height of UAV001 is adjusted to 45 meters, and the height of UAV002 is adjusted to 55 meters.

[0077] If vertical adjustment is not feasible or insufficient to maintain a safe distance, horizontal position adjustment is performed. Horizontal adjustment is based on the current positions of the UAVs, the flight direction, and the task target point, and calculates the optimal adjustment direction and distance, taking into account multiple factors, including avoiding new conflicts with other UAVs, maintaining proximity to the task target, and minimizing energy consumption. For example, UAV001 is instructed to shift 3 meters west and 2 meters south, while the position of UAV002 remains unchanged, and the distance between them increases to about 12 meters, exceeding the safety threshold.

[0078] The adjustment process is iterative, and the distances between all UAVs are recalculated after each adjustment until all safety distance requirements are met. For complex situations, such as clusters formed by multiple UAVs, priority rules are used to determine the adjustment order, for example, adjusting UAVs with higher power or those farther from the task target first.

[0079] An initial spatial distribution of the UAV swarm is obtained, in which all UAVs maintain a safety distance of at least 10 meters. The initial spatial distribution serves as the basis for subsequent task planning and path generation, ensuring that the UAV swarm starts from a safe initial state when executing tasks.

[0080] In this embodiment, by obtaining the current position coordinates, flight speed, and remaining power of each UAV in the UAV swarm, the overall state of the UAV swarm can be grasped in real time, providing accurate basic data for subsequent spatial distribution optimization. By obtaining task area information, the spatial range of the task execution area and the position coordinates of the task target point are determined, ensuring that the spatial distribution of the UAV swarm meets the task execution requirements and improves task completion efficiency. By calculating the spatial distance between adjacent UAVs and comparing it with the preset safety threshold, risk areas where collisions may occur are identified in a timely manner, and adjacent UAVs are kept at a safe distance by adjusting their flight height or horizontal position, effectively avoiding collision risks between UAVs and improving the flight safety of the UAV swarm.

[0081] In an alternative embodiment,

[0082] The task area information is divided into multiple sub-tasks and initial task allocation parameters are set, communication weights of each UAV are determined according to communication quality and relative distance between UAVs, an information interaction mechanism between UAVs is constructed based on the communication weights, execution costs of UAVs for each sub-task are calculated based on the initial task allocation parameters, and an optimized task allocation scheme is obtained by minimizing total task execution costs of all UAVs, including:

[0083] The task area information is divided into multiple sub-tasks, and initial task allocation parameters are set. The task area is divided into multiple sub-task areas by an adaptive grid division method, and a task feature vector is determined for each sub-task area. The task feature vector includes position coordinates of the sub-task, a task difficulty coefficient, and a task weight.

[0084] Communication quality and relative distance between UAVs are obtained, and communication weights of each UAV are determined based on the communication quality and the relative distance. The communication weights are calculated by a product of an exponential decay function and the relative distance and the communication quality. A communication weight matrix is constructed.

[0085] An information interaction mechanism between UAVs is constructed based on the communication weights, and execution costs of UAVs for each sub-task are calculated based on the initial task allocation parameters. The execution costs are calculated by a weighted sum of energy consumption, task execution time, and task load degree.

[0086] A benefit matrix is constructed based on the communication weight matrix, the information interaction mechanism, and the execution costs. An optimized task allocation scheme is obtained by Hungarian algorithm iteration under the conditions of meeting each sub-task being allocated to only one UAV, a single UAV task quantity upper limit, and a UAV remaining energy constraint.

[0087] The topographic map, obstacle distribution map and target point distribution map of the task area are obtained as basic data, and initial task allocation parameters are set, including area division granularity parameters, task difficulty evaluation factors and task weight calculation coefficients. The adaptive grid division method dynamically adjusts the grid size according to the complexity of the area, and determines the grid division precision by calculating the complexity indicators such as terrain change rate and obstacle density. When the area complexity indicator value is higher than the preset threshold, the current grid is subdivided into smaller subgrids; otherwise, the larger grid size is maintained. For example, in a forest resource monitoring task, the 4 square kilometer area is initially evenly divided into 16 1 square kilometer grids. For areas with dense vegetation and large terrain undulations, the grid is further subdivided into 0.25 square kilometers, while flat and open areas maintain the original size. The final area is divided into 28 subtask areas, with the number of grids in the complex terrain area significantly increasing. After completing the grid division, the task feature vector is calculated for each subtask area, including the center point coordinates of the area as the position coordinates, the task difficulty coefficient calculated according to the terrain complexity and obstacle distribution, and the task weight determined according to the monitoring importance. For example, the subtask feature vector of T5 is: position coordinates (2.35, 1.78), task difficulty coefficient 1.6, task weight 2.0; and the subtask feature vector of T12 is: position coordinates (3.42, 2.15), task difficulty coefficient 0.8, task weight 0.9.

[0088] The communication quality and relative distance between unmanned aerial vehicles are obtained, and a communication weight matrix is constructed. The communication quality is evaluated by multiple indicators, including signal strength, signal stability, data transmission rate and packet loss rate, to form a normalized score value between 0 and 1. The relative distance between unmanned aerial vehicles is calculated by their respective three-dimensional coordinates. The communication weight is calculated using an exponential decay function, i.e. the communication weight is equal to the communication quality multiplied by the distance decay factor, which decreases exponentially with increasing distance. In a five-vehicle formation flight task, the relative distance between unmanned aerial vehicles 1 and 2 is 800 meters, and the communication quality score is 0.9, resulting in a communication weight of 0.82 calculated by the exponential decay function; while the relative distance between unmanned aerial vehicles 1 and 5 is 2500 meters, and the communication quality score is 0.75, resulting in a communication weight of 0.32. By calculating the communication weight between all unmanned aerial vehicles, a 5x5 symmetric matrix is constructed as the communication weight matrix. The diagonal elements in the matrix (representing self-communication) are all 1, and the non-diagonal elements reflect the communication efficiency between different unmanned aerial vehicles.

[0089] An information exchange mechanism between UAVs is constructed based on the communication weight matrix, and a hierarchical data transmission strategy is realized. For UAV pairs with high communication weight (e.g., communication weight higher than 0.8), a high-frequency information exchange channel is established to share detailed task information and state data. For UAV pairs with medium communication weight (e.g., communication weight between 0.5 and 0.8), a medium-frequency information exchange channel is established to mainly share key state information. For UAV pairs with low communication weight (e.g., communication weight lower than 0.5), an emergency information exchange channel is established only when necessary. At the same time, the execution cost of each UAV for each subtask is calculated. The execution cost considers three main factors: energy consumption, task execution time, and task load degree. Energy consumption includes the flight energy consumption of the UAV from the current location to the subtask area and the energy consumption in the subtask area; task execution time includes flight arrival time and task completion time; task load degree reflects the occupation degree of the UAV's computing resources and sensor resources. The execution cost is calculated by the weighted sum of the three indexes, and the weight coefficients are pre-set according to the task characteristics. In the foregoing monitoring task, the flight energy consumption of UAV No. 1 for executing the T5 subtask is 12% of the battery capacity, the working energy consumption is 15%, and the total energy consumption is 27%; the flight time is 5 minutes, the task completion time is 15 minutes, and the total execution time is 20 minutes; the task load degree is 0.7. Assuming that the weights of energy consumption, task execution time, and task load degree are 0.4, 0.35, and 0.25 respectively, the execution cost of UAV No. 1 for executing the T5 subtask is calculated as 0.327. The execution cost of all UAVs for all subtasks is calculated in the same way.

[0090] An efficiency matrix is constructed based on the communication weight matrix, information interaction mechanism and execution cost. The efficiency matrix is a two-dimensional matrix of UAVs and sub-tasks, and the matrix elements represent the comprehensive efficiency value of a specific UAV executing a specific sub-task. The efficiency value calculation considers two factors: one is the conversion value of the execution cost, usually taking the negative value of the execution cost; the other is the synergy benefit gain, which is calculated according to the communication weight between the UAV and the UAV executing the adjacent sub-task. When two UAVs with high communication weight are assigned to adjacent sub-tasks, the overall efficiency will increase. In the monitoring task, the base efficiency of UAV 1 executing sub-task T5 is -0.327 (the negative value of the execution cost); if UAV 2 executes adjacent sub-task T6, and the communication weight between UAV 1 and UAV 2 is high (0.82), the efficiency of UAV 1 executing T5 will increase by 0.05, and the adjusted efficiency is -0.277. After constructing the complete efficiency matrix, the Hungarian algorithm is applied to solve the optimal task allocation scheme under the constraint conditions. The constraint conditions include: each sub-task can only be assigned to one UAV, the number of tasks of each UAV does not exceed the preset upper limit (such as 6), and the remaining energy of the UAV after executing the task is not lower than the safety threshold (such as 25%). The Hungarian algorithm finds the optimal match through row and column transformation and marking operation, and iterates multiple rounds until the optimal solution is found. In the monitoring task, the task allocation scheme finally output by the Hungarian algorithm is: UAV 1 is responsible for sub-tasks T5, T8, T15, T21 and T24, and the total energy consumption is expected to be 72% of the battery capacity, with 28% remaining; UAV 2 is responsible for sub-tasks T1, T6, T13, T18 and T25, and the total energy consumption is expected to be 69% of the battery capacity, with 31% remaining;

[0091] When the flight environment of the UAV changes or the task demand is updated, the task allocation scheme is dynamically adjusted. The change may include adding new task points, communication quality decreasing due to weather condition change, UAV energy state anomaly, etc. The adjustment strategy adopts the principle of local optimization, that is, the original allocation scheme is kept as much as possible, and only the affected part is redistributed, reducing the additional overhead brought by adjustment. For example, during the monitoring process, when UAV 3 encounters strong air flow, causing the energy consumption rate to increase and the remaining energy to rapidly decrease to the vicinity of the safety threshold, the energy required for the UAV to complete the remaining tasks is immediately calculated. It is found that completing all allocated tasks will cause the energy to be lower than the safety threshold, so the two lowest weight sub-tasks of this UAV are redistributed to UAVs 1 and 5, ensuring the successful completion of the task and the safe return of all UAVs.

[0092] In this embodiment, the task area is divided into multiple sub-task areas by an adaptive grid division method, and a task feature vector is introduced to describe the characteristics of each sub-task, realizing fine division of the task, providing accurate task feature information for subsequent task allocation, constructing a communication weight matrix based on communication quality and relative distance, using an exponential decay function to describe the decay characteristics of communication effect with distance, accurately reflecting the communication constraints between unmanned aerial vehicles, ensuring the communication reliability of the allocation scheme, calculating the execution cost by the weighted sum of energy consumption, task execution time and task load degree, and comprehensively considering multiple influencing factors to make the task allocation more reasonable.

[0093] In an optional implementation,

[0094] Based on the communication weight matrix, the information interaction mechanism and the execution cost, a benefit matrix is constructed, and under the conditions of meeting each sub-task being allocated to only one unmanned aerial vehicle, the upper limit of the number of tasks of a single unmanned aerial vehicle and the remaining energy constraint of the unmanned aerial vehicle, an optimized task allocation scheme is obtained by iterative solution of the Hungarian algorithm, including:

[0095] According to the flight task of the unmanned aerial vehicle group, a communication weight coefficient and a task execution cost coefficient are set, a communication influence coefficient is generated based on an information interaction mechanism, elements in the communication weight matrix are multiplied by the communication weight coefficient and the communication influence coefficient, the execution cost is multiplied by the task execution cost coefficient, and the benefit matrix is obtained by addition;

[0096] The uniqueness constraint data that each sub-task can be allocated to only one unmanned aerial vehicle, the maximum task carrying quantity constraint data of each unmanned aerial vehicle, and the remaining energy constraint data are determined, the communication quality evaluation data is obtained based on the information interaction mechanism between unmanned aerial vehicles, the weight increment data of adjacent unmanned aerial vehicles is calculated according to the communication weight change rate of adjacent unmanned aerial vehicles, the communication quality evaluation data is multiplied by the weight increment data and accumulated to obtain the state information update data;

[0097] Each row element in the benefit matrix is subtracted by the minimum value of the current row, each column element in the matrix after subtracting the minimum value is subtracted by the minimum value of the current column, the independent zero elements in the modified benefit matrix are determined and all zero elements are covered with the least number of covering lines, when the number of covering lines is less than the order of the matrix, the uncovered elements are subtracted by the minimum value in the uncovered elements, the elements covered twice are added by the minimum value in the uncovered elements, and the process is repeated to obtain the optimized task allocation scheme meeting the constraint data.

[0098] The communication weight coefficient and the task execution cost coefficient are determined according to the flight task nature of the UAV group, and the two coefficients respectively reflect the importance of communication efficiency and task execution efficiency in the overall task. The communication weight coefficient is usually set in the range of 0.3 to 0.7, and the value increases with the increase of the task cooperation demand; the task execution cost coefficient is set in the range of 0.3 to 0.7, and the value increases with the increase of the task timeliness requirement. In the forest fire monitoring task, since real-time information sharing is crucial to cooperative monitoring, the communication weight coefficient is set to 0.6, and the task execution cost coefficient is set to 0.4. The communication influence coefficient is generated based on the pre-established information interaction mechanism, and the communication influence coefficient describes the influence degree of the communication quality between UAVs on the overall task completion efficiency. The communication influence coefficient is obtained by analyzing the influence of the change of communication quality in the historical task on the task completion time and quality, and is set to 1.2 in the monitoring task. The elements in the communication weight matrix are multiplied by the communication weight coefficient and the communication influence coefficient respectively to obtain the adjusted communication weight value. For a 5-UAV formation, the original communication weight between UAV 1 and UAV 2 is 0.85, and the adjusted communication weight is 0.85x0.6x1.2=0.612. The adjusted communication weight of all UAV pairs is calculated. The execution cost is multiplied by the task execution cost coefficient to obtain the adjusted execution cost, such as the original execution cost of UAV 1 executing subtask T3 is 0.38, and the adjusted execution cost is 0.38x0.4=0.152. The adjusted communication weight and the adjusted execution cost are added to construct a complete benefit matrix. In the matrix, each row represents a UAV, and each column represents a subtask, and the matrix element value represents the comprehensive benefit of a specific UAV executing a specific subtask.

[0099] The task allocation optimization process needs to consider multiple constraint conditions, determine the uniqueness constraint data of each subtask that can only be assigned to one UAV, and represent it as each column in the task allocation matrix can only have one element as 1 and the rest as 0. Set the maximum task carrying capacity constraint data of each UAV, which is determined according to the differences in UAV models and performance. In the monitoring task, the maximum carrying capacity of a large fixed-wing UAV is 6 subtasks, and the maximum carrying capacity of a quadcopter UAV is 4 subtasks. The remaining energy constraint data is calculated according to the battery capacity of the UAV and the energy consumption of the task, and the remaining energy after the task is executed is required to be not less than the safety threshold, which is usually set to 25% of the total capacity. Based on the information exchange mechanism between UAVs, communication quality evaluation data is obtained, which reflects the real-time communication status, including signal strength, data transmission rate and connection stability. The communication quality evaluation adopts a 0-1 scoring system, with 1 representing the best communication state. The weight increment data of adjacent UAVs is calculated according to the communication weight change rate of adjacent UAVs. When two UAVs are assigned to adjacent task areas, the communication weight increment is the communication weight multiplied by the adjacent coefficient, which is determined according to the distance between task areas. In the monitoring task, the adjacent coefficient of adjacent areas is set to 0.15, the adjacent coefficient of non-adjacent but moderately distant areas is set to 0.08, and the adjacent coefficient of distant areas is 0. Multiply and accumulate the communication quality evaluation data and the weight increment data to obtain state information update data, which is used to adjust the decision basis in the task allocation process.

[0100] The Hungarian algorithm is applied to optimize the solution of the benefit matrix. Each row element in the benefit matrix is subtracted by the current row minimum value. The purpose of the operation is to make at least one zero element appear in each row, while keeping the relative size relationship unchanged. In the benefit matrix of a certain monitoring task, the benefit value of UAV No. 1 executing each subtask is [0.765, 0.612, 0.843, 0.528, 0.697], and the row minimum value is 0.528. After subtraction, [0.237, 0.084, 0.315, 0, 0.169] is obtained. Subtract the minimum value of the current column from each column element in the matrix after subtracting the row minimum value, so that at least one zero element appears in each column. Label the zero elements in the modified benefit matrix to determine the independent zero elements (i.e. zero elements not in the same row and column), and cover all zero elements with the minimum number of horizontal and vertical lines. When the number of covering lines is less than the order of the matrix, it indicates that the complete matching scheme has not been found, and further optimization is needed. Subtract the minimum value of the uncovered elements from the uncovered elements to generate new zero elements in the uncovered area. Add the minimum value of the uncovered elements to the elements covered twice (i.e. elements covered by horizontal and vertical lines at the same time) to maintain the relative relationship of the matrix. Repeat the process of labeling zero elements, determining covering lines, and adjusting uncovered elements until the number of covering lines equals the order of the matrix, at which time the optimal task allocation scheme that meets the constraint conditions is found.

[0101] Exemplarily, in the forest fire prevention monitoring task, the task allocation scheme obtained by the above method is: the first unmanned aerial vehicle is responsible for subtasks T2, T7, T13, T18 and T24, the total energy consumption is expected to be 71% of the battery capacity, and the remaining 29%; the second unmanned aerial vehicle is responsible for subtasks T1, T6, T12, T19 and T23, the total energy consumption is expected to be 68% of the battery capacity, and the remaining 32%; the third unmanned aerial vehicle is responsible for subtasks T3, T8, T14, T20, T25 and T29, the total energy consumption is expected to be 73% of the battery capacity, and the remaining 27%; the fourth unmanned aerial vehicle is responsible for subtasks T4, T9, T15, T21 and T26, the total energy consumption is expected to be 69% of the battery capacity, and the remaining 31%; and the fifth unmanned aerial vehicle is responsible for subtasks T5, T10, T16, T22, T27 and T30, the total energy consumption is expected to be 75% of the battery capacity, and the remaining 25%, the scheme meets all the constraint conditions and realizes the maximization of the overall task execution benefit. After optimization, the unmanned aerial vehicle group can maintain a good communication connection state, the task load is evenly distributed, the energy utilization efficiency is high, and the collaborative monitoring efficiency is greatly improved.

[0102] During the task execution process, the unmanned aerial vehicle state and the communication quality change are continuously monitored, and the dynamic adjustment mechanism is triggered when an abnormal situation is detected. The abnormal situation includes abnormal energy consumption of the unmanned aerial vehicle, sudden change of communication quality, addition of high-priority tasks, etc. The dynamic adjustment adopts a local optimization strategy, and only the task allocation of the affected area is recalculated, thereby reducing the adjustment overhead. Exemplarily, when the fourth unmanned aerial vehicle encounters strong air flow in a task, the energy consumption increases by 20%, the energy constraint data is updated in real time, and the two lowest-priority subtasks T15 and T21 responsible for the unmanned aerial vehicle are redistributed to the third and fifth unmanned aerial vehicles, respectively, thereby ensuring the task continuity and the safety of the unmanned aerial vehicle.

[0103] In this embodiment, by introducing the communication weight coefficient and the task execution cost coefficient, and combining the communication influence coefficient to dynamically adjust the benefit matrix, the flexible trade-off between communication efficiency and execution cost in the task allocation process is realized. The state information update data is calculated based on the communication quality evaluation data and the weight increment data between the unmanned aerial vehicles, a dynamic feedback mechanism in the unmanned aerial vehicle group allocation process is established, the task allocation scheme can adapt to the change of the communication environment, and the optimal task allocation scheme is obtained by iteration optimization under the premise of ensuring to meet the uniqueness constraint, the task bearing constraint and the energy constraint, thereby improving the convergence efficiency of the algorithm and the optimization degree of the scheme.

[0104] Figure 2 The unmanned aerial vehicle task allocation optimization flowchart for the unmanned aerial vehicle group formation and cooperative and task allocation optimization method of the embodiment of the present application.

[0105] In an alternative embodiment,

[0106] According to the optimized task allocation scheme, the expected flight route of each UAV for executing the allocated task is calculated, and initial formation position information is generated, which comprises:

[0107] The task execution sequence of each UAV in the optimized task allocation scheme is obtained, and an ant colony algorithm is used to calculate the shortest flight path of each UAV based on the task execution sequence, wherein the shortest flight path meets the task timing constraint and the UAV turning radius constraint;

[0108] The flight time of each UAV to reach the first task execution point is calculated according to the shortest flight path, the UAV with the longest flight time is determined as the key UAV, and the delay takeoff time of other UAVs is calculated based on the flight time of the key UAV;

[0109] The initial position coordinates of each UAV are determined based on the delay takeoff time, and the initial formation position information meeting the minimum distance constraint of the UAV is constructed.

[0110] The task execution sequence of each UAV is determined. The task execution sequence refers to the order arrangement of the UAV accessing multiple sub-task areas responsible by the UAV, which directly affects the flight path length and task completion efficiency. For n sub-tasks allocated to each UAV, there are n! possible access sequences, and the optimal execution sequence is selected by the ant colony algorithm. In the implementation process of the ant colony algorithm, multiple key parameters are set, including the number of ants, the initial value of pheromone, the pheromone evaporation coefficient, the heuristic factor and the iteration number. The number of ants is usually set to 2-3 times the number of task points, the initial value of pheromone is set to a uniform small positive number, the pheromone evaporation coefficient is controlled between 0.2 and 0.5, the heuristic factor reflects the influence of path distance on selection probability, and the iteration number is determined according to the problem size. For example, in a certain border patrol task, UAV No. 1 is responsible for 5 sub-task points with coordinates (5.2, 3.7), (8.6, 5.1), (6.3, 7.8), (4.5, 6.2) and (9.1, 4.3), the number of ants is set to 15 by the ant colony algorithm, the initial value of pheromone is 0.5, the pheromone evaporation coefficient is 0.3, the heuristic factor is 2.0, and the iteration number is 100. During the algorithm execution process, in the path construction stage of each iteration, each ant calculates the transition probability based on the pheromone concentration and distance factors on the path, and selects the next access point. After completing a round of path construction, the total length of the path constructed by each ant is calculated, and the pheromone is updated according to the path length, and the shorter the path, the greater the pheromone increment. After 100 iterations, the pheromone distribution stabilizes, and the optimal access sequence is obtained as (5.2, 3.7)→(9.1, 4.3)→(8.6, 5.1)→(6.3, 7.8)→(4.5, 6.2), and the total length of the route is 14.3 kilometers.

[0111] In the implementation process of the ant colony algorithm, the task timing constraint and the turning radius constraint of the UAV are considered. The task timing constraint refers to the requirement that specific tasks must be executed in a certain order, such as reconnaissance tasks must be executed before attack tasks. The method for implementing the timing constraint is to add a legality check during path construction. When a task point that violates the timing constraint is selected, the probability of the selection is set to zero. The turning radius constraint of the UAV is determined by the physical characteristics of the UAV. Different types of UAVs have different minimum turning radii. Fixed-wing UAVs have a larger turning radius due to the need to maintain a certain forward speed, typically ranging from 50 to 200 meters. Multi-rotor UAVs can achieve small-radius turning, with a minimum turning radius of 5 to 20 meters. The method for implementing the turning radius constraint is to check the angle formed by the adjacent three points during path generation. The corresponding turning radius is calculated. If it is less than the minimum turning radius of the UAV, the constraint is satisfied by adjusting the position of the middle point or inserting a transition point. In the aforementioned patrol task, UAV No. 1 is a fixed-wing UAV with a minimum turning radius of 100 meters. After checking, the turning radius of the original path from (9.1, 4.3) to (8.6, 5.1) to (6.3, 7.8) is 83 meters, which does not meet the constraint, and a transition point (7.8, 5.9) needs to be inserted. The revised path is (5.2, 3.7) to (9.1, 4.3) to (8.6, 5.1) to (7.8, 5.9) to (6.3, 7.8) to (4.5, 6.2), which satisfies the turning radius constraint.

[0112] According to the calculated shortest route path, the flight time of each UAV from the takeoff point to the first task execution point is calculated. The flight time calculation needs to consider the takeoff speed curve, cruising speed, and environmental factors of the UAV. The takeoff speed curve describes the process of the UAV accelerating from a stationary state to a cruising speed, which is usually divided into three stages: initial acceleration section, linear speed-up section, and approach to cruising speed section. The cruising speed is the standard speed of the UAV in a steady flight state, and the environmental factors mainly include wind speed, wind direction, air pressure, and temperature. The flight time calculation method is to divide the flight path into several small sections, calculate the flight time of each section according to the takeoff stage, wind direction factor, and UAV performance parameters, and accumulate the total flight time. In the patrol task, the takeoff point of the 6 UAVs is located at the coordinates (2.0, 1.0), and the first task point of UAV No. 1 is (5.2, 3.7). The straight-line distance is 4.18 kilometers, the cruising speed is 20 meters / second, and considering the influence of the northwest wind of 5 meters / second, the actual effective flight speed is 17.6 meters / second. The takeoff acceleration stage takes 20 seconds to reach the cruising speed, and the calculated flight time is 257 seconds. The flight times of other UAVs are calculated as follows: 202 seconds for UAV No. 2, 275 seconds for UAV No. 3, 238 seconds for UAV No. 4, 226 seconds for UAV No. 5, and 198 seconds for UAV No. 6. Comparing the flight times of each UAV, it is found that UAV No. 3 has the longest flight time of 275 seconds, so it is determined as the key UAV.

[0113] The time delay take-off time of other UAVs is calculated based on the flight time of the key UAV, the arrival time of each UAV at the first task point is coordinated, and the synchronization of the beginning of the task is improved. The calculation method of the time delay take-off time is to subtract the flight time of the current UAV from the flight time of the key UAV, and the result is the time value of the time delay take-off. If the calculation result is negative, it means that the UAV cannot arrive at the same time as the key UAV even if it takes off immediately, at this time, the delay time is set to zero, and the early arrival time of the UAV is recorded for subsequent task planning adjustment. For example, in the patrol task, the flight time of the key UAV 3 is 275 seconds, so the time delay take-off time of each UAV is respectively: UAV 1 delays 18 seconds (275-257), UAV 2 delays 73 seconds (275-202), UAV 3 delays 0 seconds, UAV 4 delays 37 seconds (275-238), UAV 5 delays 49 seconds (275-226), and UAV 6 delays 77 seconds (275-198). Through the implementation of the time delay take-off strategy, each UAV is expected to arrive at the first task execution point at the same time around 275 seconds, realizing the synchronized start of task execution.

[0114] The initial position coordinates of each UAV are determined based on the time delay take-off time, and the initial formation position information that meets the minimum distance constraint of the UAV is constructed. The initial position planning needs to consider the minimum distance required for the safe take-off of the UAV. The minimum safe distance requirements of different types of UAVs are different. The minimum horizontal distance between multi-rotor UAVs is usually 5 to 8 meters, and the minimum horizontal distance between fixed-wing UAVs is 15 to 20 meters. The initial formation shape can be selected from multiple forms such as straight line, matrix, triangle or circle, and can be flexibly selected according to the task characteristics and take-off site conditions. In the implementation process, the formation shape and reference distance are determined, and then the initial position coordinate matrix is generated according to the number and type of UAVs. For example, in the patrol task, a double-column matrix formation is adopted, the horizontal distance between adjacent UAVs is set to 15 meters, and the initial position coordinates are: UAV 1 (2.0, 1.0), UAV 2 (2.0, 16.0), UAV 3 (2.0, 31.0), UAV 4 (17.0, 1.0), UAV 5 (17.0, 16.0), and UAV 6 (17.0, 31.0). Not only does it meet the safety distance requirement, but it also facilitates the take-off of UAVs in the order of time delay take-off. In the vertical direction, the reference flight height is set to 120 meters, and the height difference between adjacent UAVs is 10 meters, forming a three-dimensional formation space, further improving the flight safety. After the UAV takes off, it takes off from the initial position according to the predetermined time delay take-off time, and flies to the first task execution point according to the calculated shortest route. All UAVs successfully arrive at the task area according to the predetermined plan, realizing the efficient and coordinated beginning of the task.

[0115] In the embodiment, the ant colony algorithm is used to calculate the shortest route path meeting the task timing constraint and the turning radius constraint, ensuring the feasibility of the route planning, realizing the optimization of the route length, improving the task execution efficiency, realizing the time synchronization of the UAV group to the task point by identifying the key UAV and calculating the delayed take-off time of other UAVs based on the flight time, avoiding the task execution conflict caused by the inconsistent arrival time among the UAVs, determining the initial formation position information based on the delayed take-off time and meeting the minimum distance constraint, ensuring the safety of the UAV group in the take-off stage and providing a reasonable initial spatial layout for the subsequent formation flight.

[0116] In an optional implementation manner,

[0117] Based on the initial formation position information, the environment parameters in the flight process of the UAV group are collected, and the optimal distance between the UAVs is calculated based on the environment parameters.

[0118] Based on the initial formation position information, the UAV group topology space is constructed, the spatial connection structure is constructed according to the relative distance relationship between the UAVs, and the topology invariants are calculated to obtain the environment parameters.

[0119] The stability of the environment parameters is analyzed by using the topological persistence atlas, and the persistence measure value is calculated, the topological complexity evaluation index is constructed, and the local topological complexity corresponding to each UAV is determined, the aerodynamic energy consumption caused by the airflow interference between the UAVs is calculated by using the flight dynamics method, the communication cost is calculated based on the communication link quality and the information transmission demand of the UAVs, and the optimal distance between the UAVs is obtained by taking the topological complexity evaluation index, the aerodynamic energy consumption and the communication cost as the optimization target.

[0120] The three-dimensional coordinate information of each unmanned aerial vehicle is obtained, including longitude, latitude and height data, to form a complete set of spatial position points. A spatial connection structure is constructed according to the relative distance relationship between the unmanned aerial vehicles, and a distance threshold is set. When the distance between two unmanned aerial vehicles is less than the threshold, a connection is established, and when the distance is greater than the threshold, no connection is established. In an exemplary six-unmanned aerial vehicle formation flight task, the initial formation position is a matrix arrangement, and the coordinates of the unmanned aerial vehicles are as follows: unmanned aerial vehicle 1 (0, 0, 100), unmanned aerial vehicle 2 (15, 0, 105), unmanned aerial vehicle 3 (30, 0, 110), unmanned aerial vehicle 4 (0, 15, 100), unmanned aerial vehicle 5 (15, 15, 105), and unmanned aerial vehicle 6 (30, 15, 110), with the distance unit being meters. The distance threshold is set to 25 meters, the distance between each unmanned aerial vehicle is calculated, and the connection relationship is established. The distance between unmanned aerial vehicle 1 and unmanned aerial vehicle 2 is 15.8 meters, and a connection is established. The distance between unmanned aerial vehicle 1 and unmanned aerial vehicle 3 is 31.6 meters, and no connection is established. The distance between unmanned aerial vehicle 1 and unmanned aerial vehicle 4 is 15 meters, and a connection is established. The connection relationship between all unmanned aerial vehicles is determined, and a complete topological connection diagram is formed. Based on the established connection diagram, topological invariants are calculated, including the Betti number, the Euler characteristic number and the number of connected components. In this formation case, the Betti number is 2, the Euler characteristic number is 3, and the number of connected components is 1. The topological invariants constitute the environmental parameters, which are used to evaluate the stability and robustness of the formation structure.

[0121] The stability of the environmental parameters is analyzed using the topological persistence map, and the persistence measure value is calculated. The distance threshold is gradually increased from small to large, and the changes in topological characteristics at each threshold are recorded to form a two-dimensional persistence map. In the topological persistence map, the horizontal axis represents the threshold at which the feature is generated, and the vertical axis represents the threshold at which the feature disappears. The point pair (b, d) represents the feature generated at threshold b disappearing at threshold d. The persistence measure value is defined as the difference between the disappearance threshold and the generation threshold. The larger the value, the more stable the feature. In the six-vehicle formation case, the distance threshold range is set to 5 to 50 meters, with a step size of 5 meters. When the threshold is 15 meters, two hollow features are generated. When the threshold increases to 30 meters, one of the hollows disappears. When the threshold increases to 45 meters, the other hollow also disappears. Therefore, the persistence measure values of the two hollows are 15 (30-15) and 30 (45-15) respectively, indicating that the second hollow feature is more stable. Based on the analysis results of the persistence map, a topological complexity evaluation index is constructed. This index considers the number of topological features and the persistence measure value. The calculation formula is the weighted sum of the persistence measure values of each feature, and the weight is related to the importance of the feature. In this embodiment, the weights of the two hollow features are set to 0.4 and 0.6 respectively, and the topological complexity evaluation index is calculated to be 24 (15x0.4+30x0.6).

[0122] For each UAV, its local topology complexity is calculated, which reflects the complexity of the location where the single UAV is located. The calculation method is to set an observation radius centered on the UAV, and analyze the topology structure characteristics within the range. In the six-UAV formation, taking UAV No. 1 as the center, the observation radius is set to 20 meters, which contains UAV No. 2 and No. 4, forming a triangular connection structure, and the local topology complexity is calculated to be 5.2; taking UAV No. 3 as the center, the observation radius is 20 meters, which only contains UAV No. 2, forming a line segment connection structure, and the local topology complexity is calculated to be 2.1. The aerodynamic energy consumption caused by airflow interference between UAVs is calculated by flight dynamics method. The downwash airflow generated by the rotor UAV during flight will affect the flight stability and energy consumption of the nearby UAV. The degree of aerodynamic interference is inversely proportional to the distance between UAVs, and proportional to the rotor diameter and rotation speed. The aerodynamic energy consumption calculation is based on the computational fluid dynamics model, considering the velocity field distribution of the downwash airflow and the force condition of the UAV. During formation flight, when the distance between two quadrotors is 10 meters, the additional power consumption of the downstream UAV increases by about 15%; when the distance increases to 20 meters, the additional power consumption decreases to about 5%; when the distance reaches 30 meters, the aerodynamic interference can be ignored.

[0123] The communication cost is calculated based on the UAV communication link quality and information transmission demand. The communication link quality is related to transmission distance, antenna characteristics and environmental interference, and is usually evaluated by signal-to-noise ratio and bit error rate indicators. Information transmission demand depends on task type and degree of cooperation, and highly cooperative tasks require more frequent information exchange. The communication cost calculation considers communication energy consumption, bandwidth occupation and delay, which approximately increases in quadratic relationship with transmission distance. For example, in formation tasks, when the distance between two UAVs is 15 meters, the communication link quality score is 0.95 (full score 1), and the communication energy consumption is 0.5 watts; when the distance increases to 30 meters, the link quality decreases to 0.82, and the communication energy consumption increases to 1.2 watts; when the distance reaches 45 meters, the link quality further decreases to 0.68, and the communication energy consumption increases to 2.3 watts. The topology complexity evaluation index, aerodynamic energy consumption and communication cost are taken as optimization objectives to build a multi-objective optimization problem to find the optimal distance between UAVs. The multi-objective optimization uses the weighted sum method to convert the three objective functions into a single objective function, and the weights are set according to the task characteristics and environmental conditions. In the environmental exploration task, the weights of topology complexity, aerodynamic energy consumption and communication cost are set to 0.3, 0.4 and 0.3 respectively. The optimization algorithm uses the particle swarm optimization method, sets the number of particles to 50, the number of iterations to 100, and the initial velocity range to [-2, 2], and finally converges to the optimal solution by continuously updating the particle position and velocity.

[0124] In the six-machine formation optimization case, the initial spacing is set to be horizontal 15 meters and vertical 5 meters. After multi-objective optimization calculation, the optimal spacing is obtained as horizontal 22.5 meters and vertical 7.5 meters. Under this spacing configuration, the topology complexity evaluation index is 18.3, the average aerodynamic energy consumption is reduced by 8.5%, the communication cost is increased by 3.2%, and the comprehensive score is optimal.

[0125] In this embodiment, by constructing the UAV group topology space and calculating the topological invariant, the mathematical description of the spatial structure characteristics of the UAV group is realized, which provides a theoretical basis for subsequent spacing optimization. The environmental parameter stability is analyzed by using the topological persistence graph, and the local topological complexity is introduced, which can accurately evaluate the complexity of the spatial structure of the UAV group and ensure the stability of the formation configuration. The aerodynamic energy consumption caused by airflow interference and the communication cost caused by communication link quality are considered, and the topological complexity, energy consumption and communication are taken as multi-objective optimization indexes. The optimal spacing obtained not only ensures the energy efficiency of formation flight, but also ensures the communication quality between UAVs.

[0126] In an alternative embodiment,

[0127] According to the optimal spacing, the relative positions of the UAVs are dynamically adjusted, the aerodynamic interference is predicted based on the relative motion between the UAVs, and real-time compensation is performed to obtain an optimized formation scheme, which includes:

[0128] Based on the optimal spacing, the distance error between adjacent UAVs is calculated, the motion speed and direction of the UAVs are determined according to the distance error, the motion speed and direction are corrected based on the change trend of the local topological complexity, the position adjustment instructions of the UAVs are obtained and executed;

[0129] During the execution of the position adjustment instructions, the evolution process of the formation configuration is analyzed by the topological persistence of the environmental parameters, the stable topological structure characteristics are extracted, the topological invariant characteristic value and the persistence measurement characteristic value are calculated according to the stable topological structure characteristics, the motion trend between adjacent UAVs is analyzed based on the characteristic values, and the aerodynamic interference between the UAVs is predicted;

[0130] According to the aerodynamic interference, the aerodynamic compensation control amount is calculated, the tracking error between the actual position and the target position of the UAV is calculated, the aerodynamic compensation control amount and the tracking compensation control amount are taken as the control input of the UAV, and the optimized formation scheme is obtained.

[0131] Real-time position data of each UAV is acquired, and high-precision three-dimensional coordinates are obtained through global positioning system and airborne inertial measurement unit fusion positioning. For adjacent UAV pairs, the difference between the actual distance and the optimal distance is the distance error. For example, in a six-UAV formation flight task, the optimal distance is determined to be 22.5 meters horizontally and 7.5 meters vertically. At a certain moment, the position of UAV No. 1 is (125.36, 83.42, 102.15), and the position of UAV No. 2 is (146.21, 85.76, 108.93). The actual distance between the two UAVs is calculated to be 21.38 meters, which is 1.12 meters less than the optimal horizontal distance of 22.5 meters, indicating that the distance between the two UAVs needs to be increased. The movement speed and direction of the UAV are determined according to the distance error. The proportional control method based on distance error is adopted, and the movement speed is proportional to the distance error, and the direction is along the line connecting the two UAVs. When the distance error is negative, the two UAVs move away from each other; when the distance error is positive, the two UAVs move closer to each other. In the aforementioned case, the distance error between UAV No. 1 and UAV No. 2 is -1.12 meters, and the control coefficient is set to 0.3. The movement speed of UAV No. 1 in the direction away from UAV No. 2 is calculated to be 0.336 meters per second.

[0132] The movement speed and direction are corrected based on the change trend of local topological complexity, which reflects the structural characteristics of the local environment of the UAV. The correction process first calculates the local topological complexity of the area around the current UAV, predicts the complexity change in different moving directions, and selects the moving direction that keeps the complexity within a reasonable range. In the application scenario, if it is detected that moving UAV No. 1 in the predetermined direction will cause the local topological complexity to rise sharply, indicating that an unstable structure may appear, then the direction is deviated by 15 degrees based on the original direction, and the moving speed is reduced to 0.28 meters per second, ensuring the stable change of the formation structure. The position adjustment instruction of the UAV is obtained after correction, which includes the target direction vector and the target speed scalar, and is sent to the corresponding UAV for execution through the communication link. In the process of executing the position adjustment instruction, the UAV uses a smooth trajectory planning algorithm to avoid flight instability caused by sudden acceleration.

[0133] During the execution of the position adjustment instruction, the formation configuration evolution process is analyzed through the topological persistence of the environmental parameters, a series of topological spaces are constructed in the continuous time window, the generation and extinction moments of topological features are recorded, and the persistence atlas is generated. For example, during the adjustment process of a six-vehicle formation, the formation position data is collected every 200 milliseconds, and the persistence atlas in a 10-second window is constructed. The triangular connection structure formed in the center area of the formation has a higher persistence measure value of 45.6, while the persistence measure value of the connection structure in the edge area is only 12.3, indicating that the structure in the center area is more stable. The stable topological structure features are extracted, and the features with a persistence measure value greater than a certain threshold are focused on. The features remain stable during the formation adjustment process and can be used as the priority maintenance objects for formation control. The topological invariant feature values and persistence measure feature values are calculated based on the stable topological structure features. In the six-vehicle formation, the Betti number of the stable triangular structure is 1, the Euler characteristic number is 0, and the persistence measure feature value is 45.6. These feature values constitute an index system for evaluating the stability of the formation.

[0134] Based on the feature value analysis of the motion trend between adjacent unmanned vehicles, the time series analysis method is used to process the continuously collected position data to extract the motion pattern and trend features. In the analysis process, it is found that the distance between No. 3 and No. 4 unmanned vehicles continues to decrease at a rate of 0.28 meters per second, and it is predicted that the distance will be less than the safety threshold of 18 meters after 30 seconds. The prediction of aerodynamic interference between unmanned vehicles needs to consider four factors: relative position, relative speed, rotor parameters, and environmental wind field. Based on the simplified computational fluid dynamics principle, the downwash airflow velocity field is approximated as a Gaussian distribution, the airflow intersection area is calculated through the relative position of the unmanned vehicles, and then the aerodynamic interference force and moment are estimated. In the relative position relationship between No. 3 and No. 4 unmanned vehicles, No. 4 unmanned vehicle is at the edge of the downwash airflow area of No. 3 unmanned vehicle, and it is predicted that No. 4 unmanned vehicle will be subjected to downward and outward aerodynamic interference, with an additional sinking force in the vertical direction of 6.2% of the standard gravity and an outward thrust in the horizontal direction of 3.8% of the standard gravity.

[0135] The aerodynamic compensation control quantity is calculated according to aerodynamic interference, and a feedforward compensation strategy is adopted to directly offset the predicted aerodynamic interference force and torque. The calculation of the aerodynamic compensation control quantity is based on the unmanned aerial vehicle dynamics model, considering the rotor aerodynamic characteristics and the force balance of the vehicle body. For example, for the No. 4 unmanned aerial vehicle, in order to offset the sinking force, the vertical direction thrust needs to be increased by 6.2%, which corresponds to an increase of about 3.1% in the motor speed; in order to offset the horizontal outward thrust, a horizontal thrust of 3.8% in the opposite direction needs to be applied, which is realized by adjusting the differential speed of each rotor. The tracking compensation control quantity is calculated according to the tracking error between the actual position and the target position of the unmanned aerial vehicle, and the tracking error is the difference between the actual position vector and the target position vector. The tracking compensation control adopts a proportional-integral-derivative (PID) controller, which adjusts the control output in real time through the error signal. In the position control of the No. 4 unmanned aerial vehicle, the tracking error at a certain time is (0.82, -0.53, 0.45) meters, and the PID parameters are set to P=0.6, I=0.1, and D=0.3. The tracking compensation control quantity calculated is (0.65, -0.42, 0.36) in standard gravity percentage.

[0136] The aerodynamic compensation control quantity and the tracking compensation control quantity are combined as the control input of the unmanned aerial vehicle using linear superposition. In the control of the No. 4 unmanned aerial vehicle, the horizontal direction control input is (-0.42+0, 0.65-3.8) percentage, and the vertical direction control input is (0.36+6.2) percentage. After converting into specific motor control signals, they are sent to the flight control system for execution. After multiple rounds of position adjustment and control optimization, the unmanned aerial vehicle group gradually converges to the predetermined formation configuration, forming a stable flight formation. For example, in a certain forest patrol task, the final formation of the six-vehicle formation after adjustment has an error between adjacent unmanned aerial vehicles of less than 0.5 meters, and the position keeping accuracy is ±0.3 meters. Even under 5-level wind conditions, the stable formation can be maintained. The complete optimization formation scheme not only includes the target position coordinates of each unmanned aerial vehicle, but also includes dynamic adjustment strategies, communication topology structure, and abnormal handling mechanisms, forming a complete formation control solution.

[0137] In this embodiment, the motion speed and direction of the unmanned aerial vehicle are dynamically adjusted based on the distance error and local topology complexity, realizing real-time optimization of the formation configuration and ensuring the convergence of the spatial structure of the unmanned aerial vehicle group. Through topology persistence analysis and stable topology structure feature extraction, the evolution process of the formation configuration can be effectively monitored, and the aerodynamic interference can be predicted based on the topology invariant feature value and the persistence measurement feature value, improving the predictability of the formation control. The introduction of the aerodynamic compensation control quantity and the tracking compensation control quantity as the comprehensive control input not only compensates for the influence of aerodynamic interference, but also ensures the position tracking accuracy, realizing accurate control during formation flight and improving the robustness and stability of the formation scheme.

[0138] Figure 3The unmanned aerial vehicle formation control efficiency comparison chart of the unmanned aerial vehicle group formation cooperation and task allocation optimization method of the embodiment of the present application shows the control efficiency comparison of the technical solution (topological persistence analysis method), the traditional position control method and the basic formation algorithm in four different test scenarios.

[0139] In a standard flight environment, the efficiency of the present solution is 85.3%, which is much higher than the 68.2% of the traditional method and the 52.1% of the basic formation algorithm. The technical solution extracts stable topological structure features through topological persistence analysis of environmental parameters, so that the structural stability can be maintained during the formation adjustment process.

[0140] Under the condition of 5-level wind, the technical solution can accurately calculate the aerodynamic compensation control amount to offset the predicted disturbance force and moment, instead of relying only on the after-feedback adjustment of the traditional method.

[0141] In the dense formation scenario (88.9%), the technical solution adjusts the motion parameters based on the change trend of local topological complexity to avoid structural instability. When the topological complexity is detected to rise sharply, the direction and speed are intelligently adjusted to ensure smooth transition of the formation.

[0142] In the long-time inspection task (94.7%), the technical solution effectively deals with the cumulative error in long-time flight by continuously monitoring the formation configuration evolution and combining with the real-time adjustment of the PID controller, and can maintain high-precision position control of ±0.3 meters even in complex environments.

[0143] In summary, the technical solution combines topological analysis with traditional control methods to form a multi-level and multi-dimensional cooperative control strategy, which significantly improves the flight stability and control efficiency of the unmanned aerial vehicle formation in various environments.

[0144] In an alternative embodiment,

[0145] The optimized formation scheme is sent to the unmanned aerial vehicle group for execution, and obstacle information in the flight process is monitored in real time. If an obstacle is detected, an obstacle avoidance strategy is calculated and the optimized formation scheme is updated until the task is completed, including:

[0146] The optimized formation scheme is sent to the unmanned aerial vehicle group, and the unmanned aerial vehicle group flies in formation according to the optimized formation scheme;

[0147] Obstacle information is monitored in real time during the formation flight, and obstacle position coordinates are obtained according to the obstacle information. If an obstacle is detected, a safe flight path for each unmanned aerial vehicle is calculated based on the obstacle position coordinates, and the minimum distance between the safe flight path and the obstacle is greater than a preset safety threshold.

[0148] According to the safe flight path, an updated optimal formation scheme is generated, the updated optimal formation scheme including a flight trajectory of the UAV flying around the obstacle and a reconstruction formation strategy after obstacle avoidance, and the updated optimal formation scheme is sent to the UAV group for execution until the task is completed.

[0149] The optimal formation scheme is distributed to each UAV in the form of a data packet by a ground control station or a lead UAV, the data packet containing formation geometry information, relative position coordinates, flight speed parameters, and communication topology. The data transmission adopts an encrypted wireless communication protocol to ensure information security and reliability. For example, in a certain mountain power line inspection task, the data packet size of the optimal formation scheme of a six-UAV formation is 26.4 KB, containing complete formation configuration description and 120 seconds of trajectory prediction data. The data packet is distributed by the lead UAV as a relay node, with an average transmission delay of 78 milliseconds and a transmission success rate of 99.7%. After receiving the formation scheme, the UAV performs data analysis and verification, extracts the corresponding position instructions and speed parameters, integrates with the local navigation system, and generates specific flight control instructions. The UAV flies in formation according to the optimal formation scheme, adjusts the position according to the received three-dimensional coordinate information, and forms the predetermined formation configuration. The flight control system adopts a cascade PID control architecture, with an inner loop for attitude stabilization and an outer loop for position accuracy, and the control frequencies are 200 Hz and 50 Hz, respectively. In the inspection task, the six-UAV formation successfully constructed a "reverse V" formation configuration, with a horizontal distance between each UAV of 22.5 meters and a vertical distance of 7.5 meters, and the position keeping accuracy was better than ±0.4 meters.

[0150] Real-time monitoring of obstacle information is an important measure to ensure flight safety during formation flight. The obstacle monitoring system includes an active sensing unit and a passive sensing unit. The active sensing unit uses a laser radar, a millimeter wave radar, or an ultrasonic sensor to actively emit a detection signal and detect obstacles by receiving the reflected signal. The passive sensing unit uses an optical camera and an infrared camera to passively receive environmental images and identify obstacles through computer vision algorithms. For example, in the patrol task, UAVs 1 and 3 are equipped with a 23-line laser radar with a horizontal field of view of 360 degrees, a vertical field of view of 30 degrees, and a detection distance of 100 meters. UAVs 2 and 5 are equipped with a binocular stereo camera with a field of view of 90 degrees and an effective recognition distance of 60 meters. UAVs 4 and 6 are equipped with a millimeter wave radar with a field of view of 60 degrees and a detection distance of 150 meters, which can penetrate fog and light rain. When obtaining the obstacle position coordinates from the obstacle information, a multi-sensor data fusion technology is used to improve the positioning accuracy. The data fusion process includes coordinate system conversion, time synchronization, and Kalman filtering. In the case where the same obstacle is detected by multiple sensors, weights are assigned according to the measurement error characteristics of each sensor, and a more accurate obstacle position is obtained by weighted averaging. During the patrol, the monitoring system detects a microwave communication tower 80 meters ahead, with a tower height of 62 meters, a base coordinate of (785.3, 342.6, 125.8), and a top coordinate of (785.3, 342.6, 187.8), with a positioning accuracy of ±1.2 meters.

[0151] If an obstacle is detected, the safe flight path for each UAV is calculated based on the obstacle position coordinates. The safe flight path planning uses the artificial potential field method, representing obstacles as repulsive fields and target waypoints as attractive fields, guiding UAVs to avoid obstacles through the combined force field. To improve computational efficiency, obstacles are simplified as bounding boxes or bounding spheres, and the communication tower is simplified as a cylindrical body with a height of 62 meters and a radius of 8 meters. The preset safety threshold is 25 meters, i.e., the minimum distance between the planned safe flight path and the surface of the obstacle is not less than 25 meters. Considering the performance differences of different UAVs, the safety threshold can be dynamically adjusted according to the flight speed, with higher flight speed resulting in a larger safety threshold. In the flight path planning, the potential field parameters are optimized to avoid UAVs getting stuck in local minimum points, while considering UAV dynamics constraints to ensure that the generated trajectory meets the minimum turning radius requirement. A smooth trajectory is generated through cubic spline interpolation to reduce acceleration discontinuity. The calculation results show that UAV 1 chooses to pass from the left side of the communication tower with a closest approach distance of 33.7 meters; UAVs 2 to 5 choose to pass from the right side with closest approach distances of 36.2 meters, 32.5 meters, 34.8 meters, and 35.1 meters, respectively; and UAV 6 chooses to pass from above the tower top with a closest approach distance of 28.3 meters. The safe flight paths of all UAVs meet the preset safety threshold requirements.

[0152] According to the safe flight path, an updated optimal formation scheme is generated, which includes the flight trajectory of the UAVs around the obstacle and the reconstruction strategy of the formation after obstacle avoidance. The flight trajectory planning needs to consider the overall structure of the formation and the relative position relationship between the UAVs to avoid new collision risks between the UAVs during obstacle avoidance. The formation flight strategy has two ways: overall flight and grouped flight. When the obstacle size is small or the formation scale is small, the overall flight strategy is adopted to maintain the formation configuration unchanged and adjust the flight path as a whole. When the obstacle size is large or the shape is complex, the grouped flight strategy is adopted to temporarily divide the formation into multiple sub-formations, and different paths are selected for flight. In the foregoing example, the six-formation is divided into three groups: No. 1 alone from the left side, Nos. 2-5 four machines in a group from the right side, and No. 6 alone from above. The flight trajectory is described by a three-order Bezier curve to ensure smooth and continuous trajectory. The reconstruction strategy of the formation after obstacle avoidance defines how the UAVs recover to the original formation configuration. The reconstruction strategy includes three parts: convergence point selection, timing coordination and trajectory planning. The convergence point selection determines a suitable position for re-converging on the flight path outside the safe distance of the obstacle; the timing coordination ensures that each UAV or sub-formation reaches the convergence point at the appropriate time; and the trajectory planning generates a smooth trajectory from the separation point to the convergence point. In the foregoing example, a position 150 meters downwind of the obstacle is selected as the convergence point, and the expected arrival time is calculated according to the flight distance and speed of each group. The No. 1 UAV is expected to take 18.6 seconds, the Nos. 2-5 group is expected to take 19.2 seconds, and the No. 6 UAV is expected to take 17.3 seconds. To coordinate the arrival time, the flight speed of each group is adjusted. The speed of No. 1 UAV is adjusted to 15.3 m / s, the speed of Nos. 2-5 group is adjusted to 16.8 m / s, and the speed of No. 6 UAV is adjusted to 14.5 m / s, so that the three groups of UAVs can arrive at the convergence point within a time difference of 2 seconds, facilitating the reconstruction of the formation.

[0153] The updated optimal formation scheme is sent to the UAV group for execution. The update scheme is transmitted to each UAV in real time through a wireless communication network, and a high-priority data packet is used to ensure fast processing. Data integrity check is implemented during transmission to prevent information loss or errors. After receiving the update scheme, the UAV immediately switches to the obstacle avoidance mode and executes the flight according to the new trajectory. During execution, each UAV continuously monitors the position change of the obstacle and the position error of itself, and adjusts the local trajectory if necessary. At the same time, the communication within the group is maintained, and the position and speed information is exchanged to ensure coordination. When all UAVs safely pass through the obstacle area and reconstruct the formation at the convergence point, the system returns to the normal task execution mode and continues to fly along the planned path until the task is completed.

[0154] In this embodiment, by monitoring the obstacle information in real time and calculating the safe flight path, the safety of obstacle avoidance during the formation flight of the UAV group is ensured, the setting of the safety threshold provides sufficient obstacle avoidance margin for the UAV, the updated formation scheme containing the bypass trajectory and reconstruction strategy is dynamically generated, the autonomous obstacle avoidance and formation reconstruction of the UAV group when encountering obstacles are realized, the continuity of task execution is ensured, the closed-loop control mechanism of continuous monitoring and scheme updating is adopted, the UAV group can adjust the flight strategy in real time according to the environmental changes, and the adaptability of the formation system to complex environment and the reliability of task completion are improved.

[0155] In a second aspect of the embodiment of the application, a UAV group formation coordination and task allocation optimization system is provided, comprising:

[0156] A first unit is configured to acquire flight state information and task area information of UAVs in a UAV group, determine an initial spatial distribution of the UAV group based on the flight state information, and determine an initial formation position of the UAV group based on the initial spatial distribution.

[0157] A second unit is configured to divide the task area information into a plurality of subtasks, set initial task allocation parameters, determine communication weights of the UAVs according to communication quality and relative distances between the UAVs, construct an information interaction mechanism between the UAVs based on the communication weights, calculate execution costs of the UAVs for executing the subtasks based on the initial task allocation parameters, and obtain an optimized task allocation scheme by minimizing total task execution costs of all the UAVs.

[0158] A third unit is configured to calculate expected flight paths of each UAV for executing the allocated tasks according to the optimized task allocation scheme, and generate initial formation position information.

[0159] A fourth unit is configured to acquire environmental parameters in a flight process of the UAV group based on the initial formation position information, calculate optimal distances between the UAVs based on the environmental parameters, dynamically adjust relative positions of the UAVs according to the optimal distances, predict aerodynamic interference effects based on relative motions between the UAVs and perform real-time compensation, and obtain an optimized formation scheme.

[0160] A fifth unit is configured to send the optimized formation scheme to the UAV group for execution, monitor obstacle information in a flight process in real time, calculate an obstacle avoidance strategy and update the optimized formation scheme if an obstacle is monitored, and perform the task until completion.

[0161] In a third aspect of the embodiment of the application, a computer readable storage medium is provided, which stores computer program instructions, and the computer program instructions are executed by a processor to implement the method described above.

[0162] The present application can be a method, an apparatus, a system, and / or a computer program product. The computer program product can include a computer readable storage medium (or media) having computer readable program instructions thereon for performing various aspects of the present application.

[0163] Finally, it should be noted that the above-described embodiments are merely intended to illustrate the technical solutions of the present application, and are not intended to limit the present application; even though the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or equivalently replace some or all of the technical features thereof; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for optimizing the cooperation and task allocation of a UAV swarm formation, characterized in that, The application relates to a method for optimizing task allocation and formation of a UAV group, comprising the following steps: acquiring flight state information and task area information of UAVs in the UAV group, determining initial spatial distribution of the UAV group based on the flight state information; dividing the task area information into multiple sub-tasks and setting initial task allocation parameters, determining communication weights of the UAVs according to communication quality and relative distance between the UAVs, constructing an information interaction mechanism between the UAVs based on the communication weights, calculating execution cost of the UAVs in executing the sub-tasks based on the initial task allocation parameters, and obtaining an optimized task allocation scheme by minimizing total task execution cost of all the UAVs, wherein the communication weights are calculated by multiplying a product of the relative distance and the communication quality by an exponential decay function; calculating an expected flight path of each UAV in executing the allocated task according to the optimized task allocation scheme, and generating initial formation position information; acquiring environmental parameters in a flight process of the UAV group based on the initial formation position information, calculating optimal distances between the UAVs based on the environmental parameters, dynamically adjusting relative positions of the UAVs according to the optimal distances, predicting aerodynamic interference between the UAVs based on relative motion of the UAVs and performing real-time compensation, and obtaining an optimized formation scheme; sending the optimized formation scheme to the UAV group for execution, monitoring obstacle information in a flight process in real time, calculating an obstacle avoidance strategy and updating the optimized formation scheme if an obstacle is monitored, and stopping until a task is completed.

2. The method of claim 1, wherein, The method comprises the following steps: acquiring flight state information of each UAV in the UAV group, wherein the flight state information comprises current position coordinates, flight speed and remaining power of each UAV; acquiring task area information, wherein the task area information comprises spatial range of a task execution area and position coordinates of a task target point; calculating spatial distances between each UAV and adjacent UAVs, and keeping a safe distance between the adjacent UAVs by adjusting flight height or horizontal position of the UAVs when the spatial distances are less than a preset safety threshold, to obtain initial spatial distribution of the UAV group.

3. The method of claim 1, wherein, The method comprises the following steps: dividing the task area information into multiple sub-tasks and setting initial task allocation parameters, determining communication weights of the UAVs according to communication quality and relative distance between the UAVs, constructing an information interaction mechanism between the UAVs based on the communication weights, calculating execution cost of the UAVs in executing the sub-tasks based on the initial task allocation parameters, and obtaining an optimized task allocation scheme by minimizing total task execution cost of all the UAVs, wherein the communication weights are calculated by multiplying a product of the relative distance and the communication quality by an exponential decay function; acquiring communication quality and relative distance between the UAVs, determining communication weights of the UAVs based on the communication quality and the relative distance, and constructing a communication weight matrix. constructing an information interaction mechanism among the UAVs based on the communication weight, calculating an execution cost of the UAVs for executing each subtask based on the initial task allocation parameter, the execution cost being calculated by a weighted sum of energy consumption, task execution time and task load degree; constructing a benefit matrix based on the communication weight matrix, the information interaction mechanism and the execution cost, and obtaining an optimized task allocation scheme by iteratively solving the benefit matrix by using the Hungarian algorithm under the conditions of each subtask being allocated to only one UAV, a maximum number of tasks of a single UAV and a remaining energy constraint of the UAVs.

4. The method of claim 3, wherein, constructing a benefit matrix based on the communication weight matrix, the information interaction mechanism and the execution cost, and obtaining an optimized task allocation scheme by iteratively solving the benefit matrix by using the Hungarian algorithm under the conditions of each subtask being allocated to only one UAV, a maximum number of tasks of a single UAV and a remaining energy constraint of the UAVs includes: setting a communication weight coefficient and a task execution cost coefficient according to a flight task of the UAV group, generating a communication influence coefficient based on an information interaction mechanism, multiplying elements in the communication weight matrix by the communication weight coefficient and the communication influence coefficient, multiplying the execution cost by the task execution cost coefficient, and adding the results to obtain a benefit matrix; determining unique constraint data that each subtask can be allocated to only one UAV, maximum task carrying quantity constraint data of each UAV, and remaining energy constraint data, obtaining communication quality evaluation data based on the information interaction mechanism among the UAVs, calculating weight increment data of adjacent UAVs according to a communication weight change rate of adjacent UAVs, multiplying the communication quality evaluation data by the weight increment data and accumulating the results to obtain state information update data; subtracting a minimum value in a current row from each element in a row of the benefit matrix, subtracting a minimum value in a current column from each element in a column of the matrix after the subtraction, determining independent zero elements in the modified benefit matrix and covering all zero elements with a minimum number of covering lines, when the number of covering lines is less than a matrix order, subtracting a minimum value in un-covered elements from un-covered elements, adding a minimum value in un-covered elements to elements covered twice, and repeating the execution to obtain an optimized task allocation scheme satisfying the constraint data.

5. The method of claim 1, wherein, calculating an expected flight route of each UAV for executing an allocated task according to the optimized task allocation scheme, and generating initial formation position information includes: obtaining a task execution sequence of each UAV in the optimized task allocation scheme, calculating a shortest flight path of each UAV based on the task execution sequence by using an ant colony algorithm, the shortest flight path satisfying a task timing constraint and a UAV turning radius constraint; calculating a flight time of each UAV to a first task execution point according to the shortest flight path, determining a key UAV as a UAV with the longest flight time, and calculating a delayed take-off time of other UAVs based on the flight time of the key UAV; determining initial position coordinates of each UAV based on the delayed take-off time, and constructing initial formation position information satisfying a minimum distance constraint of the UAVs.

6. The method of claim 1, wherein, acquiring environmental parameters in a flight process of the UAV group based on the initial formation position information, and calculating an optimal distance between the UAVs based on the environmental parameters includes: constructing a UAV group topology space based on the initial formation position information, constructing a space connection structure according to relative distance relationships between UAVs and calculating a topological invariant to obtain the environmental parameter; adopting a topological persistence graph to analyze stability of the environmental parameter and calculating a persistence metric value, constructing a topological complexity evaluation index and determining a local topological complexity corresponding to each UAV, calculating aerodynamic energy consumption generated by airflow interference between UAVs through a flight dynamics method, calculating a communication cost based on a UAV communication link quality and information transmission demand, taking the topological complexity evaluation index, the aerodynamic energy consumption and the communication cost as optimization objectives, and solving to obtain an optimal distance between UAVs.

7. The method of claim 1, wherein, dynamically adjusting relative positions of UAVs according to the optimal distance, predicting aerodynamic interference based on relative motion between UAVs and performing real-time compensation to obtain an optimized formation scheme including: calculating distance errors between adjacent UAVs based on the optimal distance, determining motion speeds and directions of UAVs according to the distance errors, correcting the motion speeds and directions based on a change trend of the local topological complexity, obtaining position adjustment instructions of the UAVs and executing the position adjustment instructions; during execution of the position adjustment instructions, analyzing a formation configuration evolution process through topological persistence analysis of the environmental parameter, extracting stable topological structure features, calculating topological invariant feature values and persistence metric feature values based on the stable topological structure features, analyzing motion trends between adjacent UAVs based on the feature values, and predicting aerodynamic interference between the UAVs; calculating an aerodynamic compensation control amount according to the aerodynamic interference, calculating a tracking compensation control amount according to a tracking error between actual positions and target positions of the UAVs, taking the aerodynamic compensation control amount and the tracking compensation control amount as control inputs of the UAVs, and obtaining the optimized formation scheme.

8. The method of claim 1, wherein, sending the optimized formation scheme to the UAV group for execution, monitoring obstacle information in a flight process in real time, calculating an obstacle avoidance strategy and updating the optimized formation scheme if an obstacle is monitored, and completing a task until the task is completed including: sending the optimized formation scheme to the UAV group, and the UAV group flying in formation according to the optimized formation scheme; monitoring obstacle information in real time during the formation flight, obtaining obstacle position coordinates according to the obstacle information, calculating a safe flight path of each UAV based on the obstacle position coordinates if an obstacle is detected, and a minimum distance between the safe flight path and the obstacle being greater than a preset safety threshold; generating an updated optimized formation scheme according to the safe flight path, the updated optimized formation scheme including a flight trajectory of the UAVs flying around the obstacle and a reconstruction formation strategy after obstacle avoidance, sending the updated optimized formation scheme to the UAV group for execution, and completing the task until the task is completed.

9. A UAV swarm formation coordination and task allocation optimization system for implementing the method of any one of the preceding claims 1-8, characterized in that, including: a first unit configured to obtain flight state information and task area information of UAVs in a UAV group, and determine an initial spatial distribution of the UAV group based on the flight state information; The second unit is configured to divide the task area information into a plurality of subtasks and set initial task allocation parameters, determine a communication weight of each unmanned aerial vehicle according to a communication quality and a relative distance between the unmanned aerial vehicles, construct an information interaction mechanism between the unmanned aerial vehicles based on the communication weight, calculate an execution cost of the unmanned aerial vehicles for executing the subtasks based on the initial task allocation parameters, obtain an optimized task allocation scheme by minimizing a total task execution cost of all the unmanned aerial vehicles, and calculate the communication weight by using an exponential decay function and a product of the relative distance and the communication quality; The third unit is configured to calculate an expected flight route of each unmanned aerial vehicle for executing the allocated tasks according to the optimized task allocation scheme, and generate initial formation position information; The fourth unit is configured to collect an environmental parameter in a flight process of the unmanned aerial vehicle group based on the initial formation position information, calculate an optimal distance between the unmanned aerial vehicles based on the environmental parameter, dynamically adjust relative positions of the unmanned aerial vehicles according to the optimal distance, predict aerodynamic interference between the unmanned aerial vehicles based on relative motions of the unmanned aerial vehicles and perform real-time compensation, and obtain an optimized formation scheme; The fifth unit is configured to send the optimized formation scheme to the unmanned aerial vehicle group for execution, monitor obstacle information in the flight process in real time, calculate an obstacle avoidance strategy and update the optimized formation scheme if the obstacle information is monitored, and perform the task until completion.

10. A computer-readable storage medium having stored thereon computer program instructions, wherein, The computer program instructions, when executed by the processor, implement the method of any one of claims 1 to 7.

Images