Intelligent unmanned aerial vehicle swarm performance evaluation method

Abstract

The invention relates to an intelligent unmanned aerial vehicle colony performance evaluation method, belongs to the technical field of unmanned aerial vehicle colony control, and solves the problem that an existing unmanned aerial vehicle colony performance evaluation method is one-sided, poor in environmental adaptability and insufficient in accuracy. The invention discloses an intelligent unmanned aerial vehicle swarm performance evaluation method. The method comprises the following steps: collecting original data of support parameters, emergence parameters and efficiency parameters; comprehensively evaluating the support parameters by adopting a dynamic grey analytic hierarchy process to generate a support comprehensive score; calculating a comprehensive score of the emergencies based on the emergencies parameters; inputting the efficiency parameters into a pre-trained efficiency prediction model, and predicting an efficiency comprehensive score; and obtaining a performance evaluation result based on the support comprehensive score, the emergence comprehensive score, the efficiency comprehensive score and the task scene data.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) cluster control, and in particular to a method for evaluating the performance of an intelligent UAV swarm. Background Art

[0002] Drone swarm technology is becoming a hot topic in research and application. By collaborating with multiple drones, drone swarms can accomplish complex missions such as reconnaissance, surveillance, and strike. However, with the continuous development of drone swarm technology, the need for its performance evaluation is becoming increasingly urgent. Traditional drone performance evaluation methods primarily focus on the performance metrics of individual drones, such as flight speed and endurance. These methods have significant limitations when evaluating the overall performance of a drone swarm. The performance of a drone swarm depends not only on the performance of individual drones but also on multiple factors, including the collaborative capabilities between drones, the swarm's organizational structure, and the mission environment. Therefore, a performance evaluation method that comprehensively considers these factors is needed.

[0003] Although the existing drone swarm performance evaluation methods can reflect the performance of drone swarms to a certain extent, they cannot fully consider all aspects of the drone swarm's parameters, resulting in one-sided and inaccurate evaluation results. The existing evaluation methods are also unable to flexibly adjust the weights of various parameters according to different mission scenarios and formation sizes, resulting in poor environmental adaptability of the evaluation results and reduced accuracy and reliability of the evaluation results. Summary of the Invention

[0004] In view of the above analysis, the embodiments of the present invention aim to provide a method for evaluating the performance of an intelligent UAV swarm, so as to solve the problem that the existing UAV swarm performance evaluation method is one-sided, has poor environmental adaptability, and leads to insufficient accuracy.

[0005] An embodiment of the present invention provides a method for evaluating the performance of an intelligent drone swarm, comprising the following steps:

[0006] Collect raw data of supporting parameters, emergent parameters and performance parameters;

[0007] A dynamic grey hierarchical analysis method is used to comprehensively evaluate the support parameters and generate a comprehensive support score;

[0008] Calculate the emergence composite score based on the emergence parameters;

[0009] Input the efficacy parameters into the pre-trained efficacy prediction model to predict the comprehensive efficacy score;

[0010] The performance evaluation results are obtained based on the support comprehensive score, emergence comprehensive score, effectiveness comprehensive score and task scenario data.

[0011] As a further improvement to this application, generating performance evaluation results includes:

[0012] Extract formation size data and mission target data based on mission scenario data;

[0013] The weights of the supportive and emergent comprehensive scores are dynamically allocated based on the formation size data and mission objective data; the weight of the effectiveness comprehensive score is a fixed value;

[0014] The support comprehensive score, emergence comprehensive score and effectiveness comprehensive score are weighted and integrated to obtain the performance evaluation results.

[0015] As a further improvement to this application, the weights of the supportive comprehensive score and the emergent comprehensive score are dynamically allocated based on the formation size data and mission objective data, including:

[0016] Determine the initial weights of the supportive comprehensive score and the emergent comprehensive score based on the formation size data;

[0017] The initial weights of the supportive comprehensive score and the emergent comprehensive score are revised based on the task target data; among which, the sum of the revised supportive comprehensive score, emergent comprehensive score, and effectiveness comprehensive score is 1.

[0018] As a further improvement to this application, the initial weights of the supportive comprehensive score and the emergent comprehensive score are determined based on the formation size data, including:

[0019] The weight of the supportive comprehensive score is set to the first initial weight value, and the weight of the emergent comprehensive score is set to the second initial weight value;

[0020] If the formation size data is less than the preset threshold, the weight of the supportive comprehensive score is maintained at the first initial weight value, and the weight of the emergent comprehensive score is maintained at the second initial weight value;

[0021] If the formation size data is greater than or equal to a preset threshold, the first initial weight value and the second initial weight value are adjusted according to the difference between the preset threshold and the formation size data; so that the first initial weight value increases as the formation size data increases, and the second initial weight value decreases as the formation size data increases, and the sum of the adjusted first initial weight and second initial weight remains unchanged.

[0022] As a further improvement to this application, the initial weights of the supportive and emergent comprehensive scores are modified based on the task objective data, including:

[0023] If the mission target is a reconnaissance target, the adjusted first initial weight value is increased and the adjusted second initial weight value is decreased;

[0024] If the mission target is a strike target, the adjusted second initial weight value is increased and the adjusted first initial weight value is decreased.

[0025] As a further improvement of the present application, a dynamic grey hierarchical analysis method is used to comprehensively evaluate the support parameters to generate a comprehensive support score including:

[0026] Construct a hierarchical evaluation system for supporting parameters, defining primary and secondary parameters; the primary parameters include platform performance parameters, deployment performance parameters, communication performance parameters, navigation performance parameters, perception performance parameters, and damage performance parameters; each primary parameter is associated with several corresponding secondary parameters;

[0027] Calculating a weighted score for each first-level parameter based on the weighted scores of all second-level parameters corresponding to the first-level parameter;

[0028] The weighted scores of all first-level parameters are fused to obtain the supportive comprehensive score.

[0029] As a further improvement of the present application, the emergent parameters include the number of node failures, formation reconstruction time, and communication delay time;

[0030] The emergence composite score is calculated based on the emergence parameters including:

[0031] Generate a preliminary emergence score based on the number of node failures;

[0032] The initial emergence score is adjusted based on the formation reconstruction time and communication delay time to obtain the comprehensive emergence score.

[0033] As a further improvement of the present application, the performance parameters include collaborative performance parameters, environmental performance parameters, and task performance parameters; the performance prediction model is shown in the following formula:

[0034]

[0035] Among them, K σ is the dynamic kernel function, b is the bias term, α, α * is a pair of Lagrange multipliers, x is a support vector, x i is the performance parameter of the i-th training sample, and N is the total number of training samples.

[0036] As a further improvement of the present application, the dynamic kernel function is shown in the following formula:

[0037]

[0038] Among them, x c is the collaborative support subvector, x e is the environment support vector, x tis the environment support vector, is the synergistic efficiency parameter vector of the i-th sample, is the environmental performance parameter vector of the i-th sample, is the task effectiveness parameter vector of the i-th sample, σ c is the kernel function parameter corresponding to the synergistic effectiveness parameter vector, σ e is the kernel function parameter corresponding to the environmental performance parameter vector, σ t is the kernel function parameter corresponding to the synergistic effectiveness parameter vector.

[0039] As a further improvement of the present application, the objective function of the pre-trained performance prediction model is:

[0040]

[0041] Among them, N is the total number of training samples, y n is the historical performance label corresponding to the nth training sample, f' n (x) is the predicted output of the support vector regression model.

[0042] The embodiments of the present invention have at least the following beneficial effects:

[0043] 1. By collecting raw data on supporting parameters, emergent parameters, and effectiveness parameters, this method not only covers the basic performance of drone swarms, but also considers the collaborative capabilities between drones and their actual performance in specific mission scenarios. This more comprehensive evaluation of drone swarm performance can effectively solve the problem of existing technologies' single evaluation methods that cannot fully reflect drone swarm performance, and improve the accuracy of performance evaluation results.

[0044] 2. By using the dynamic grey hierarchical analysis method to conduct a comprehensive evaluation of supporting parameters, the hierarchical relationship and weight distribution between parameters can be handled more scientifically, further improving the accuracy of the evaluation results. By dynamically allocating weights and generating performance evaluation results, the weights of each parameter can be flexibly adjusted according to different mission scenarios and formation sizes, thereby enhancing the environmental adaptability and accuracy of the evaluation results.

[0045] In the present invention, the above-mentioned technical solutions can be combined with each other to achieve more preferred combinations. Other features and advantages of the present invention will be described in the following description, and some advantages will become apparent from the description or be learned through practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings are only used for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. Throughout the drawings, the same reference symbols denote the same components.

[0047] Figure 1 A schematic flow chart of a method for evaluating the performance of an intelligent drone swarm provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0048] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, and are not used to limit the scope of the present invention.

[0049] A specific embodiment of the present invention discloses a method for evaluating the performance of an intelligent drone swarm. Figure 1 A method for evaluating the performance of an intelligent drone swarm includes the following steps:

[0050] Step 101: Collect the original data of supporting parameters, emergent parameters and performance parameters.

[0051] Specifically, supporting parameters refer to the basic performance parameters of each drone in a swarm. These parameters include primary and secondary parameters. Primary parameters include platform performance parameters, deployment performance parameters, communication performance parameters, navigation performance parameters, perception performance parameters, and damage performance parameters. Each primary parameter is associated with several corresponding secondary parameters. Supporting parameters can be obtained from technical manuals, product specifications, or measurements using pre-set sensors.

[0052] Secondary parameters for platform performance parameters include flight speed, endurance, payload capacity, engine performance, wing design, structural strength, and takeoff weight. Secondary parameters for deployment performance parameters include deployment speed, deployment accuracy, deployment flexibility, deployment range, deployment stability, and deployment adaptability. Secondary parameters for communication performance parameters include communication range, communication bandwidth, communication stability, communication protocol compatibility, anti-interference capability, and data transmission rate. Secondary parameters for navigation performance parameters include navigation accuracy, positioning speed, anti-interference capability, navigation algorithm efficiency, and multipath effect resistance. Secondary parameters for perception performance parameters include sensor accuracy, perception range, data processing speed, multi-sensor fusion capability, and target recognition capability. Secondary parameters for damage performance parameters include target hit probability, ammunition compatibility, target recognition accuracy, and combat radius.

[0053] Effectiveness parameters are metrics that measure the effectiveness and efficiency of a drone swarm in completing specific tasks. They include collaborative effectiveness parameters, environmental effectiveness parameters, and mission effectiveness parameters. Acquisition effectiveness parameters involve real-time collection of collaborative operation data, environmental disturbance data, and mission execution data through onboard sensors. Collaborative operation data refers to data generated when drones collaborate with each other, such as communication status and formation information. Environmental disturbance data refers to data affected by external environmental factors during flight, such as wind speed and electromagnetic interference. Mission execution data refers to data generated during the completion of a drone mission, such as coverage area and mission completion time.

[0054] Emergent parameters include the number of node failures, formation reconfiguration time, and communication delay. Node failure refers to the number of drones that fail to operate normally or lose contact during a swarm's mission due to various reasons (such as mechanical failure, electronic component damage, or communication interruption). This can be determined by monitoring the communication signals between the drones and the control station. Formation reconfiguration time refers to the time it takes for the swarm to reorganize and restore to its intended formation after its formation is disrupted by node failure, mission change, or other interference. This can be calculated by timing the time it takes for drones to reconfigure their formation during the swarm's mission. Communication delay refers to the time it takes for data and commands to be transmitted between drones or between a drone and the control station, from the time the data is sent by the transmitter to the time it is successfully received by the receiver. The communication delay can be calculated by adding a timestamp to the data packet at the transmitter and recording the current timestamp at the receiver upon receipt. Alternatively, a network analyzer can be used to capture data packets and automatically calculate the delay.

[0055] Step 102: A dynamic grey analytic hierarchy process is used to comprehensively evaluate the support parameters to generate a comprehensive support score.

[0056] The comprehensive supportability score is derived from a comprehensive evaluation of the supportability parameters of a drone swarm using the dynamic grey analytic hierarchy process (AHP). It reflects the swarm's performance in terms of platform, communication, and navigation capabilities. The dynamic grey analytic hierarchy process (AHP) is a comprehensive evaluation method that combines grey system theory and the analytic hierarchy process (AHP). Supportability parameters include multiple primary parameters and corresponding secondary parameters.

[0057] Specifically, the dynamic grey analytic hierarchy process is used to comprehensively evaluate the support parameters to generate a comprehensive support score including:

[0058] Step 1021: construct a hierarchical evaluation system for supporting parameters, define primary parameters and secondary parameters; and calculate the weighted score of each primary parameter based on the weighted scores of all secondary parameters corresponding to each primary parameter.

[0059] Specifically, the weighted score of each primary parameter is calculated based on the weighted scores of all secondary parameters corresponding to each primary parameter. The calculation of the weighted score requires first determining the weight of each secondary parameter, which can be determined based on expert experience or historical data. Each secondary parameter score is multiplied by its corresponding weight, and then these weighted scores are added together to obtain the weighted score for the primary parameter.

[0060] In step 1022, the weighted scores of all first-level parameters are integrated to obtain a supporting comprehensive score.

[0061] By calculating the weighted score of each secondary parameter, the corresponding weighted score of the primary parameter is obtained. The weighted scores of all primary parameters are then integrated to generate a supporting comprehensive score.

[0062] Step 103: Calculate an emergent comprehensive score based on the emergent parameters. The emergent parameters include the number of node failures, formation reconfiguration time, and communication delay time.

[0063] The emergence composite score is calculated based on the emergence parameters including:

[0064] Step 1031: Generate a preliminary emergence score based on the number of node failures.

[0065] Specifically, generating a preliminary emergence score based on the number of node failures includes:

[0066] The swarm working phase is divided into multiple time windows according to the swarm working task; the failure ratio of each time window is calculated based on the number of failed swarm nodes in each time window and the total number of drone swarm nodes; each drone in the swarm corresponds to a swarm node; based on the failure ratio of each time window, the average value of the failure ratio in the window and the fluctuation range of the failure ratio are calculated; the average value is compared with a preset failure number threshold to generate a first communication stability index, and the fluctuation range is compared with the fluctuation threshold to generate a second communication stability index; a preliminary emergence score is generated based on the first communication stability index and the second communication stability index.

[0067] First, the swarm's work phase is divided into multiple time windows based on the swarm's tasks. Within each time window, the failure ratio for each time window is calculated by calculating the ratio of the number of failed swarm nodes to the total number of drone swarm nodes. The average failure ratio for each time window reflects the degree of node failure. The fluctuation range of the failure ratio reflects the changing degree of failure. This fluctuation range can be obtained by calculating the difference between the maximum and minimum failure ratios across all time windows.

[0068] The average value is compared with a preset failure count threshold to generate a first communication stability indicator, and the fluctuation amplitude is compared with the fluctuation threshold to generate a second communication stability indicator. The failure count threshold and the fluctuation threshold can be set through experiments or historical data. The evaluation rules for the first communication stability indicator are set based on the failure count threshold, and the evaluation rules for the second communication stability indicator are set based on the fluctuation threshold.

[0069] Compare the average failure rate with a preset failure threshold. If the average is lower than the preset failure threshold, the swarm has good overall communication stability. The first communication stability indicator is set to c1, with a range of 1.1-1.3. Otherwise, the first communication stability indicator is set to c2, with a range of 0.7-0.9. Compare the fluctuation amplitude with the fluctuation threshold. If the fluctuation amplitude is lower than the preset fluctuation threshold, the swarm's failure rate is relatively stable. The second communication stability indicator is set to d1, with a range of 1.1-1.3. Otherwise, the swarm's failure rate fluctuates significantly. The second communication stability indicator is set to d2, with a range of 0.7-0.9.

[0070] By comparing the average failure rate with the preset failure number threshold to generate a first communication stability index, and by comparing the fluctuation amplitude with the fluctuation threshold to generate a second communication stability index, preliminary emergent scores can be generated from the two dimensions of overall communication stability and failure change, making the preliminary emergent scores more accurate.

[0071] Generating a preliminary emergence score based on the first communication stability indicator and the second communication stability indicator includes: normalizing the first communication stability indicator and the second communication stability indicator respectively; performing harmonic averaging on the normalized first communication stability indicator and the second communication stability indicator; and exponentially mapping the harmonic average of the normalized first communication stability indicator and the second communication stability indicator to generate a preliminary emergence score.

[0072] Specifically, the values of the first communication stability indicator and the second communication stability indicator are mapped to a common scale, typically in the range of 0 to 1. The normalized first communication stability indicator and the second communication stability indicator are harmonically averaged, which can be obtained by multiplying the two normalized indicator values, multiplying by 2, and then dividing by the sum of the two indicator values.

[0073] Exponential mapping uses an exponential function to map the harmonic mean to a target range. For example, you can multiply the harmonic mean by 100 and then multiply by the square of the harmonic mean to obtain a preliminary emergent score. The harmonic mean can be used to avoid overly high or low individual indicators from significantly influencing the results. Exponential mapping allows the harmonic mean to be mapped to a more intuitive scoring range, such as 0 to 100.

[0074] Step 1032: Adjust the preliminary emergence score based on the formation reconstruction time and the communication delay time to obtain a comprehensive emergence score.

[0075] Specifically, the initial emergence score is adjusted based on the formation reconstruction time and communication delay time to obtain the emergence comprehensive score including:

[0076] Calculating the deviation between the formation reconstruction time and the ideal formation reconstruction time, and calculating the deviation between the communication delay time and the ideal communication delay time;

[0077] If the deviation between the formation reconstruction time and the ideal formation reconstruction time is lower than the preset reconstruction time threshold or the deviation between the communication delay time and the ideal communication delay time is lower than the preset delay time threshold, the preliminary emergence score is multiplied by the first scoring coefficient to obtain the emergence comprehensive score;

[0078] If the deviation between the formation reconstruction time and the ideal formation reconstruction time is higher than the preset reconstruction time threshold and the deviation between the communication delay time and the ideal communication delay time is higher than the preset delay time threshold, the preliminary emergence score is multiplied by the second scoring coefficient to obtain the comprehensive emergence score.

[0079] Formation reconfiguration time refers to the time it takes for a swarm to re-form a valid formation after a node failure, while communication delay refers to the delay in communication within the swarm. Ideal formation reconfiguration time refers to the time it takes for a swarm to re-form a valid formation under optimal conditions, while ideal communication delay refers to the delay in communication under optimal conditions. Both ideal formation reconfiguration time and ideal communication delay can be determined based on historical data statistics or expert experience.

[0080] The deviation of the formation reconstruction time from the ideal formation reconstruction time is calculated by calculating the ratio of the difference between the formation reconstruction time and the ideal formation reconstruction time to the ideal formation reconstruction time. The deviation of the communication delay time from the ideal communication delay time is calculated by calculating the ratio of the difference between the communication delay time and the ideal communication delay time to the ideal communication delay time. If the formation reconstruction time deviation is lower than a preset reconstruction time threshold, or the communication delay time deviation is lower than a preset delay time threshold, the preliminary emergence score is multiplied by a first scoring coefficient, where the first scoring coefficient is greater than or equal to 1. If the formation reconstruction time deviation is higher than the preset reconstruction time threshold, and the communication delay time deviation is higher than the preset delay time threshold, the preliminary emergence score is multiplied by a second scoring coefficient, where the second scoring coefficient is less than 1.

[0081] Correcting the initial emergence score based on the formation reconstruction time and communication delay time can more comprehensively reflect the emergence level of the swarm, so that the emergence composite score can reflect its adaptability to interference and failure in actual tasks, thereby enhancing the accuracy and credibility of the emergence composite score.

[0082] Step 104 : Input the performance parameters into the pre-trained performance prediction model to predict the comprehensive performance score.

[0083] Step 1041 , calculating collaborative performance parameters based on the collaborative operation data, calculating environmental effectiveness parameters based on the environmental disturbance data, and calculating task effectiveness parameters based on the task execution data.

[0084] Specifically, collaborative operation data includes the communication adjacency matrix, actual formation position data, and drone positioning data. The communication adjacency matrix describes the communication connection status between drones, the actual formation position data reflects the actual position of the drones in the formation, and the drone positioning data provides the precise location of the drones.

[0085] Environmental disturbance data includes three-dimensional wind speed vector data, obstacle distribution grid data, and electromagnetic spectrum data. The three-dimensional wind speed vector data describes the magnitude and direction of wind speed, the obstacle distribution grid data represents the spatial distribution of obstacles, and the electromagnetic spectrum data reflects the interference of the electromagnetic environment.

[0086] Mission execution data includes the coordinates of the drone's coverage area, actual path length, and mission completion time. The coordinates of the drone's coverage area determine the drone's coverage range, the actual path length reflects the actual length of the drone's flight path, and the mission completion time is the time it takes the drone to complete the mission.

[0087] Calculation of collaborative performance parameters includes:

[0088] The communication stability coefficient is calculated by the algebraic connectivity of the communication adjacency matrix.

[0089] The communication adjacency matrix is a two-dimensional matrix whose elements indicate whether a communication connection exists between drones. A value of 1 in the matrix indicates normal communication between the two drones, while a value of 0 indicates a communication interruption. The connectivity of each drone is calculated using the communication adjacency matrix. The average connectivity of the entire swarm is then calculated to serve as the communication stability coefficient.

[0090] The formation maintenance rate is calculated by calculating the Hausdorff distance between the actual formation position data and the ideal formation position. The actual formation position data records the actual position coordinates of the UAV in the formation and can be expressed in the form of two-dimensional or three-dimensional coordinates. The ideal formation position refers to the position where the UAV should be in the formation. The formation maintenance rate is calculated by calculating the Hausdorff distance between the actual formation position data and the ideal formation position. The smaller the Hausdorff distance, the higher the formation maintenance rate. Specifically, the actual formation position data and the ideal formation position can be represented as point sets respectively, and the Hausdorff distance between the two point sets can be calculated to quantify the formation maintenance.

[0091] The collaborative positioning accuracy is determined by calculating the root mean square error (RMS) of the relative positions between drones based on their positioning data. The real-time position of the drones can be determined using satellite positioning or other positioning methods. The RMS error is calculated by calculating the RMS value of the deviation between the actual relative positions of all drones and their ideal relative positions, thereby determining the collaborative positioning accuracy.

[0092] The communication stability coefficient, formation retention rate, and collaborative positioning accuracy are weighted and fused to obtain collaborative performance parameters. During the weighted fusion process, weights can be set based on mission requirements and experience.

[0093] Furthermore, the computing environment performance parameters include:

[0094] The modulus of the gradient vector of the three-dimensional wind speed vector data is calculated to determine the intensity of the wind speed disturbance. Three-dimensional wind speed vector data refers to the wind speed magnitude and direction information of the environment in which the drone is located during flight, usually represented as a three-dimensional vector. The modulus of the gradient vector of the three-dimensional wind speed vector data is calculated to determine the intensity of the wind speed disturbance. The wind speed disturbance intensity reflects the severity of the wind speed change. When calculating the wind speed disturbance intensity, the three-dimensional wind speed vector data can be first subjected to time series analysis to extract the time derivative of the wind speed change, that is, the rate at which the wind speed changes over time. The modulus of the gradient vector of the time derivative is then calculated to quantify the intensity of the wind speed disturbance.

[0095] Calculate the density of each grid cell based on obstacle distribution raster data. Obstacle distribution raster data refers to the spatial distribution of obstacles in the environment obtained by sensors. It is usually represented in a rasterized form, with each grid cell containing obstacle density information. The density of each grid cell is calculated based on the obstacle distribution raster data. For obstacle distribution raster data, different grid resolutions can be set according to task requirements. For example, a higher resolution grid is used in complex terrain areas to more accurately reflect the distribution of obstacles.

[0096] The electromagnetic interference intensity is calculated by integrating the energy of the interference frequency bands of electromagnetic spectrum data. Electromagnetic spectrum data refers to the electromagnetic signal strength and frequency distribution in the drone's environment, reflecting the intensity and frequency band of electromagnetic interference. The electromagnetic interference intensity is calculated by integrating the energy of the interference frequency bands of electromagnetic spectrum data. This intensity reflects the impact of the electromagnetic environment on the drone's communication and navigation systems. When processing electromagnetic spectrum data, electromagnetic interference in different frequency bands can be classified. For example, the interference frequency bands can be divided into low-frequency, medium-frequency, and high-frequency ranges, and the energy integrals of these three ranges can be calculated. The interference intensity of different frequency bands can then be assigned different weights based on mission requirements.

[0097] The environmental performance parameter (EPP) is obtained by taking the harmonic mean of wind speed disturbance intensity, grid cell density, and electromagnetic interference intensity. By taking the harmonic mean of wind speed disturbance intensity, obstacle density, and electromagnetic interference intensity, a comprehensive EEP parameter is obtained. This EEP parameter more comprehensively reflects the impact of the environment on the execution of drone swarm missions.

[0098] Furthermore, the calculation task performance parameters include:

[0099] The area of the drone's coverage area is determined based on the drone's coverage area coordinates. The intersection of the drone's coverage area and the target area is calculated, and the target coverage rate is calculated based on the ratio of the intersection area to the target area. The drone's coverage area coordinates refer to the geographic area covered by the drone during the mission, expressed in coordinate form. When calculating the target coverage rate, the drone's coverage area coordinates can be rasterized, dividing the target area into multiple small grid cells. The number of grid cells covered by the drone is then counted to calculate the target coverage rate.

[0100] Calculate the ratio of the theoretical optimal path length to the actual path length to determine path selection reliability. Path selection reliability measures the degree to which the actual path is close to the theoretical optimal path. When calculating path selection reliability, a higher path selection reliability indicates that the actual path is closer to the theoretical optimal path and that the path selection is more reasonable.

[0101] Mission completion efficiency is calculated by calculating the ratio of the total planned time to the mission completion time. Mission completion time refers to the total time it takes for a drone to actually complete a mission, and is typically greater than the total planned time. When calculating mission completion efficiency, the total planned time can be dynamically adjusted based on the mission's urgency and priority. For high-priority tasks, the total planned time can be appropriately shortened. Mission completion time is typically greater than the total planned time. A higher mission completion efficiency value indicates a shorter mission completion time and higher mission execution efficiency.

[0102] The task effectiveness parameter is calculated based on the target coverage rate, path selection reliability and task completion efficiency. The task effectiveness parameter can be obtained based on the weighted sum of the target coverage rate, path selection reliability and task completion efficiency.

[0103] Step 1042 , constructing a support vector regression model, constructing a training sample set based on historical task data, and optimizing the kernel function parameters of the support vector regression model using an improved particle swarm optimization algorithm based on the training sample set to obtain a performance prediction model.

[0104] When building a support vector regression model, a training sample set is first constructed based on historical task data. Specifically, a three-dimensional input feature vector is constructed based on historical collaboration indicators, historical environmental indicators, and historical task indicators. A training sample set is then constructed based on the three-dimensional input feature vector and historical performance labels. Next, the search space of the improved particle swarm algorithm is defined, where the search space is the range of values for the kernel function parameters. The kernel function parameters of the support vector regression model are iteratively updated using the improved particle swarm algorithm to minimize the mean squared error (MSE) between the predicted output of the support vector regression model and the historical performance labels. Finally, a performance prediction model is generated based on the updated kernel function parameters. Specifically, at each iteration, the model's predicted output is calculated based on the current parameters and compared with the historical performance labels to calculate the mean squared error (MSE). When the MSE is minimized or the preset number of iterations is met, iterations are terminated, resulting in the optimized kernel function parameters. Historical task data is extracted from a historical database and includes historical collaboration indicators, historical environmental indicators, and historical task indicators corresponding to historical tasks, as well as pre-labeled historical performance labels. Historical coordination indicators, historical environmental indicators, and historical mission indicators are calculated based on historical mission data and are used to reflect the coordination, environment, and mission execution of drone swarms in historical missions. Pre-labeled historical performance labels are performance level labels pre-labeled based on the completion of historical missions and are used for model training.

[0105] Furthermore, a training sample set is constructed based on historical task data. Based on the training sample set, an improved particle swarm optimization algorithm is used to optimize the kernel function parameters of the support vector regression model to obtain an efficiency prediction model, including:

[0106] A three-dimensional input feature vector is constructed based on historical collaboration indicators, historical environment indicators, and historical task indicators, and a training sample set is constructed based on the three-dimensional input feature vector and historical performance labels; the three-dimensional input feature vector includes a historical collaboration sub-vector, a historical environment sub-vector, and a historical task sub-vector.

[0107] Historical mission data is data accumulated during previous drone swarm missions, including historical collaboration, environmental, and mission metrics, as well as corresponding pre-labeled historical performance tags. An improved particle swarm optimization algorithm is used to optimize the kernel function parameters in the support vector regression model to improve its predictive performance. Kernel function parameters determine the model's complexity and fit. The three-dimensional input feature vector consists of a historical collaboration subvector, a historical environmental subvector, and a historical mission subvector, corresponding to collaboration, environmental, and mission characteristics, respectively.

[0108] The mean vector of all historical collaborative subvectors in the training sample set is calculated as the collaborative support subvector, the mean vector of all historical environment subvectors in the training sample set is calculated as the environment support subvector, and the mean vector of all historical task subvectors in the training sample set is calculated as the task indicator support subvector.

[0109] The support vector of the kernel function is constructed based on the collaborative support subvector, the environment support subvector and the task support subvector.

[0110] Define the search space of the improved particle swarm optimization algorithm, which is the value range of the kernel function parameters.

[0111] Based on the training sample set, the kernel function parameters of the support vector regression model are iteratively updated by the improved particle swarm algorithm to minimize the mean square error between the predicted output of the support vector regression model and the historical performance label.

[0112] The kernel function parameters are iteratively updated using an improved particle swarm optimization algorithm. Each iteration calculates the predicted output of the support vector regression model based on the current kernel function parameters and calculates the mean squared error between the predicted output and the historical performance labels. When the mean squared error is minimized or the preset number of iterations is met, the iteration stops, resulting in the optimized kernel function parameters and the generation of a performance prediction model.

[0113] Generate an efficacy prediction model based on the updated kernel function parameters.

[0114] The efficacy prediction model is shown in the following formula:

[0115]

[0116] Among them, K σ is the dynamic kernel function, b is the bias term, α, α *is a pair of Lagrange multipliers, x is a support vector, x i is the three-dimensional input feature vector of the i-th sample, and N is the total number of training samples.

[0117] The dynamic kernel function is used to map input data into a high-dimensional space, thereby improving the model's fitting ability. It can also dynamically adjust kernel function parameters based on differences in input feature vectors to better adapt to different input conditions. Lagrange multiplier pairs are parameters introduced during the SVR optimization process to balance model complexity and fitting error.

[0118] The dynamic kernel function is shown in the following formula:

[0119]

[0120] Among them, x c is the collaborative support subvector, x e is the environment support vector, x t is the task support subvector, is the synergistic efficiency parameter vector of the i-th sample, is the environmental performance parameter vector of the i-th sample, is the task effectiveness parameter vector of the i-th sample, σ c is the kernel function parameter corresponding to the synergistic effectiveness parameter vector, σ e is the kernel function parameter corresponding to the environmental performance parameter vector, σ t is the kernel function parameter corresponding to the synergistic effectiveness parameter vector.

[0121] The parameters of the dynamic kernel function include the kernel parameter σc corresponding to the synergistic effectiveness parameter vector, the kernel parameter σe corresponding to the environmental effectiveness parameter vector, and the kernel parameter σt corresponding to the task effectiveness parameter vector. The kernel parameters σc, σe, and σt control the influence of the three subvectors in the kernel function. The dynamic kernel function dynamically adjusts its shape and width based on the specific characteristics of the input feature vector, thus better adapting to different input data.

[0122] The objective function of iteratively updating the kernel function parameters by the improved particle swarm algorithm is:

[0123]

[0124] Among them, N is the total number of training samples, y n is the historical performance label corresponding to the nth training sample, f' n (x) is the predicted output of the support vector regression model.

[0125] Step 1043 , input the performance parameters into the trained performance prediction model to obtain a comprehensive performance score.

[0126] Step 105 , obtaining a performance evaluation result based on the support comprehensive score, the emergence comprehensive score, the effectiveness comprehensive score, and the task scenario data.

[0127] Specifically, the performance evaluation results generated based on the support comprehensive score, emergence comprehensive score, effectiveness comprehensive score and task scenario data include:

[0128] Step 1051 extracts formation size data and mission objective data based on the mission scenario data. The mission scenario data includes formation size data and mission objective data. Formation size data refers to the number of drones in the drone swarm, and mission objective data refers to the specific objectives of the drone swarm's mission, such as reconnaissance, strike, or cruise.

[0129] Step 1052: Dynamically assign weights of the supportive comprehensive score and the emergent comprehensive score based on the formation size data and the mission objective data; the weight of the effectiveness comprehensive score is a fixed value.

[0130] Generally, since the comprehensive efficiency score directly reflects the actual effect of the drone swarm in mission execution and has high stability and accuracy, the weight of the comprehensive efficiency score is set to a fixed value, and the weight of the comprehensive efficiency score is set to [0.4-0.5];

[0131] Specifically, the weights of the supportive comprehensive score and the emergent comprehensive score are dynamically allocated based on the formation size data and mission objective data.

[0132] The initial weights of the supportive comprehensive score and the emergent comprehensive score are determined based on the formation size data; the initial weights of the supportive comprehensive score and the emergent comprehensive score are revised based on the mission objective data; among which, the sum of the revised supportive comprehensive score, emergent comprehensive score, and effectiveness comprehensive score is 1.

[0133] The initial weights of the supportive comprehensive score and the emergent comprehensive score are determined based on the formation size data. When the formation size is small, the weight of the supportive comprehensive score is higher, and the basic parameters of the swarm have a greater impact on the swarm performance. When the formation size is large, the weight of the emergent comprehensive score will increase, because the self-organization and adaptability of large-scale swarms have a more important impact on the swarm performance.

[0134] Specifically, the initial weights of the supportive comprehensive score and the emergent comprehensive score are determined based on the formation size data, including:

[0135] The weight of the supportive comprehensive score is set to a first initial weight value, and the weight of the emergent comprehensive score is set to a second initial weight value; the first initial weight value and the second initial weight value are the same;

[0136] If the formation size data is less than the preset threshold, the weight of the supportive comprehensive score is maintained at the first initial weight value, and the weight of the emergent comprehensive score is maintained at the second initial weight value;

[0137] If the formation size data is greater than or equal to a preset threshold, the first initial weight value and the second initial weight value are adjusted according to the difference between the preset threshold and the formation size data; so that the first initial weight value increases as the formation size data increases, and the second initial weight value decreases as the formation size data increases, and the sum of the adjusted first initial weight and second initial weight remains unchanged.

[0138] Specifically, the initial weight value can be set based on a combination of historical data, and then the initial weight value can be dynamically adjusted based on the comparison of the formation size data with the preset threshold. The weight of the supporting comprehensive score is set to the first initial weight value, and the weight of the emergent comprehensive score is set to the second initial weight value. First, the first initial weight value and the second initial weight value are set to be the same. The preset threshold is a critical value determined based on experience or historical data. When the formation size data is less than the preset threshold, the initial weight value remains unchanged; when the formation size data is greater than or equal to the preset threshold, the initial weight value is adjusted based on the difference between the formation size data and the preset threshold, so that the weight of the supporting comprehensive score increases with the increase of the formation size, and the weight of the emergent comprehensive score decreases with the increase of the formation size, while ensuring that the total weight after adjustment remains unchanged. The preset threshold can be set to 10 aircraft. If the formation size data is less than 10 aircraft, the initial weight is maintained, and the first initial weight value and the second initial weight value are equal, which is [0.25-0.3]. If the fleet size is greater than or equal to 10 aircraft, the weights are adjusted based on the difference between the fleet size and the preset threshold. For example, for each additional drone, the support weight increases by 0.02 and the emergence weight decreases by 0.02. For a fleet size of 15 aircraft, the first initial weight is 0.6 and the second initial weight is 0.4.

[0139] The initial weights of the supportive and emergent comprehensive scores are modified based on the task objective data, including:

[0140] If the mission target is a reconnaissance target, the adjusted first initial weight value is increased and the adjusted second initial weight value is decreased;

[0141] If the mission target is a strike target, the adjusted second initial weight value is increased and the adjusted first initial weight value is decreased.

[0142] The initial weights of the support and emergence composite scores are adjusted based on mission objective data. During implementation, reconnaissance missions rely more heavily on the drone's fundamental performance, so the support composite score's weight increases. Meanwhile, strike missions, which place greater demands on drone formation capabilities, require the emergence composite score to be adaptively increased. When the mission objective is reconnaissance, the support composite score's weight increases, while the emergence composite score's weight decreases. When the mission objective is strike, the emergence composite score's weight increases, while the support composite score's weight decreases. The total of these adjusted weights remains 1.

[0143] The initial weight is corrected, and the correction amount is shown in the following formula;

[0144] Δ = Δ0 × K; where Δ is the correction value, Δ0 is the baseline correction value, and K is the mission complexity coefficient. The mission complexity coefficient is determined by the complexity of the mission objective and ranges from 1 to 3. For example, for a simple reconnaissance mission, K = 1, and for a complex reconnaissance mission, K = 2; for a simple strike mission, K = 1.5, and for a complex strike mission, K = 3. The baseline correction value is the baseline correction value for the initial weight and is set based on experience.

[0145] When the mission objective is a reconnaissance target, the K value is determined based on the complexity of the mission objective. After the K value is determined, the correction value Δ is calculated based on the K value. Δ is subtracted from the second initial weight value and Δ is added to the first initial weight value to complete the weight correction. When the mission objective is an attack target, the K value is determined based on the complexity of the mission objective. After the K value is determined, the correction value Δ is calculated based on the K value. Δ is subtracted from the first initial weight value and Δ is added to the second initial weight value to complete the weight correction.

[0146] Step 1053 , weightedly integrate the support comprehensive score, the emergence comprehensive score, and the effectiveness comprehensive score to obtain a performance evaluation result.

[0147] Weighted fusion involves multiplying the supportive, emergent, and effectiveness scores by their respective weights, then summing these weighted scores to arrive at a performance evaluation result. This performance evaluation result is then compared with pre-set performance evaluation rules to determine the performance level of the swarm and whether it meets the expected performance standards.

[0148] The above-mentioned embodiments of the present invention include at least the following technical effects: by collecting the original data of supporting parameters, emergent parameters and efficiency parameters, the method not only covers the basic performance of the UAV swarm, but also takes into account the coordination ability between UAVs and the actual performance in specific mission scenarios; a more comprehensive evaluation of the performance of the UAV swarm can effectively solve the problem that the evaluation method in the existing technology is single and cannot fully reflect the performance of the UAV swarm, thereby improving the accuracy of the performance evaluation results; by adopting the dynamic gray hierarchical analysis method to comprehensively evaluate the supporting parameters, the hierarchical relationship and weight distribution between parameters can be handled more scientifically, further improving the accuracy of the evaluation results, and by dynamically allocating weights and generating performance evaluation results, the weights of each parameter can be flexibly adjusted according to different mission scenarios and formation sizes, so that the environmental adaptability of the evaluation results is enhanced and the accuracy is improved.

[0149] Those skilled in the art will appreciate that all or part of the process steps of the above-described embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, such as a magnetic disk, an optical disk, a read-only memory, or a random access memory.

[0150] 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 changes or substitutions that can be easily thought of by any technician familiar with this technical field within the technical scope disclosed by the present invention should be covered by the scope of protection of the present invention.

Claims

1. A method for evaluating the performance of an intelligent drone swarm, characterized in that: The following steps are involved: Collect raw data of supporting parameters, emergent parameters and performance parameters; A dynamic grey hierarchical analysis method is used to comprehensively evaluate the support parameters and generate a comprehensive support score; Calculate the emergence composite score based on the emergence parameters; Input the efficacy parameters into the pre-trained efficacy prediction model to predict the comprehensive efficacy score; The performance evaluation results are obtained based on the support comprehensive score, emergence comprehensive score, effectiveness comprehensive score and task scenario data.

2. The method according to claim 1, characterized in that The performance evaluation results based on the support comprehensive score, emergence comprehensive score, effectiveness comprehensive score and task scenario data include: Extract formation size data and mission target data based on mission scenario data; The weights of the supportive and emergent comprehensive scores are dynamically allocated based on the formation size data and mission objective data; the weight of the effectiveness comprehensive score is a fixed value; The support comprehensive score, emergence comprehensive score and effectiveness comprehensive score are weighted and integrated to obtain the performance evaluation results.

3. The method according to claim 2, characterized in that The weights of the supportive and emergent comprehensive scores are dynamically allocated based on the formation size data and mission objective data. Determine the initial weights of the supportive comprehensive score and the emergent comprehensive score based on the formation size data; The initial weights of the supportive comprehensive score and the emergent comprehensive score are revised based on the task target data; among which, the sum of the revised supportive comprehensive score, emergent comprehensive score, and effectiveness comprehensive score is 1.

4. The method according to claim 3, characterized in that The initial weights for determining the supportive and emergent comprehensive scores based on fleet size data include: The weight of the supportive comprehensive score is set to the first initial weight value, and the weight of the emergent comprehensive score is set to the second initial weight value; If the formation size data is less than the preset threshold, the weight of the supportive comprehensive score is maintained at the first initial weight value, and the weight of the emergent comprehensive score is maintained at the second initial weight value; If the formation size data is greater than or equal to a preset threshold, the first initial weight value and the second initial weight value are adjusted according to the difference between the preset threshold and the formation size data; so that the first initial weight value increases as the formation size data increases, and the second initial weight value decreases as the formation size data increases, and the sum of the adjusted first initial weight and second initial weight remains unchanged.

5. The method according to claim 4, characterized in that The initial weights of the supportive and emergent comprehensive scores are modified based on the task objective data, including: If the mission target is a reconnaissance target, the adjusted first initial weight value is increased and the adjusted second initial weight value is decreased; If the mission target is a strike target, the adjusted second initial weight value is increased and the adjusted first initial weight value is decreased.

6. The method according to claim 1, characterized in that The dynamic grey analytic hierarchy process is used to comprehensively evaluate the support parameters and generate a comprehensive support score including: Construct a hierarchical evaluation system for supporting parameters, defining primary and secondary parameters; the primary parameters include platform performance parameters, deployment performance parameters, communication performance parameters, navigation performance parameters, perception performance parameters, and damage performance parameters; each primary parameter is associated with several corresponding secondary parameters; Calculating a weighted score for each first-level parameter based on the weighted scores of all second-level parameters corresponding to the first-level parameter; The weighted scores of all first-level parameters are fused to obtain the supportive comprehensive score.

7. The method according to claim 1, characterized in that Emergent parameters include the number of node failures, formation reconstruction time, and communication delay time; The emergence composite score is calculated based on the emergence parameters including: Generate a preliminary emergence score based on the number of node failures; The initial emergence score is adjusted based on the formation reconstruction time and communication delay time to obtain the comprehensive emergence score.

8. The method according to claim 1, characterized in that The performance parameters include collaborative performance parameters, environmental performance parameters, and task performance parameters. The performance prediction model is shown in the following formula: Among them, K σ is the dynamic kernel function, b is the bias term, α, α * is a pair of Lagrange multipliers, x is a support vector, x i is the performance parameter of the i-th training sample, and N is the total number of training samples.

9. The method according to claim 8, characterized in that The dynamic kernel function is shown in the following formula: Among them, x c is the collaborative support subvector, x e is the environment support vector, x t is the task support subvector, is the synergistic efficiency parameter vector of the i-th sample, is the environmental performance parameter vector of the i-th sample, is the task effectiveness parameter vector of the i-th sample, σ c is the kernel function parameter corresponding to the synergistic effectiveness parameter vector, σ e is the kernel function parameter corresponding to the environmental performance parameter vector, σ t is the kernel function parameter corresponding to the synergistic effectiveness parameter vector.

10. The method according to claim 9, characterized in that The objective function of the pre-trained performance prediction model is: Among them, N is the total number of training samples, y n is the historical performance label corresponding to the nth training sample, f' n (x) is the predicted output of the support vector regression model.

Images