Method for intelligent task allocation and cooperative decision of heterogeneous cluster facing anti- unmanned aerial vehicle

Abstract

The application provides a heterogeneous cluster intelligent task allocation and cooperative decision method for unmanned aerial vehicle countermeasures, and belongs to the technical field of unmanned aerial vehicle target allocation. The application is used for solving the problems of inaccurate evaluation quantification caused by significant platform capability difference in a heterogeneous unmanned aerial vehicle cluster countermeasure system, and low solving efficiency and influence on task execution efficiency in traditional unmanned aerial vehicle task allocation and cooperative planning in a dynamic environment. The application comprises the following steps: S1, constructing a heterogeneous unmanned aerial vehicle performance evaluation model and a target importance evaluation model according to the performance attribute of the heterogeneous countermeasure platform and the state information of the task target; S2, constructing a judgment matrix to obtain the importance score of the task, constructing a comprehensive cost matrix according to the task importance score and the comprehensive performance score of the platform, and performing task allocation by means of an improved Hungarian algorithm; and S3, performing dynamic reallocation based on a contract net protocol to obtain an optimal task allocation scheme under a sudden situation.

Description

Technical Field

[0001] This invention relates to a heterogeneous swarm intelligent task allocation and collaborative decision-making method for countering unmanned aerial vehicles (UAVs), belonging to the field of UAV target allocation technology. Background Technology

[0002] With the rapid development and widespread adoption of drone technology, drones have demonstrated enormous application potential in civilian fields such as logistics delivery, aerial photography, and agricultural plant protection, while also bringing increasingly severe challenges.

[0003] Current drone threats exhibit significant characteristics such as multi-target mission execution, low-cost, high-frequency operation, and swarm collaboration. Traditional methods suffer from limitations such as limited operational range, a limited number of targets to be addressed, and low cost-effectiveness. Heterogeneous drone swarm countermeasure systems, by integrating fixed countermeasure platforms, achieve complementary advantages and collaborative mission execution, becoming an effective solution for multi-task execution in complex scenarios. However, how to achieve efficient, real-time, and robust task allocation and collaborative decision-making based on the capability differences of heterogeneous platforms and the dynamically changing importance of mission objectives has become a key challenge and urgent need for drone mission execution.

[0004] Task allocation is a key technology for achieving efficient collaborative countermeasures in heterogeneous UAV swarms. Essentially, it involves establishing the optimal mapping relationship between the platform and the target, while meeting platform capability constraints and mission requirements, to maximize overall countermeasure effectiveness. In the past five years, domestic and international scholars have conducted extensive research on the task allocation problem for UAV swarms, resulting in various technical approaches based on intelligent optimization algorithms, game theory, reinforcement learning, and distributed negotiation.

[0005] In recent years, the Hungarian algorithm and the Contract Network protocol have been widely used in the field of UAV swarm task allocation due to their unique advantages. As a classic combinatorial optimization method, the Hungarian algorithm can solve the global optimal solution of the allocation problem in polynomial time complexity O(n³), and has significant advantages such as strong algorithm determinism and high solution efficiency. Applying the Hungarian algorithm to the allocation of multi-UAV cooperative reconnaissance tasks, optimal path allocation can be achieved by constructing a distance cost matrix.

[0006] Existing methods often rely on simple distance or energy consumption as bidding criteria, failing to fully quantify the comprehensive performance of heterogeneous platforms and the importance level of targets, thus affecting the rationality of task matching and countermeasure effectiveness. Furthermore, while the Hungarian algorithm alone can obtain an initial globally optimal solution, it requires a global recalculation in dynamic environments when facing countermeasure platform failures or the emergence of new targets, resulting in slow response times and a lack of flexibility. Conversely, while the Contract Network protocol alone possesses dynamic adjustment capabilities, its poor initial allocation quality can negatively impact overall operational effectiveness. Therefore, how to organically integrate the advantages of both methods to achieve rapid and robust reallocation in dynamic scenarios while ensuring the global optimality of the initial allocation is a critical issue that urgently needs to be addressed. Summary of the Invention

[0007] This invention addresses the problems of inaccurate evaluation and quantification caused by significant differences in platform capabilities in heterogeneous UAV swarm countermeasure systems, as well as the low efficiency of traditional UAV task allocation and collaborative planning in dynamic environments, which affects task execution efficiency. Therefore, it proposes a heterogeneous swarm intelligent task allocation and collaborative decision-making method for UAV countermeasures.

[0008] The technical solution adopted by the present invention to solve the above problems is as follows: The present invention includes the following steps:

[0009] Step 1: Based on the performance attributes of the heterogeneous countermeasure platform and the status information of the mission objectives, construct a two-layer quantitative evaluation system, which includes an evaluation index system and a mission importance evaluation index system. Step 2: Obtain the comprehensive performance score of the corresponding heterogeneous countermeasure platform based on the evaluation index system; Step 3: Construct a judgment matrix based on expert scores to determine the importance of evaluation indicators. Perform normalization and consistency checks on the judgment matrix in sequence to obtain the importance score of the task objective. Step 4: Establish a comprehensive cost matrix based on the comprehensive performance score of the heterogeneous countermeasure platform and the importance score of the task objective output by the two-layer quantitative evaluation system, and solve the task allocation scheme by improving the Hungarian algorithm; Step 5: During task execution, based on Step 3, according to the dynamically triggered events, the distributed negotiation mechanism based on the Contract Network protocol is executed to realize the dynamic redistribution of tasks, so as to update the task allocation scheme and obtain the optimal task allocation scheme.

[0010] Furthermore, step 2 specifically includes: Step 2.1: Establish the original decision matrix based on the heterogeneous countermeasure platform and evaluation indicators; Step 2.2: Normalize the vectors in the original decision matrix and introduce index weight vectors. Calculate the weighted normalized matrix; Step 2.4: Calculate the positive ideal solution based on the weighted normalization matrix and negative ideal solution , where the positive ideal solution represents the best combination of all evaluation indicators, and the negative ideal solution represents the worst combination of all evaluation indicators; Step 2.5: Calculate the Euclidean distances from the corresponding heterogeneous countermeasure platform to the positive ideal solution and the negative ideal solution respectively, and calculate the comprehensive performance score of the platform based on the Euclidean distances from the corresponding heterogeneous countermeasure platform to the positive ideal solution and the negative ideal solution. The expression for the ideal solution is: (1); The expression for the negative ideal solution is: (2); The formulas for calculating the Euclidean distance from the heterogeneous countermeasure platform to the positive and negative ideal solutions are as follows: (3); (4); In formulas (4) and (5), The Euclidean distance from the heterogeneous countermeasure platform to the positive ideal solution. The Euclidean distance from the heterogeneous countermeasure platform to the negative ideal solution; The formula for calculating the overall performance score is as follows: (5).

[0011] Furthermore, step 3 specifically includes: Construct a judgment matrix A Determine the importance of the key states of the target under heterogeneous countermeasure missions; The weight vector is calculated using the column normalization-row averaging method for the judgment matrix. A Each column is normalized to obtain the weight vector. ; Calculate the consistency index CI and the consistency ratio CR, and perform consistency tests on them; For weight vector The importance score of the mission objective is obtained by weighted summation of the calculation results from the velocity, payload, and distance dimensions. ; The expression for the judgment matrix is: (6); In formula (6), For the first i The first on the platform j Expert scores for each evaluation indicator m To determine the number of rows in a matrix, nTo determine the number of columns in a matrix; The consistency index (CI) is: (7); The consistency ratio (CR) is: (8); In formulas (7) and (8), The largest eigenvalue of the judgment matrix is ​​n, the order of the judgment matrix is ​​n, and RI is the random consistency index. The expression for the velocity dimension is: (9); The expression for the load dimension is: (10); The expression for the distance dimension is: (11); In formulas (9)-(11), For the goal i speed, For the fastest countermeasure speed of heterogeneous countermeasure platforms, For the goal i The load capacity, The maximum load of the target. For the goal i Distance to the task execution area This represents the maximum detection range of the heterogeneous countermeasure platform. Task objective importance rating The calculation formula is: (12); In formula (12), The weights for speed, payload, and distance of each drone are calculated by normalizing and averaging the scores from the expert evaluation matrix. This is a dimensional index obtained by combining the enemy drone's speed, payload, and distance parameters with dimensional expression.

[0012] Furthermore, step 4 specifically includes: Step 4.1: Establish a comprehensive cost matrix based on the overall performance score, the importance score of the mission objective, and the differences in spatial distance and speed; Step 4.2: Simplify the overall cost matrix by subtracting the minimum value of each row; Step 4.3: Simplify the overall cost matrix by subtracting the minimum value of each column; Step 4.4: Cover the comprehensive cost matrix with zero elements by using the k-tuned line to cover all zero elements; Step 4.5: Determine optimality. If k = n, there is a perfect match among the zero elements, and proceed to Step 4.6; if k < n, proceed to Step 4.7; Step 4.6: Find the perfect match. Find n independent zeros in each row and each column at the zero element positions to construct the task assignment plan, and the algorithm ends; Step 4.7: Find the smallest element δ not covered by the straight lines, and adjust the matrix. Subtract δ from the elements not covered, add δ to the elements covered by two lines, and keep the elements covered by one line unchanged to obtain the adjusted comprehensive cost matrix, return to Step 4.4 and obtain the task assignment plan; The expression of the comprehensive cost matrix is: (13); In formula (13), m is the number of rows of the cost matrix, and n is the number of columns of the cost matrix.

[0013] Furthermore, the construction of the cost matrix in Step 4.1 specifically includes: Calculate the performance matching cost of the UAV according to the comprehensive performance score , where the higher the comprehensive performance score, the lower the ability matching cost; Calculate the task importance priority cost according to the importance score of the task objective , where the lower the task importance score, the lower the task importance priority cost; Calculate the time cost according to the spatial distance between the heterogeneous countermeasure platform and the UAV target, the maximum countermeasure speed of the heterogeneous countermeasure platform, and the target speed and perform normalization processing; Perform weighted summation on the performance matching cost, task importance priority cost, and normalized time cost according to the preset weights to calculate the cost function of the comprehensive cost matrix, and calculate the element values in the comprehensive cost matrix based on the cost function; The expression of the cost function is: (14); In formula (14), , and are the preset weights of the performance matching cost, task importance priority cost, and normalized time cost respectively, is the constraint compensation; The performance matching cost of the UAV is: (15); The task importance priority cost is: (16); The time cost for: (17); In formula (17), For drones u i With the goal t j Euclidean distance, To achieve the maximum countermeasure speed for heterogeneous countermeasure platforms, For the target drone, To prevent division by zero of small constants; The expression for time cost normalization is: (18); In formula (18), The maximum value of time cost. This represents the minimum time cost.

[0014] Furthermore, step 5 specifically includes: The simulated task allocation scheme triggers a task announcement phase based on events such as interception platform failure, the emergence of new targets, or interception failure, in which the administrator node broadcasts task information to the network. Based on the task information, qualified executor nodes independently calculate the bidding score and submit the bid to the manager node. The bidding score is determined by type capability, distance, power, and current status. The specific weights are adjusted according to the actual operation of the algorithm. Based on the collected bids, the management node evaluates and ranks the bids, selects the best bidder, and sends a contract award notice. According to the contract award notice, the successful bidder executes the interception task and periodically reports the execution status. The management node monitors the task progress and handles abnormal situations, realizes dynamic allocation of the plan, and obtains the optimal task allocation plan.

[0015] Furthermore, the system independently calculates bid scores and submits bids, including: Based on the current status information of the executor nodes, determine whether the platform availability and load status meet the task requirements; Based on the executor nodes that meet the requirements, a comprehensive bid score is calculated, which takes into account the platform performance matching degree, the time required to reach the target, and the current task execution status. A bid is generated based on the comprehensive bid score and submitted to the administrator node.

[0016] The beneficial effects of this invention are: 1. The two-layer evaluation model constructed using TOPSIS and AHP in this invention can accurately quantify the platform capabilities and the importance of the target; the improved Hungarian algorithm proposed in this invention can integrate the evaluation results into the comprehensive cost matrix, and achieve the global optimal allocation through iterative constraint solving, thus greatly improving the constraint satisfaction rate. 2. The dynamic reallocation strategy based on the contract network protocol designed in this invention can achieve millisecond-level fast response in fault and new target scenarios, significantly improving response speed compared to centralized methods. This method extends traditional static allocation to the field of dynamic reallocation, has strong scalability, and is more closely aligned with actual task execution scenarios. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a heterogeneous swarm intelligent task allocation and collaborative decision-making method for countering unmanned aerial vehicles (UAVs). Figure 2 This is a flowchart of the contract network protocol. Detailed Implementation

[0018] like Figure 1 As shown, the steps of the heterogeneous swarm intelligent task allocation and collaborative decision-making method for UAV countermeasures described in this embodiment include: S1: Based on the performance attributes of the heterogeneous countermeasure platform and the status information of the mission target, construct a heterogeneous UAV performance evaluation model and a target importance evaluation model; In countering heterogeneous drone swarms, accurately assessing the comprehensive capabilities of the interception platform and the target threat level is a prerequisite for efficient task allocation. Traditional methods often rely on simple weighting or expert judgment, which suffers from incomplete assessment dimensions, strong subjectivity, and difficulty in adapting to dynamic scenarios. This chapter constructs a two-layer quantitative assessment system, using the TOPSIS method to objectively evaluate the comprehensive performance of heterogeneous platforms and the AHP method to accurately assess the target threat level, providing a scientific and reliable decision-making basis for subsequent task allocation.

[0019] The TOPSIS approximation ideal solution ranking method is a classic multi-criteria decision-making method. It ranks the evaluation objects by calculating the distances between them and the positive and negative ideal solutions, which can make full use of the original data information and objectively reflect the gaps between the evaluation schemes.

[0020] This invention selects four aspects—mobility, endurance, mission execution efficiency, and economy—of a heterogeneous UAV platform to construct an evaluation index system. It includes the following steps: S101: Construct the decision matrix; Given m heterogeneous countermeasure platforms and n evaluation indicators, construct the original decision matrix; S102: Vector normalization; To eliminate dimensional differences, this implementation method employs vector normalization: (1); In formula (1), For the first i The first on the platform j The measured values ​​of each evaluation indicator; S103: Weighted normalization; Introduce indicator weight vectors, ,in: .

[0021] Calculate the weighted normalized matrix: .

[0022] S104: Determine the ideal solution; Positive ideal solution (optimal reference point): (2); Negative ideal solution (worst reference point): (3); A positive ideal solution represents the best combination of all indicators, while a negative ideal solution represents the worst combination.

[0023] S105: Calculate distance; Countermeasures Platform i Euclidean distance to the ideal solution: (4); Countermeasures Platform i Euclidean distance to the negative ideal solution: (5); S106: Overall Ability Assessment; Countermeasures Platform i The TOPSIS score (relative closeness) is: (6).

[0024] Rating Meaning: Si→1: Platform performance is close to ideal, with strong overall capabilities. Si→0: Platform performance deviates from ideal, with weak overall capabilities.

[0025] In UAV countermeasures missions, the importance of a target is influenced by various factors, including flight speed, payload capacity, and distance from the protected target. These factors include both objective data (such as radar speed measurements) and situational assessments (such as the destructive potential of the payload type), making comprehensive evaluation difficult using only mathematical models. This invention employs the Analytic Hierarchy Process (AHP) to establish a hierarchical assessment model. By integrating expert experience and measured data, it systematically quantifies threat levels, providing a scientific prioritization basis for task allocation.

[0026] This invention selects three criteria—speed characteristics, payload capacity, and distance—of the enemy target to construct a threat assessment index system. It includes the following steps: S107: Construct the judgment matrix; (7); In formula (7), For the first i The first on the platform j Expert scores for each evaluation indicator m To determine the number of rows in a matrix, n To determine the number of columns in a matrix; S108: Calculate the weight vector; The weight vector is calculated using a summation method (column normalization-row averaging). Each column of the judgment matrix A is normalized. The weight vector is then obtained by averaging the rows of the normalized matrix A.

[0027] (8); S109: Consistency check; Calculate the consistency index (CI): (9); Calculate the consistency ratio (CR): (10); Scoring based on three dimensions: Speed ​​dimension: (11); Load dimension: (12); Distance dimension: (13); In formulas (11)-(13), For the goal i speed, For the fastest countermeasure speed of heterogeneous countermeasure platforms, For the goal i The load capacity, The maximum load of the target. For the goal i Distance to the task execution area This represents the maximum detection range of the heterogeneous countermeasure platform. S1010: Solving for the comprehensive importance of the objective; Target i The overall importance H i We obtain the following by weighted summation: (14); In formula (14), The weights for speed, payload, and distance of each drone are calculated by normalizing and averaging the scores from the expert evaluation matrix. This is a dimensional index obtained by combining the enemy drone's speed, payload, and distance parameters with dimensional expression.

[0028] S2: Construct a judgment matrix to obtain the importance score of the task, construct a comprehensive cost matrix based on the task importance score and the platform's comprehensive performance score, and allocate tasks by improving the Hungarian algorithm; Based on the performance evaluation of heterogeneous UAVs and the quantification of target importance, an efficient task allocation model needs to be established to achieve optimal matching between countermeasures and targets. Traditional Hungarian algorithms primarily address one-to-one allocation problems, while UAV countermeasure scenarios have significant unique characteristics: (1) Coordinated interception requirements: A single high-importance target usually requires multiple heterogeneous UAVs to work together to deal with it; (2) Type constraints require that different task execution methods be reasonably configured to improve the success rate; (3) Multi-objective optimization requires simultaneous consideration of multiple factors such as threat priority, interception effectiveness, and time window. To address the above issues, this section proposes an improved Hungarian algorithm, which achieves optimal task allocation in many-to-one scenarios through node expansion technology, multi-factor cost function design, and constraint satisfaction mechanism.

[0029] The algorithm solution process is as follows: S201: Input the comprehensive cost matrix: .

[0030] The cost function comprehensively considers three dimensions: performance matching degree, importance and priority of task objectives, and time. (15); In formula (15), , and These are the preset weights for performance matching cost, task importance priority cost, and normalized time cost, respectively. For constraint compensation; Calculate the performance matching cost of the drone based on the comprehensive performance score. The higher the overall performance score, the lower the capability matching cost, and the lower the performance matching cost of the drone. for: (16); Calculate the cost of task importance priority based on the importance score of the task objectives. , where the lower the task importance score, the lower the task importance priority cost; the task importance priority cost is: (17); Calculate the time cost according to the spatial distance between the heterogeneous countermeasure platform and the UAV target, the maximum countermeasure speed of the heterogeneous countermeasure platform, and the target speed And perform normalization on the time cost is: (18); In formula (18), is the UAV u i and the target t j 's Euclidean distance, is the maximum countermeasure speed of the heterogeneous countermeasure platform, is the target UAV, is a small constant to prevent division by zero; The expression for normalizing the time cost is: (19); In formula (19), is the maximum value of the time cost, is the minimum value of the time cost.

[0031] The following are the specific solution steps: S202: Perform row reduction on the comprehensive cost matrix, subtract the minimum value of each row from each row; S203: Perform column reduction on the comprehensive cost matrix, subtract the minimum value of each column from each column; S204: Cover the zero elements of the comprehensive cost matrix, use k straight lines to cover all zero elements; S205: Judge the optimality. If k = n, there is a perfect match among the zero elements, and proceed to step S206; if k < n, proceed to S207; S206: Find the perfect match, find n independent zeros in each row and column at the zero element positions to construct the task allocation scheme, and the algorithm ends; S207: Find the smallest element δ not covered by the straight lines, adjust the matrix, subtract δ from the uncovered elements, add δ to the elements covered by two lines, and keep the elements covered by one line unchanged, obtain the adjusted comprehensive cost matrix, return to S204 and obtain the task allocation scheme; To ensure the efficiency and effectiveness of heterogeneous UAV task allocation, certain constraints need to be introduced during the allocation process. For example, each target allocation formation must guarantee that at least one UAV is assigned to perform the target task to increase the probability of target task success. Simultaneously, to enhance the robustness of allocation, a backup mechanism can be activated in conflict scenarios or complex dynamic situations. During the algorithm solution process, if the allocation result given by the Hungarian algorithm does not fully satisfy the type constraints or task performance requirements, a greedy constraint strategy can be switched as a safeguard. The greedy strategy selects appropriate UAV combinations to satisfy the core constraints, thereby ensuring that tasks can be allocated in a short time, reducing the possibility of allocation failure or delay. Finally, by combining the global optimization capability of the Hungarian algorithm with the fast response characteristics of the greedy strategy, the efficiency and reliability of heterogeneous UAV task allocation are further improved, enabling the system to maintain a high task completion rate in dynamic and complex environments.

[0032] S3: Dynamically reallocate tasks based on the Contract Network protocol to obtain the optimal task allocation scheme in case of emergencies; Drone missions are highly dynamic and uncertain; targets may enter the mission area at any time, and target maneuvers and environmental interference can lead to changes in mission requirements. Traditional centralized task allocation methods struggle to handle such emergencies, posing single-point-of-failure risks and response delays. ContractNet, as a distributed task allocation mechanism, achieves decentralized decision-making through a manager-bidder negotiation model, exhibiting good robustness and scalability. This invention, based on the classic CNP framework, designs a dynamic reallocation strategy for countermeasure scenarios, introducing a global allocation state maintenance, multi-dimensional bid evaluation, and constraint-aware contract granting mechanism to achieve rapid response to dynamic events such as drone malfunctions and the emergence of new targets. It includes the following four core stages: (1) Task announcement: The manager node broadcasts task information to potential executors in the network, including task type, requirements and deadline.

[0033] (2) Bid generation: Qualified executors independently calculate their bid scores based on their own status and capabilities, and submit their bids.

[0034] Based on the current status information of the executor nodes, determine whether the platform availability and load status meet the task requirements; for executor nodes that meet the requirements, calculate the comprehensive bid score, which takes into account the platform performance matching degree, the time required to reach the target, and the current task execution status; generate a bid based on the comprehensive bid score and submit it to the manager node.

[0035] (3) Contract award: The manager evaluates and ranks the collected bids, selects the best bidder and sends a notice of award.

[0036] (4) Execution Feedback: The successful bidder performs the task and periodically reports its status. The manager monitors the task progress and handles any abnormal situations. Specifically, for example... Figure 2 As shown: This mechanism achieves task allocation through distributed negotiation, avoiding the single point of failure risk of centralized methods, but its allocation quality is highly dependent on the design of the bidding strategy.

[0037] This paper proposes an intelligent task allocation method that integrates the improved Hungarian algorithm and the Contract Network protocol, forming a two-layer decision-making framework of "globally optimal initial allocation + distributed dynamic reallocation". For static scenarios, the Hungarian algorithm is improved by constructing a comprehensive cost matrix to achieve many-to-one collaborative allocation; for dynamic scenarios, a distributed reallocation strategy is designed based on the Contract Network protocol. The organic combination of the two ensures the synergistic improvement of countermeasure effectiveness, real-time performance, and robustness.

[0038] To verify the effectiveness and robustness of the proposed intelligent task allocation method, which integrates two-layer evaluation, an improved Hungarian algorithm, and a contract net protocol, in UAV countermeasures, this implementation designed multiple sets of simulation experiments. The simulation scenarios cover typical countermeasures scenarios such as initial task allocation, UAV fault reassignment, allocation upon the emergence of new targets, and dynamic coordination of multiple targets. The performance of the proposed method was evaluated across multiple dimensions, including task completion rate, task execution success rate, response time, objective function value, resource consumption, and processing of high-importance units.

[0039] Parameter settings: The simulation experiment constructed a 10km×10km two-dimensional mission execution space, which includes defense targets and a preset no-fly zone.

[0040] The drones used in the simulation experiment were classified into types 1-3. Performance was categorized into four dimensions: maneuverability, endurance, countermeasure effectiveness, and economy. Initial positions were randomly assigned within the combat space. Specific values ​​are shown in Table 1.

[0041] Table 1

[0042] The impact of factors such as the importance of the UAV executing the mission, its payload capacity, and its distance from the target being defended is shown in Table 2. Table 2

[0043] Simulation experiment scenario: (1) Static initial task allocation Number of target drones: 3 (target drone-A, target drone-B, target drone-C).

[0044] Initial number of drones: 15.

[0045] Based on the target's importance level, location, and the capabilities of our own UAVs, task allocation was completed using the proposed method. The specific allocation results are as follows: Target UAV-A assigns UAVs: [Type 1 Mission UAV-2, Type 2 Mission UAV-1, Type 3 Mission UAV-1], with a mission assignment cost of 1.4870.

[0046] Target UAV-B assigns UAVs: [Type 1 Mission UAV-4, Type 2 Mission UAV-2, Type 2 Mission UAV-4], with a mission assignment cost of 3.5154.

[0047] Target UAV - C Assigned UAVs: [Type 1 Mission UAV - 1, Type 2 Mission UAV - 5, Type 3 Mission UAV - 2], Mission Assignment Cost is 2.8770.

[0048] The total allocation cost was 7.8793, satisfying both type diversity and task constraints, demonstrating high resource allocation efficiency and providing a solid foundation for subsequent dynamic processing scenarios. In the initial task, the target allocation success rate was 100%, and the allocation constraints were met.

[0049] The types of drones participating in the collaborative formation are also appropriate. The algorithm's runtime is less than 1ms.

[0050] (2) Unmanned Aerial Vehicle (UAV) Fault Reassignment To verify whether the proposed contract network protocol can achieve rapid task redistribution and coordination in dynamic scenarios, a drone was tested in a scenario where it would exit the mission due to power depletion or communication failure.

[0051] During the interception mission of target UAV-A, UAV-2, a Type 1 mission assigned to that target, malfunctioned and was unable to continue the mission. The system initiated a reallocation mechanism via the Contract Network Protocol.

[0052] Reassignment Notice: We are issuing a request for replacements for faulty drones, prioritizing available, idle net-catching drones of the same type.

[0053] Best alternative: Select Type 1 mission drone-3 (score 0.736, idle state) to take over the mission based on the overall score.

[0054] Updated assignment results: The mission of target drone-A has been adjusted to [Type 1 mission drone-3, Type 2 mission drone-1, Type 3 mission drone-1].

[0055] In this scenario, the task redistribution completion time is less than 1ms, which is short. The system reacts quickly, allowing the task to continue. The redistribution mechanism demonstrates efficient dynamic adjustment capabilities.

[0056] (3) New targets emerge and are allocated To verify the effectiveness of the distributed contract network protocol in handling the dynamic entry of new targets, a new high-importance target, UAV-D, was added after arbitrary allocation in Scenario 1. New target parameters: high simulated importance priority, payload capacity of 8, speed of 35 km / h, and distance from the center of the protected area is only 1.5 km.

[0057] During the mission, a high-importance target, UAV-D, is located at [9, 1], requiring urgent allocation of a UAV for mission execution. The system issues a new target allocation request, including the target's importance level and constraints.

[0058] Allocation results: The drones assigned to target drone-D are [Type 1 mission drone-5, Type 2 mission drone-6, Type 3 mission drone-7].

[0059] In dynamic response, the system utilizes reserved resources to allocate suitable drones for new tasks, ensuring that targets are dealt with in a timely manner and preventing potential threats from escalating.

[0060] (4) Complex dynamic scenarios: Target UAV-B mission failure and Target UAV-E appearance The target UAV-B experienced another malfunction during its mission, and the Type 2 mission UAV-4 assigned to target UAV-B also experienced a malfunction during its mission. The system has restarted dynamic reallocation.

[0061] Based on the evaluation of the mission status and assignment scores of the impact-type UAVs, the Type 2 mission UAV-8 was selected as a replacement.

[0062] Update result: The mission of target drone-B has been updated to [Type 1 mission drone-4, Type 2 mission drone-2, Type 2 mission drone-8].

[0063] While dealing with the malfunction of target UAV-B, a new high-importance target UAV-E enters the mission area at position [1, 9]. Given the extremely limited resources available, the system adopts a dynamic transfer strategy.

[0064] The busy mission drone-3 (from target drone-A) and mission drone-2 (from target drone-B) with less impact were dispatched to intercept the new target in a coordinated manner.

[0065] Final allocation: Target UAV-E is allocated as [Type 1 mission UAV-3, Type 2 mission UAV-3, Type 2 mission UAV-2].

[0066] In this scenario, multiple tasks are concurrent and user needs are complex. The proposed method demonstrates efficient dynamic adjustment and resource scheduling capabilities, sacrificing the performance of some tasks to improve overall efficiency.

[0067] Simulation Result Analysis: In all test tasks, the method proposed in this invention achieved a good task completion rate, demonstrating strong task adaptability and task execution effectiveness. As shown in Table 3, there were no cases of task allocation failure, fully releasing the potential of cluster task execution.

[0068] As shown in Table 4, the average response time is only 1.0 millisecond, indicating that the solution can achieve extremely fast task redistribution and dynamic coordination in sudden tasks and failure scenarios. Meanwhile, the flight distance and remaining battery power performance show that the allocation mechanism can maximize resource conservation and platform endurance while ensuring task completion, thus optimizing overall resource scheduling efficiency.

[0069] The final allocation status is shown in Table 5. The allocation results fully demonstrate the algorithm's flexible scheduling capability in multi-target and multi-constraint scenarios. Not only are high-threat targets all assigned to specialized and diverse platform formations, such as assigning three heterogeneous platforms to targets C and D, but the allocation formations of all targets also meet the type and quantity constraints, achieving task collaborative optimization in multi-target scenarios. Table 3

[0070] Table 4

[0071] Table 5

[0072] As shown in Table 6, the method proposed in this invention outperforms traditional methods in all key indicators, particularly in achieving 100% task completion rate and constraint satisfaction rate. It boasts a fast response speed and minimal impact of dynamic events on previously assigned tasks, ensuring the continuous and efficient execution of global tasks. In contrast, the classic Hungarian algorithm and greedy algorithm suffer from significant performance deficiencies due to their inability to accommodate type constraints or dynamic changes. While the contract net protocol demonstrates better dynamic adaptability, its initial allocation quality and constraint guarantees still fall short of the method proposed in this invention.

[0073] Table 6

[0074] These results validate the effectiveness and operational stability of the proposed method in UAV countermeasures missions, while also demonstrating a significant type constraint satisfaction rate and system adaptability. By combining multiple algorithms, task allocation and resource scheduling in dynamic scenarios are optimized. Even in complex situations such as multi-target conflicts and overlapping UAV failures, the task can still be executed efficiently, fully verifying the system's robustness and the correctness of the algorithmic approach.

[0075] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent substitutions, and improvements made to the above embodiments without departing from the scope of the present invention, based on the technical essence of the present invention and within the spirit and principles of the present invention, shall still fall within the protection scope of the present invention.

Claims

1. A heterogeneous swarm intelligent task allocation and collaborative decision-making method for countering unmanned aerial vehicles (UAVs), characterized in that, Including: Step 1: Construct a two - layer quantitative evaluation system according to the performance attributes of the heterogeneous counter - measure platform and the status information of the task objectives. The two - layer quantitative evaluation system includes an evaluation index system and a task importance evaluation index system; Step 2: Obtain the comprehensive performance score of the corresponding heterogeneous counter - measure platform based on the evaluation index system; Step 3: Construct a judgment matrix for judging the importance degree of evaluation indexes based on expert scoring, perform normalization processing and consistency test on the judgment matrix in sequence, and obtain the importance score of the task objective; Step 4: Establish a comprehensive cost matrix according to the comprehensive performance score of the heterogeneous counter - measure platform and the importance score of the task objective output by the two - layer quantitative evaluation system, and solve the task assignment scheme through an improved Hungarian algorithm; Step 5: During the task execution process, based on Step 4, according to the dynamically triggered event, execute a distributed negotiation mechanism based on the contract net protocol to achieve dynamic re - allocation of tasks, update the task assignment scheme, and obtain the optimal task assignment scheme.

2. The heterogeneous swarm intelligent task allocation and collaborative decision-making method for UAV countermeasures as described in claim 1, characterized in that, Step 2 specifically includes: Step 2.1: Establish an original decision matrix based on the heterogeneous counter - measure platform and evaluation indexes; Step 2.2: Normalize the vectors in the original decision matrix and introduce index weight vectors. Calculate the weighted normalized matrix; Step 2.4: Calculate the positive ideal solution based on the weighted normalization matrix and negative ideal solution , where the positive ideal solution represents the best combination of all evaluation indicators, and the negative ideal solution represents the worst combination of all evaluation indicators; Step 2.5: Calculate the Euclidean distances from the corresponding heterogeneous counter - measure platform to the positive ideal solution and the negative ideal solution respectively, and calculate the comprehensive performance score of the platform based on the Euclidean distances from the corresponding heterogeneous counter - measure platform to the positive ideal solution and the negative ideal solution; The expression of the positive ideal solution is: （1）； The expression of the negative ideal solution is: （2）； The calculation formulas for the Euclidean distances from the heterogeneous counter - measure platform to the positive ideal solution and the negative ideal solution are respectively: （3）； （4）； In formulas (4) and (5), The Euclidean distance from the heterogeneous countermeasure platform to the positive ideal solution. The Euclidean distance from the heterogeneous countermeasure platform to the negative ideal solution; The calculation formula for the comprehensive performance score is: （5）。 3. The heterogeneous swarm intelligent task allocation and collaborative decision-making method for UAV countermeasures as described in claim 1, characterized in that, Step 3 specifically includes: Construct a judgment matrix A Determine the importance of the key states of the target under heterogeneous countermeasure missions; The weight vector is calculated using the column normalization-row averaging method for the judgment matrix. A Each column is normalized to obtain the weight vector. ; Calculate the consistency index CI and the consistency ratio CR, and conduct a consistency test on them; For weight vector The importance score of the mission objective is obtained by weighted summation of the calculation results from the velocity, payload, and distance dimensions. ; The expression of the judgment matrix is: （6）； In formula (6), For the first i The first on the platform j Expert scores for each evaluation indicator m To determine the number of rows in a matrix, n To determine the number of columns in a matrix; The consistency index CI is: （7）； The consistency ratio CR is: （8）； In formulas (7) and (8), The largest eigenvalue of the judgment matrix is ​​n, the order of the judgment matrix is ​​n, and RI is the random consistency index. The expression of the speed dimension is: （9）； The expression of the payload dimension is: （10）； The expression of the distance dimension is: （11）； In formulas (9)-(11), For the goal i speed, For the fastest countermeasure speed of heterogeneous countermeasure platforms, For the goal i The load capacity, The maximum load of the target. For the goal i Distance to the task execution area This represents the maximum detection range of the heterogeneous countermeasure platform. Task objective importance rating The calculation formula is: （12）； In formula (12), The weights for speed, payload, and distance of each drone are calculated by normalizing and averaging the scores from the expert evaluation matrix. This is a dimensional index obtained by combining the enemy drone's speed, payload, and distance parameters with dimensional expression.

4. The heterogeneous swarm intelligent task allocation and collaborative decision-making method for UAV countermeasures as described in claim 1, characterized in that, Step 4 specifically includes: Step 4.1: Establish a comprehensive cost matrix according to the comprehensive performance score results, the importance score results of the task objectives, and the spatial distance and speed differences; Step 4.2: Perform row reduction on the comprehensive cost matrix, subtract the minimum value of each row from each row; Step 4.3: Perform column reduction on the comprehensive cost matrix, subtract the minimum value of each column from each column; Step 4.4: Cover the zero elements of the comprehensive cost matrix, and cover all zero elements with k straight lines; Step 4.5: Judge the optimality. If k = n, there is a perfect match among the zero elements, and proceed to Step 4.6; if k < n, proceed to Step 4.7; Step 4.6: Find a perfect match, find n independent zeros in each row and each column at the zero element positions to construct the task assignment scheme, and the algorithm ends; Step 4.7: Find the minimum element δ that is not covered by the straight line, adjust the matrix, subtract δ from the elements not covered, add δ to the elements covered by two lines, and keep the elements covered by one line unchanged, obtain the adjusted comprehensive cost matrix, return to Step 4.4 and obtain the task assignment scheme; The expression of the comprehensive cost matrix is: （13）； In formula (13), m is the number of rows of the cost matrix, and n is the number of columns of the cost matrix.

5. The heterogeneous swarm intelligent task allocation and collaborative decision-making method for UAV countermeasures according to claim 4, characterized in that, The construction of the cost matrix in Step 4.1 specifically includes: Calculate the performance matching cost of the drone based on the comprehensive performance score. The higher the overall performance score, the lower the cost of capability matching; Calculate the cost of task importance priority based on the importance score of the task objectives. Among them, the lower the task importance score, the lower the cost of task importance priority; Calculate the time cost based on the spatial distance between the heterogeneous countermeasure platform and the UAV target, the maximum countermeasure speed of the heterogeneous countermeasure platform, and the target speed. And perform normalization processing; The performance matching cost, task importance priority cost, and normalized time cost are weighted and summed according to preset weights to calculate the cost function of the comprehensive cost matrix, and the element values ​​in the comprehensive cost matrix are calculated based on the cost function. The expression for the cost function is: （14）； In formula (14), , and These are the preset weights for performance matching cost, task importance priority cost, and normalized time cost, respectively. For constraint compensation; Performance matching cost of drones for: （15）； Task importance priority cost for: （16）； In formula (16), The importance of the goal; Time cost for: （17）； In formula (17), For drones u i With the goal t j Euclidean distance, To achieve the maximum countermeasure speed for heterogeneous countermeasure platforms, For the target drone, To prevent division by zero of small constants; The expression for time cost normalization is: （18）； In formula (18), The maximum value of time cost. This represents the minimum time cost.

6. The heterogeneous swarm intelligent task allocation and collaborative decision-making method for UAV countermeasures according to claim 4, characterized in that, Step 5 specifically includes: The task allocation scheme is simulated, and a task announcement phase is triggered based on events such as interception platform failure, the appearance of a new target, or interception failure, in which the manager node broadcasts task information to the network. Based on the task information, qualified executor nodes independently calculate the bidding score and submit the bid to the manager node. The bidding score is determined by type capability, distance, power, and current status. The specific weights are adjusted according to the actual operation of the algorithm. Based on the collected bids, the management node evaluates and ranks the bids, selects the best bidder, and sends a contract award notice. According to the contract award notice, the successful bidder executes the interception task and periodically reports the execution status. The management node monitors the task progress and handles abnormal situations, realizes dynamic allocation of the plan, and obtains the optimal task allocation plan.

7. The heterogeneous swarm intelligent task allocation and collaborative decision-making method for UAV countermeasures according to claim 6, characterized in that, Calculate bid scores independently and submit bids, including: Based on the current status information of the executor node, determine whether the platform availability, load status, and power supply meet the task requirements; Based on the executor nodes that meet the requirements, a comprehensive bid score is calculated, which takes into account the platform performance matching degree, the time required to reach the target, and the current task execution status. A bid is generated based on the comprehensive bid score and submitted to the administrator node.

Images