Staged unmanned cluster cooperation efficiency dynamic evaluation method

Abstract

The invention discloses a staged unmanned cluster collaborative efficiency dynamic evaluation method, which comprises the following steps of: dividing an unmanned cluster into different task stages so as to analyze collaborative efficiency in the different task stages, constructing an index system of each task stage, and determining an overall target of evaluation and a hierarchical structure of the index system; determining the definition of each index, the evaluation content and the function of each index in the unmanned cluster task, and determining an index calculation method and data required by calculation; constructing a judgment matrix and carrying out weight calculation; dynamically evaluating the staged unmanned cluster cooperation efficiency; and calculating the cooperative efficiency and the stage efficiency of the unmanned cluster according to the efficiency value at each simulation moment. According to the method, the efficiency of each stage can be dynamically evaluated in real time, effective support can be provided for unmanned cluster task planning and optimization, and the change of the collaborative efficiency can be reflected constantly in the whole task execution process.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned cluster effectiveness evaluation, and in particular to a phased unmanned cluster collaborative effectiveness dynamic evaluation method. Background Art

[0002] Unmanned swarm collaboration refers to the collaborative completion of group tasks by individuals within a swarm, without human intervention. Unmanned swarms can employ a variety of collaborative algorithms when performing tasks, such as path planning, formation control, collaborative navigation and positioning, task allocation, and swarm networking and communication. These collaborative algorithms enable unmanned swarms to effectively execute complex group tasks such as detection, obstacle avoidance, repelling, and rounding up. During collaborative tasks, efficient cooperation between individuals helps improve task completion and the overall effectiveness of the swarm. With the continuous advancement of unmanned swarm technology, evaluating the collaborative effectiveness of swarms during mission execution has become a critical issue that needs to be addressed. Effective collaborative effectiveness evaluation not only helps optimize the design of unmanned swarms but also provides a basis for mission strategy formulation.

[0003] In the modern field of performance evaluation, with the development of machine learning and artificial intelligence technologies, the use of artificial neural networks (ANNs) for performance evaluation has become a growing trend. ANNs possess powerful learning and generalization capabilities, enabling them to automatically identify complex patterns and features in unmanned swarm collaboration tasks. However, ANN performance evaluation methods also have limitations. For example, neural networks require extensive data training, and in practical applications, learning and adjustment often require extensive experimentation, which undoubtedly increases the complexity and experimental cost of the evaluation process.

[0004] Furthermore, traditional static evaluation methods cannot meet the needs of unmanned swarms in dynamic mission environments. Unmanned swarm missions are typically performed in a constantly changing environment, with mission requirements and individual states fluctuating over time. Static evaluation methods cannot reflect the real-time changes in performance during the mission. Therefore, dynamically evaluating the collaborative effectiveness of unmanned swarms during mission execution has become an important research direction. Summary of the Invention

[0005] In view of the problems existing in the prior art, the present invention discloses a phased method for dynamically evaluating the collaborative effectiveness of unmanned swarms, which specifically includes the following steps:

[0006] S1. Divide the task into several stages according to the target system to be evaluated;

[0007] Furthermore, step S1 divides the system to be evaluated into several task phases in order to more carefully monitor and evaluate the changes in the system's collaborative effectiveness under different task states, specifically including:

[0008] S1.1. Based on task type: Different types of tasks may contain multiple subtasks, each with different objectives and execution strategies. Dividing complex tasks into several stages and evaluating the effectiveness of each stage separately helps understand the system's performance in different tasks.

[0009] Furthermore, the step S1.1 specifically includes:

[0010] S1.1.1. Detection phase: The unmanned swarm conducts preliminary search and detection of the target;

[0011] S1.1.2, Approach phase: The swarm approaches the target area and adjusts its formation or position;

[0012] S1.1.3, Attack or interference phase: The clusters work together to perform the main task, such as attack, expel or interfere;

[0013] S1.1.4, retreat phase: after completing the mission, the cluster evacuates safely;

[0014] S1.2, based on time period division: If the task is carried out continuously, the entire task cycle can be divided into several time periods according to the time axis of task execution;

[0015] Furthermore, the step S1.2 specifically includes:

[0016] S1.2.1, Initial phase (0-T1): System initialization and task preparation;

[0017] S1.2.2, mid-term (T1-T2): the main execution part of the task, with frequent collaboration between clusters;

[0018] S1.2.3, Termination phase (T2-task completion): status finishing and exit after the task is completed;

[0019] S1.3. Division based on environmental changes: Unmanned swarm missions are often performed in dynamic environments, where environmental changes may affect the swarm's collaborative effectiveness. Dividing mission phases based on environmental changes (such as the emergence of enemy forces, weather changes, terrain changes, etc.) can assess the system's adaptability in different environments.

[0020] Furthermore, the step S1.3 specifically includes:

[0021] S1.3.1, Static environment stage: The environment is relatively stable during task execution, and the movement and coordination of the cluster are relatively simple;

[0022] S1.3.2, Complex environment stage: When obstacles, hostile targets or other interference factors appear, the cluster needs to make dynamic adjustments;

[0023] S1.4. Division based on system status: The internal state of the unmanned swarm (such as energy level, communication status, individual damage, etc.) will affect its collaborative effectiveness. Different mission phases can be divided according to changes in system status;

[0024] Furthermore, the step S1.4 specifically includes:

[0025] S1.4.1, High-efficiency collaboration stage: All individuals have sufficient energy, smooth communication, and good performance;

[0026] S1.4.2, Limited collaboration stage: Some individuals fail or have insufficient energy, communication is poor, and performance is reduced;

[0027] S1.4.3, Critical coordination stage: Multiple units are damaged, the mission faces the risk of failure, and coordination capabilities are limited;

[0028] S1.5. Division based on changes in mission objectives: When the mission objective changes (e.g., switching from search to tracking, or from defense to attack), the mission can be divided into different phases to evaluate the system's performance under different mission objectives.

[0029] Furthermore, the step S1.5 specifically includes:

[0030] S1.5.1. Task transition phase: The transition phase when the task objectives change and the cluster adjusts its strategy and formation;

[0031] S1.5.2, Secondary Target Execution Phase: Dealing with non-primary mission objectives, such as disrupting peripheral threats to the cluster;

[0032] S1.5.3, Main goal execution phase: The cluster concentrates its efforts on executing core tasks;

[0033] S1.6. Action Division Based on Unmanned Swarms: Unmanned swarms typically need to perform a series of actions (such as path planning, formation, coordinated strikes, etc.). The tasks can be divided into different action phases, and the collaborative effectiveness of each action phase can be evaluated.

[0034] Furthermore, the step S1.6 specifically includes:

[0035] S1.6.1, Path planning phase: The cluster calculates and executes the optimal path, avoiding obstacles;

[0036] S1.6.2, Formation Adjustment Phase: The cluster reorganizes its formation to adapt to new mission requirements or terrain;

[0037] S1.6.3, Target attack phase: Clusters coordinate to attack enemy targets;

[0038] S1.7. Division based on mission complexity and key nodes: In some complex missions, key nodes in the mission process (such as reaching a specific location, destroying a specific target, breaking through a defense line, etc.) determine the changes in system effectiveness. The mission can be divided into several stages based on these key nodes;

[0039] Furthermore, the step S1.7 specifically includes:

[0040] S1.7.1. Pre-critical node 1 phase: Cluster preparation and collaboration before approaching the first target;

[0041] S1.7.2, key node 1 to 2 phase: the cluster passes the first target and executes the new target task;

[0042] S1.7.3, post-critical phase 3: the final sprint or finishing work before the end of the task;

[0043] S2. Construct an indicator system for each stage;

[0044] Furthermore, in step S2, constructing an indicator system for each stage, it is necessary to decompose the complex performance evaluation problem into multiple levels and indicators, specifically including:

[0045] S2.1. The overall evaluation objective must be clearly defined, namely, evaluating the collaborative effectiveness of the unmanned swarm at each stage. This objective is the highest level of the indicator system, and all evaluation work is geared towards achieving this objective.

[0046] S2.2. Determine the hierarchical structure of the indicator system;

[0047] Furthermore, in step S2.2, the evaluation method divides the complex problem into multiple levels, including:

[0048] S2.2.1. Target layer (top layer): This is the final evaluation goal - the collaborative effectiveness of the unmanned swarm at each stage;

[0049] S2.2.2, Criteria layer (first-level indicators): Based on the characteristics of the collaborative tasks of unmanned swarms, the criteria layer usually includes several main criteria that affect performance;

[0050] S2.2.3, Indicator layer (secondary indicators): Under each criterion, it is refined into specific quantifiable indicators. The indicator layer is the lowest level of specific measurement indicators used to evaluate the performance of each criterion;

[0051] S3. Determine the indicator calculation method and the data required for the calculation;

[0052] Furthermore, in step S3, the calculation method of each indicator needs to take into account the relevant parameters and influencing factors in the actual task, including:

[0053] S3.1. Clarify the definition of each indicator, the content of the evaluation, and its role in unmanned swarm missions. Different stages of the mission focus on different collaborative effectiveness indicators;

[0054] S3.2 When determining the calculation method for each indicator, consider how to quantify these indicators and what data are needed to perform the calculation. The calculation methods for different indicators may involve various calculation models such as geometry, physics, and statistics;

[0055] S4, construct a judgment matrix and perform weight calculation;

[0056] Furthermore, the step S4 specifically includes:

[0057] S4.1. Compare the importance of each pair of all indicators corresponding to the first capability indicator of the indicator layer at a certain stage, and obtain the n*n order judgment matrix A = (a ij ), where a ij For the total efficiency U, U i U j The importance of is scored using the nine-scale method. The judgment matrix A after expert scoring is:

[0058]

[0059] S4.2, solve the eigenvalue and eigenvector of the judgment matrix A, and take the maximum eigenvalue λ max The corresponding eigenvector is normalized to obtain the weight index W i ;

[0060] S4.3. Perform consistency test. The consistency index CI is:

[0061] CI=(λ max -n) / (n-1)

[0062] Where n is the matrix order;

[0063] S4.4. Determine whether the consistency index is acceptable based on the consistency ratio CR. The CR formula is as follows:

[0064] CR=CI / RI

[0065] When CR<0.10, the consistency of the judgment matrix is considered acceptable, otherwise the judgment matrix should be appropriately modified, where RI is the average random consistency index;

[0066] S4.5. For the weights of the indicators corresponding to the remaining capability indicators in the indicator layer of this stage, the calculation steps are the same as S4.1 to S4.4, except that the establishment of the judgment matrix A is changed to the comparison of the importance of each pair of indicators corresponding to the capability indicator.

[0067] S4.6. For the weights of the criterion layer, the calculation steps are the same as S4.1 to S4.4, except that the establishment of the judgment matrix A is changed to the comparison of the importance of each pair of capability indicators at this stage.

[0068] S5. Dynamically evaluate the collaborative effectiveness of unmanned swarms at different stages;

[0069] Furthermore, the step S5 specifically includes:

[0070] S5.1, standardization of indicators;

[0071] Furthermore, the step S5.1 specifically includes:

[0072] S5.1.1 Indicators are categorized as benefit-based and cost-based. Indicators where we prefer the calculated value to be as large as possible are considered benefit-based, while indicators where we prefer the calculated value to be as small as possible are considered cost-based.

[0073] For quantitative benefit indicators, the evaluation value x can be determined according to the evaluation criteria. ij , and then further normalize it. The normalized formula is as follows:

[0074]

[0075] Similarly, for quantitative cost indicators, the normalized formula is as follows:

[0076]

[0077] where x ij is the secondary evaluation index u ij The initial value of i represents the number of the first-level indicator, j represents the number of the second-level indicator, max{x ij}、min{x ij} are the maximum and minimum values of similar indicators. ij for u ij Normalized, dimensionless value.

[0078] S5.2. Establish a mathematical model for evaluation;

[0079] Furthermore, the step S5.2 specifically includes:

[0080] S5.2.1. To make fuzzy judgments on the elements in the indicator layer, we need to establish a comment set and, combined with the comment levels, obtain a judgment matrix. Considering the characteristics of the indicators in the factor set, we establish the comment set V = {v1, v2, v3, v4, v5} = {Excellent, Good, Average, Poor, Bad}. The comments are divided into five levels, and the corresponding comments, level intervals, and interval medians are shown in Table 1A. The data in Table 1 satisfies 1>a>b>c>d>0:

[0081] Table 1

[0082]

[0083] Normalize the indicators to get the indicator value c of the indicator layer ij If c ij As a quantitative indicator, c can be obtained according to the level interval of the judgment set. ij By judging each secondary indicator in the first-level evaluation index against the comment set, we can get the evaluation matrix R of the secondary indicator i ;

[0084] S5.2.2. Use the weight vector and evaluation matrix to comprehensively evaluate the primary indicators and performance. The operation between the weight vector and the evaluation matrix is a composite operation, and its typical models include "weighted average", "various factors determination", and "hybrid". The weighted average model balances all factors according to their weights and is more suitable for situations requiring overall indicators. The weighted average algorithm is used here;

[0085] S5.3, graded assessment;

[0086] Furthermore, the step S5.3 specifically includes:

[0087] S5.3.1. First, conduct a comprehensive evaluation of the secondary indicators. The secondary evaluation results are calculated based on the weighted average algorithm as follows:

[0088] B i =W i ·R i

[0089] The various capabilities of the criterion layer are calculated as follows:

[0090] C=[a / 2,(b+c) / 2,(d+e) / 2,(f+g) / 2,h / 2)]

[0091] S5.3.2. Evaluate the first-level indicators, where the number of second-level indicators is n. The performance is calculated as follows:

[0092]

[0093] S6. Comprehensively evaluate the overall synergy effectiveness and stage effectiveness.

[0094] Furthermore, the process of comprehensively evaluating the global collaborative effectiveness and stage effectiveness in step S6 requires a systematic approach and reasonable calculation method to accurately reflect the overall collaborative effectiveness of the unmanned swarm during the entire mission. The algorithm obtains an effectiveness value at each simulation moment and calculates the collaborative effectiveness and stage effectiveness based on the effectiveness value at each simulation moment, specifically including:

[0095] S6.1. Calculate stage performance: At the end of each stage, calculate the stage performance. The calculation methods include weighted average and simple average.

[0096] (1) Simple average method:

[0097]

[0098] This approach is suitable when the stages are of equal importance or when there is no clear priority.

[0099] (2) Weighted average method:

[0100]

[0101] Among them, E stagej Represents the effectiveness of a certain evaluation, E global represents the stage effectiveness, N is the number of evaluations within the stage, and the weight W stagej This can be determined based on factors such as the importance of the tasks at each stage, the criticality of the task's success, and the duration of the task. For example, in an unmanned swarm mission, the swarm collaborative execution stage may have a greater weight than the preparation stage. Weighting can be performed using a nine-degree scoring method for each stage or using time-weighted performance. The nine-degree scoring method for each stage is the same as in S4.1 to S4.4. The time-weighted performance is calculated as follows:

[0102]

[0103] Where Δt j Indicates the duration of the jth stage. If the stage duration is longer, the performance value of this stage will have a greater impact on the global performance.

[0104] By adopting the above-mentioned technical solution, the present invention provides a phased dynamic evaluation method for the collaborative effectiveness of unmanned swarms. This method divides the tasks of the unmanned swarm into multiple stages, meticulously monitoring and evaluating the changes in the collaborative effectiveness of the system under different task states. This method can evaluate the collaborative effectiveness of the unmanned swarm at each stage of the task execution, not only dynamically tracking the completion of tasks at each stage, but also accurately reflecting the collaborative effectiveness performance of the unmanned swarm at different stages. This phased evaluation enables in-depth analysis of the different collaborative capabilities of the unmanned swarm at each stage. BRIEF DESCRIPTION OF THE DRAWINGS

[0105] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0106] Figure 1 Schematic diagram of the phased unmanned swarm collaboration efficiency dynamic evaluation method of the present invention;

[0107] Figure 2 Schematic diagram of the indicator system in the first stage of the embodiment of the present invention;

[0108] Figure 3 Schematic diagram of the indicator system in the second stage of the embodiment of the present invention;

[0109] Figure 4 Schematic diagram of the indicator system for the third stage in an embodiment of the present invention. DETAILED DESCRIPTION

[0110] To make the technical solutions and advantages of the present invention more clear, the technical solutions in the embodiments of the present invention are clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention:

[0111] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0112] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0113] like Figure 1The method for dynamically evaluating the collaborative effectiveness of unmanned swarms in stages includes the following steps:

[0114] S1. The target system to be evaluated is a continuous mission. The phases can be divided based on time periods into the perception and navigation phase, the strike phase, and the return and re-formation phase.

[0115] S1.1, Perception and Navigation Stage:

[0116] During this phase, the unmanned vessel needs to enter the target area or mission execution area, facing challenges such as environmental perception, navigation, and obstacle avoidance. Its perception system plays a key role in this phase, creating conditions for the subsequent strike phase by identifying the surrounding environment, detecting potential threats, navigating, and avoiding obstacles. Specific process:

[0117] S1.1.1. The unmanned vessel starts and relies on the perception system to collect environmental data.

[0118] S1.1.2. The navigation system plans the optimal route to efficiently enter the target area while ensuring the safety of the path and avoiding obstacles and threats.

[0119] S1.1.3. Ensure the unmanned vessel arrives safely at the strike area and completes mission preparation through path tracking and obstacle avoidance algorithms.

[0120] S1.2, Strike Phase:

[0121] The strike phase is the critical moment for the UAV system to perform its core mission. The system needs to quickly identify, aim at, and attack the target, which involves the coordination of weapon control and tactical execution. The effectiveness evaluation of this phase directly reflects the tactical execution capability of the UAV system. Specific process:

[0122] S1.2.1. When the unmanned vessel enters the preset attack range, the target identification and tracking system is activated.

[0123] S1.2.2. When using the strike module for precision strikes, the system must keep the target stable and execute the attack command.

[0124] S1.2.3. The system records the strike effect and adjusts the subsequent mission path to prepare for the return phase.

[0125] S1.3, Return and Re-formation Phase:

[0126] This phase aims to assess the UAV's ability to safely return to port and recover after a mission, especially when multiple UAVs are working together, where reorganization to prepare for subsequent missions is particularly important. Performance at this stage directly impacts the UAV's survivability and potential for multiple uses. Specific process:

[0127] S1.3.1. After completing the mission, the unmanned vessel begins to return home and activates the navigation and obstacle avoidance systems to ensure safety.

[0128] S1.3.2. If multiple unmanned vessels participate in a mission, the system must update their positions in real time and maintain an appropriate distance from their teammates to achieve coordinated return.

[0129] S1.3.3. The unmanned vessel returns safely and returns to its initial state, ready to re-form a team and enter the next mission or standby state.

[0130] S2. Based on the fuzzy comprehensive evaluation method, the indicator system of each stage is constructed. It is necessary to consider whether each indicator has practical significance in the system. The three stages of this example are constructed according to the target layer, criterion layer, and indicator layer. The specific indicator system is shown in Figure 2 、 Figure 3 、 Figure 4 .

[0131] S3. Calculation method for determining indicators:

[0132] Perception and navigation stage:

[0133] S3.1, traversal cycle:

[0134]

[0135] S3.2, Target type recognition accuracy:

[0136]

[0137] S3.3, Target number recognition accuracy:

[0138]

[0139] S3.4. Target positioning accuracy:

[0140]

[0141] S3.5. Assembly Time:

[0142] Assembly time = T 编队集结2 -T 编队集结1 , formation assembly end time T 编队集结2 , Formation assembly start time T 区域搜索 1S3.6, Configuration maintains scale:

[0143] The number of nodes N that can maintain state in the configuration 构型保持节点数 S3.7, Track tracking error:

[0144] N is the number of unmanned ships S3.8, and the tracking error of the flight phase angle is:

[0145] The heading angle tracking error of the j-th unmanned boat for the i-th time:

[0146] Calculate the i-th heading angle tracking error: S3.9, speed tracking error:

[0147] The speed tracking error of the j-th unmanned boat at the i-th time:

[0148] Calculate the i-th velocity tracking error: Attack phase:

[0149] S3.10. Hit probability:

[0150]

[0151] S3.11. Strike Accuracy:

[0152]

[0153] S3.12, Dynamic Reconstruction Time:

[0154] Dynamic reconstruction time = T 编队重构2 -T 编队重构1 ,Reconstruction start time T 编队重构1 , reconstruction end time T 编队重构2

[0155] S3.13, Configuration holding scale, track tracking error, heading angle tracking error, speed tracking error:

[0156] The calculation method of these four indicators in this stage is the same as S3.6 to S3.9 in the specific implementation method.

[0157] Return and re-formation phase performance:

[0158] S3.14, assembly time, configuration holding scale, track tracking error, heading angle tracking error, and speed tracking error are the same as S3.5 to S3.9 in the specific implementation manner, and dynamic reconstruction time is the same as S3.12 in the specific implementation manner.

[0159] S4. Construct a judgment matrix and calculate weights

[0160] S4.1. Taking the perception and navigation stage as an example:

[0161] The importance of all the secondary indicators (traversal cycle, target type recognition accuracy, target number recognition accuracy, and target positioning accuracy) corresponding to the first capability indicator of the indicator layer of this stage is compared pairwise to obtain a 4*4 order judgment matrix A=(a ij ), where a ij For the total efficiency U, U i U j The importance of is scored using the nine-scale method. The judgment matrix A after expert scoring is (assuming that the indicator layer has 3 indicators):

[0162]

[0163] Similarly, the first-level indicators (perception ability and control ability) of the perception and navigation stages are weighted, and the importance of these two first-level indicators is compared pairwise to obtain a 2*2 order judgment matrix.

[0164] For all the first-level and second-level indicators in the strike phase, return phase and re-formation phase, corresponding judgment matrices are also established first.

[0165] S4.2. Solve the eigenvalue and eigenvector of the judgment matrix A corresponding to all indicators, and take the maximum eigenvalue λ max =4.011, the corresponding eigenvectors are normalized to obtain weight indices W1=[0.227, 0.423, 0.227, 0.1224], W2=[0.375, 0.125, 0.125, 0.125, 0.25], and W=[0.5, 0.5];

[0166] S4.3. Perform consistency test. The consistency index CI is:

[0167] CI=(λ max -n) / (n-1)=0.00351

[0168] For all secondary indicators (traversal cycle, target type recognition accuracy, target number recognition accuracy, and target positioning accuracy) corresponding to the first capability indicator of the indicator layer of the perception and navigation phase, n is the matrix order and is 4;

[0169] Consistency testing is also required for other indicators, and the testing process is the same.

[0170] S4.4. Determine whether the consistency index is acceptable based on the consistency ratio CR. The CR formula is as follows:

[0171] CR=CI / RI=0.0039<0.1

[0172] The consistency test is performed, where RI is the average random consistency index. The value of RI is shown in Table 2 and is related to the order n of the judgment matrix A. Among all the secondary indicators (traversal cycle, target type recognition accuracy, target number recognition accuracy, and target positioning accuracy) corresponding to the first capability indicator of the indicator layer in the perception and navigation stages, n is 0.90.

[0173] Table 2

[0174]

[0175] Make corresponding values for other indicators n and judge whether the consistency indicators are acceptable.

[0176] S5. Dynamically evaluate the collaborative effectiveness of unmanned vessels at different stages;

[0177] S5.1, standardization of indicators;

[0178] All secondary indicators in the perception and navigation stages - traversal cycle, target type recognition accuracy, target quantity recognition accuracy, target positioning accuracy, assembly time, configuration retention scale, track tracking error, heading angle tracking error, speed tracking error, are classified according to benefit type and cost type.

[0179] Among them, the target type recognition accuracy, target quantity recognition accuracy, and configuration retention scale are classified as benefit indicators; the traversal cycle, target positioning accuracy, assembly time, track tracking error, heading angle tracking error, and speed tracking error are classified as cost indicators.

[0180] The traversal period is a cost-based indicator, and its value is generally between 0.03-0.07ms. The traversal period of a system simulation in the perception and navigation phase of this system is 0.0416ms. The standardization of this indicator is as follows

[0181]

[0182] max{x ij}、min{x ij} is the maximum and minimum value of the traversal cycle

[0183] S5.2. Establish a fuzzy comprehensive evaluation mathematical model;

[0184] In order to make fuzzy judgments on the elements in the indicator layer, it is necessary to establish a comment set and combine the comment levels to obtain a judgment matrix.

[0185] Normalize the indicators to get the indicator value c of the indicator layer ij =0.71, when the data scoring level is v3, the judgment vector is [0,0,1,0,0].

[0186] The results of normalizing each secondary indicator of the perception and navigation phases of this instance are 0.71, 0.96, 0.73, and 0.98; the results of normalizing each secondary indicator of the control capability are 0.81, 0.66, 0.79, 0.73, and 0.58. By comparing the evaluation set, we can obtain the evaluation matrix R1 and R2 of the secondary indicators:

[0187]

[0188] The same applies to the judgment matrices in other stages.

[0189] S5.3, graded assessment;

[0190] A comprehensive evaluation of the perception and navigation phases of this instance was conducted, and the first-level evaluation result was obtained based on the weighted average algorithm as follows:

[0191] B1=W1·R1=[0.5454,0,0.454,0,0]

[0192] Similarly, we can calculate B2 = [0, 0.5, 0.25, 0.25, 0]. The various capabilities of the criterion layer are calculated as follows:

[0193] C=(0.95,0.84,0.69,0.5,0.2)

[0194] B1'=B1·C T =0.8314, B2'=B2·C T =0.7075

[0195] The evaluation result of perception ability is B1', and the evaluation result of control ability is B2'. According to Table 1, the evaluation result of perception ability is good, and the evaluation result of control ability is average.

[0196] The effectiveness of the perception and navigation stages is evaluated and calculated as follows:

[0197]

[0198] B=W·R=[0.2727,0.25,0.352,0.125,0], B′=B·C T =0.7745

[0199] B' is the performance of the perception and navigation phase. Table 1 shows that the calculated results are within the range of [0.78, 0.6], indicating that the performance of the perception and navigation phase is fair.

[0200] The assessment of the strike phase and the return and re-formation phases is the same as the perception and transit phase. S6. Comprehensively evaluate the overall coordination effectiveness and phase effectiveness.

[0201] In this example, the weighted average method is used to evaluate the global synergy effectiveness, where E stagej Represents the effectiveness of a certain evaluation, E global represents the stage performance, and N is the number of evaluations within the stage:

[0202]

[0203] Among them, the weight W stagej Using the nine-degree scoring method, the importance of these three stages is compared in pairs to obtain a third-order judgment matrix. The eigenvector corresponding to the maximum eigenvalue of this matrix is calculated and normalized to obtain the weight vector of the stage.

[0204] In this example, each stage is not further subdivided, so the stage performance adopts the simple average method:

[0205]

[0206] The present invention provides a phased dynamic evaluation method for the collaborative effectiveness of unmanned swarms. By evaluating the mission process of the unmanned swarm in phases, the unmanned swarm system is modeled and analyzed in phases, and the collaborative effectiveness value of each phase is calculated using a fuzzy comprehensive evaluation method. The effectiveness evaluation results of each phase can accurately reflect the collaborative performance of the unmanned swarm in mission execution. At the same time, by comprehensively evaluating the effectiveness values of each phase, the global collaborative effectiveness can be obtained. The present invention can not only dynamically evaluate the effectiveness of each phase in real time, but also provide effective support for unmanned swarm mission planning and optimization, reflecting the changes in collaborative effectiveness at all times during the entire mission execution process.

[0207] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A phased dynamic evaluation method for unmanned swarm collaboration effectiveness, characterized by include: Divide the unmanned swarm into different mission stages to analyze the collaborative effectiveness in different mission stages. Construct an indicator system for each task stage and clearly define the overall evaluation objectives and the hierarchical structure of the indicator system, where the hierarchy includes the goal layer, the criteria layer, and the indicator layer; Determine the definition of each indicator, the content of the evaluation, and its role in unmanned swarm missions, as well as the indicator calculation method and the data required for calculation; Construct a judgment matrix for the indicator layer and criterion layer of each stage and calculate the weights, and perform consistency test on the judgment matrix; Each indicator is normalized to obtain a normalized indicator value, and a comment set is established to evaluate the level of elements in the indicator layer. The collaborative effectiveness of the unmanned swarm at each moment is dynamically evaluated in stages. Based on the efficiency value of the unmanned cluster at each moment, the collaborative efficiency and stage efficiency of the unmanned cluster are comprehensively calculated.

2. The method for dynamic evaluation of unmanned swarm collaboration effectiveness in stages according to claim 1, characterized in that: When dividing the unmanned swarm into stages: based on the mission type, it is divided into the detection stage, approach stage, attack or interference stage, and retreat stage; based on the time period, it is divided into the initial stage, mid-term stage, and termination stage; based on environmental changes, it is divided into the static environment stage and the complex environment stage; based on the system state, it is divided into the efficient coordination stage, the restricted coordination stage, and the critical coordination stage; based on the change of mission objectives, it is divided into the mission conversion stage, the secondary objective execution stage, and the primary objective execution stage; based on the actions of the unmanned swarm, it is divided into the path planning stage, the formation adjustment stage, and the target strike stage; based on the complexity of the mission and the key nodes, it is divided into the pre-key node 1 stage, the key node 1 to 2 stage, and the post-key node 3 stage.

3. The phased unmanned swarm collaboration efficiency dynamic evaluation method according to claim 1, characterized in that: When constructing the judgment matrix for the indicator layer and criterion layer of each stage: compare the importance of all indicators corresponding to the first capability indicator of the indicator layer of a certain stage, and obtain the n*n order judgment matrix A=(a ij ), where a ij For the total efficiency U, U i U j The importance of is scored using the nine-scale method. The judgment matrix A after expert scoring is: Solve the eigenvalue kernel eigenvector of the judgment matrix A and take the maximum eigenvalue λ max The corresponding eigenvector is normalized to obtain the weight index W i .

4. The phased unmanned swarm collaboration efficiency dynamic evaluation method according to claim 1, characterized in that: Indicators are divided into benefit indicators and cost indicators. For quantitative benefit indicators, the evaluation value x is determined according to the evaluation criteria. ij , and then perform the following normalization processing: For quantitative cost indicators, the normalized formula is as follows: where x ij is the secondary evaluation index u ij The initial value of i represents the number of the first-level indicator, j represents the number of the second-level indicator, max{x ij }、min{x ij } are the maximum and minimum values of similar indicators.

5. The phased unmanned swarm collaboration efficiency dynamic evaluation method according to claim 1 is characterized by: When calculating stage performance: At the end of each stage, use the simple average method or weighted average method to calculate the stage performance: The simple averaging method is: The weighted average method is: Determine the weight W based on the importance of tasks at each stage, the criticality of task success, and the length of time. stagej , The time-weighted performance is calculated as follows: Among them, E stagej Represents the effectiveness of a certain evaluation, E global represents the stage effectiveness, N is the number of evaluations within the stage, Δt j Indicates the duration of the jth stage.

Images