A method and device for evaluating performance of unmanned clusters by combining numerical simulation and reality

Abstract

The application discloses a kind of unmanned cluster performance numerical-real combination simulation evaluation method and device, the method includes: the performance actual test result information set of unmanned cluster is collected, and the performance simulation test result information set is obtained;The performance actual test result information set includes the actual test result value sequence of each performance index of unmanned cluster;The performance simulation test result information set includes the simulation test result value sequence of each performance index of unmanned cluster;The performance actual test result information set and performance simulation test result information set are preprocessed, and the preprocessed information set is obtained;The preprocessed information set is carried out performance comprehensive evaluation processing, and the comprehensive performance evaluation value of unmanned cluster is obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned swarm and the technical field of simulation evaluation, and particularly relates to a numerical-real combined simulation evaluation method and device for performance of an unmanned swarm. BACKGROUND

[0002] With the rapid development of unmanned technology, unmanned swarms are increasingly widely used in fields such as logistics, exploration and resource management. Performance evaluation of unmanned swarms, as a key link to ensure their efficient and reliable operation, has attracted much attention from researchers and industry practitioners. Currently, performance evaluation of unmanned swarms mainly adopts separate actual testing or simulation testing methods. Although actual testing alone can truly reflect the performance of unmanned swarms in actual scenarios, it has problems such as high testing cost, long testing cycle, great influence of environmental factors and difficulty in reproducing testing conditions. While simulation testing alone has advantages such as low cost, strong repeatability and flexible setting of testing scenarios, the accuracy and reliability of simulation testing results are difficult to guarantee due to the differences between simulation models and actual scenarios. In addition, existing evaluation methods often ignore the correlation between actual testing results and simulation testing results, and fail to fully utilize the complementarity of the two, making the evaluation results unable to comprehensively and accurately reflect the true performance of unmanned swarms. SUMMARY

[0003] The present application mainly solves the problem of how to effectively fuse various types of testing data of unmanned swarms and accurately estimate the performance of unmanned swarms. The present application discloses a numerical-real combined simulation evaluation method and device for performance of an unmanned swarm.

[0004] In a first aspect, the present application discloses a numerical-real combined simulation evaluation method for performance of an unmanned swarm, comprising:

[0005] S1, collecting performance actual testing result information set and performance simulation testing result information set of the unmanned swarm; the performance actual testing result information set comprises actual testing result value sequence of each performance index of the unmanned swarm; the performance simulation testing result information set comprises simulation testing result value sequence of each performance index of the unmanned swarm;

[0006] S2, pre-processing the performance actual testing result information set and the performance simulation testing result information set to obtain a pre-processing information set;

[0007] S3, performing performance comprehensive evaluation processing on the pre-processing information set to obtain a comprehensive performance evaluation value of the unmanned swarm.

[0008] The pre-processing of the performance actual testing result information set and the performance simulation testing result information set to obtain the pre-processing information set comprises:

[0009] S21, performing data cleaning processing on the performance actual test result information set and the performance simulation test result information set to obtain a first information set;

[0010] S22, performing category detection processing on the first information set to obtain a second information set;

[0011] S23, performing numerical-real joint discrimination processing on the second information set to obtain a pretreatment information set.

[0012] The numerical-real joint discrimination processing on the second information set to obtain a pretreatment information set comprises:

[0013] S231, calculating a corresponding average value for each performance index in the second information set respectively based on a simulation test result value sequence and an actual test result value sequence;

[0014] S232, subtracting the corresponding average value from the simulation test result value sequence and the actual test result value sequence of each performance index in the second information set respectively to obtain a corresponding difference sequence;

[0015] S233, performing deviation upper limit estimation processing on the difference sequence of each performance index to obtain a corresponding deviation upper limit value; and constructing a deviation range information of each performance index based on the deviation upper limit value;

[0016] S234, discriminating whether the elements of the simulation test result value sequence and the actual test result value sequence of each performance index in the second information set are within the corresponding deviation range information, and deleting the elements not within the deviation range information from the second information set to obtain a pretreatment information set.

[0017] The deviation upper limit estimation processing on the difference sequence of each performance index to obtain a corresponding deviation upper limit value comprises:

[0018] An autoregressive-moving average model is established for the independent variable and the dependent variable of the difference sequence corresponding to the simulation test result value sequence of each performance index with the collection time of the elements of the simulation test result value sequence as the independent variable and the elements of the difference sequence as the dependent variable to obtain a regression model of the simulation test result value sequence;

[0019] An autoregressive-moving average model is established for the independent variable and the dependent variable of the difference sequence corresponding to the actual test result value sequence of each performance index with the collection time of the elements of the actual test result value sequence as the independent variable and the elements of the difference sequence as the dependent variable to obtain a regression model of the actual test result value sequence;

[0020] The regression model of the simulation test result value sequence and the regression model of the actual test result value sequence are fused to obtain a fusion regression model;

[0021] The collection time of all elements of the simulation test result value sequence is calculated by using the fusion regression model to obtain a deviation value sequence;

[0022] The maximum value in the deviation value sequence is determined as the deviation upper limit value of the performance index.

[0023] The performance comprehensive evaluation processing on the preprocessed information set is performed to obtain a comprehensive performance evaluation value of the unmanned cluster, and the performance comprehensive evaluation processing on the preprocessed information set comprises:

[0024] S31, an index standard value set is obtained; the index standard value set comprises a standard value of each performance index;

[0025] S32, actual test performance evaluation processing is performed on the performance actual test result information set in the preprocessed information set and the index standard value set to obtain an actual test performance evaluation value and a statistical fluctuation value;

[0026] S33, simulation performance evaluation processing is performed on the performance simulation test result information set in the preprocessed information set and the index standard value set based on the statistical fluctuation value to obtain a simulation performance evaluation value;

[0027] S34, weighted summation processing is performed on the actual test performance evaluation value and the simulation performance evaluation value to obtain a comprehensive performance evaluation value of the unmanned cluster.

[0028] The actual test performance evaluation processing on the performance actual test result information set in the preprocessed information set and the index standard value set comprises:

[0029] S321, the actual test result value sequence of each performance index in the performance actual test result information set in the preprocessed information set is subtracted from the standard value of the corresponding performance index in the index standard value set to obtain a corresponding difference value sequence;

[0030] S322, a difference matrix is constructed by using the difference value sequence of each performance index as a row vector;

[0031] S323, performance evaluation processing is performed on the difference matrix to obtain an actual test performance evaluation value;

[0032] S324, statistical fluctuation estimation processing is performed on the difference matrix to obtain a statistical fluctuation value.

[0033] The performance evaluation processing on the difference matrix comprises:

[0034] Eigenvalue solving processing is performed on the difference matrix to obtain an average eigenvalue λ0 and a maximum eigenvalue λ max ;

[0035] Each row vector of the difference matrix is respectively subjected to variational modal decomposition to obtain a transformation matrix Y; a row vector of the transformation matrix Y is a variational modal decomposition result of each row vector of the difference matrix;

[0036] Fusion feature calculation is performed on the transformation matrix and the difference matrix to obtain a fusion matrix A;

[0037] Logarithmic calculation processing is performed on diagonal elements of the fusion matrix A to obtain a measured performance evaluation value.

[0038] In a second aspect of the embodiment of the present application, a device for simulating and evaluating the performance of an unmanned cluster in a combination of numerical and actual data is disclosed, and the device comprises:

[0039] A memory in which executable program codes are stored;

[0040] A processor coupled with the memory;

[0041] The processor calls the executable program codes stored in the memory to execute the method for simulating and evaluating the performance of an unmanned cluster in a combination of numerical and actual data.

[0042] In a third aspect of the embodiment of the present application, a computer storage medium is disclosed, and the computer storage medium stores computer instructions, which are called by a computer to execute the method for simulating and evaluating the performance of an unmanned cluster in a combination of numerical and actual data.

[0043] In a fourth aspect of the embodiment of the present application, an information data processing terminal is disclosed, and the information data processing terminal is used to implement the method for simulating and evaluating the performance of an unmanned cluster in a combination of numerical and actual data.

[0044] The present application has the following beneficial effects:

[0045] The unmanned cluster performance numerical-real combined simulation evaluation method provided by the embodiment has multiple significant advantages. In the data processing link, through data cleaning, category detection and numerical-real combined discrimination and other preprocessing operations on the performance actual test result information set and the performance simulation test result information set, especially based on autoregressive-moving average modeling and model fusion calculation to determine the deviation upper limit, abnormal data is effectively eliminated, the data quality and reliability are improved, and a foundation is laid for subsequent accurate evaluation. In the performance evaluation process, the actual test result and the simulation test result are respectively evaluated with the index standard value set, and the statistical fluctuation value is considered, and the difference matrix is processed by using the variational mode decomposition and eigenvalue solving methods, so that the actual performance and the simulation performance are accurately evaluated. Finally, the actual performance evaluation value and the simulation performance evaluation value are weighted and summed, the advantages of the actual test and the simulation test are fully given, and the comprehensive performance evaluation value can more comprehensively and accurately reflect the real performance of the unmanned cluster. The method not only reduces the limitations brought by a single test method, but also improves the evaluation efficiency and reduces the evaluation cost, and provides a scientific and reliable basis for the optimal design and efficient operation of the unmanned cluster. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 The embodiment flowchart of the method of the present application. DETAILED DESCRIPTION

[0047] In order to better understand the content of the present application, an embodiment is given here.

[0048] Figure 1 The embodiment flowchart of the method of the present application.

[0049] In the first aspect of the embodiment of the present application, a kind of unmanned cluster performance numerical-real combined simulation evaluation method is disclosed, comprising:

[0050] S1, the performance actual test result information set and the performance simulation test result information set of unmanned cluster are acquired;The performance actual test result information set includes the actual test result value sequence of each performance index of unmanned cluster;The performance simulation test result information set includes the simulation test result value sequence of each performance index of unmanned cluster;

[0051] S2, the performance actual test result information set and the performance simulation test result information set are preprocessed, and the preprocessed information set is obtained;

[0052] S3, the performance comprehensive evaluation processing of the preprocessed information set is carried out, and the comprehensive performance evaluation value of unmanned cluster is obtained.

[0053] The pre-processing of the performance actual test result information set and the performance simulation test result information set includes:

[0054] S21, data cleaning processing is performed on the performance actual test result information set and the performance simulation test result information set to obtain a first information set;

[0055] S22, category detection processing is performed on the first information set to obtain a second information set;

[0056] S23, numerical-real joint discrimination processing is performed on the second information set to obtain a pre-processed information set.

[0057] The numerical-real joint discrimination processing on the second information set to obtain a pre-processed information set includes:

[0058] S231, the average value corresponding to each performance index in the second information set is calculated by subtracting the simulation test result value sequence and the actual test result value sequence of each performance index in the second information set from the corresponding average value;

[0059] S232, the difference sequence corresponding to each performance index is obtained by subtracting the simulation test result value sequence and the actual test result value sequence of each performance index in the second information set from the corresponding average value;

[0060] S233, the deviation upper limit value of each performance index is obtained by performing deviation upper limit estimation processing on the difference sequence of each performance index; and the deviation range information of each performance index is constructed based on the deviation upper limit value.

[0061] S234, the elements of the simulation test result value sequence and the actual test result value sequence of each performance index in the second information set are discriminated to determine whether the values are within the corresponding deviation range information; the elements not within the deviation range information are deleted from the second information set to obtain a pre-processed information set.

[0062] The deviation range information of each performance index is constructed based on the deviation upper limit value.

[0063] The deviation range information of each performance index is constructed by taking the mean value of the actual test result value sequence of each performance index as the center value and taking the deviation upper limit value as the fluctuation range; and the expression of the deviation range information is [a-b, a, a+b], where a is the mean value of the actual test result value sequence and b is the deviation upper limit value.

[0064] The deviation upper limit value of each performance index is obtained by performing deviation upper limit estimation processing on the difference sequence of each performance index.

[0065] corresponding to the simulation test result value sequence of each performance index, taking the collection time of the elements of the simulation test result value sequence corresponding to the difference sequence as the independent variable, taking the elements of the difference sequence as the dependent variable, performing autoregressive-moving average modeling on the independent variable and the dependent variable to obtain a regression model of the simulation test result value sequence;

[0066] corresponding to the actual test result value sequence of each performance index, taking the collection time of the elements of the actual test result value sequence corresponding to the difference sequence as the independent variable, taking the elements of the difference sequence as the dependent variable, performing autoregressive-moving average modeling on the independent variable and the dependent variable to obtain a regression model of the actual test result value sequence;

[0067] performing fusion calculation on the regression model of the simulation test result value sequence and the regression model of the actual test result value sequence to obtain a fusion regression model;

[0068] using the fusion regression model to calculate the collection time of all elements of the simulation test result value sequence to obtain a deviation value sequence;

[0069] determining the maximum value in the deviation value sequence as an upper deviation limit value.

[0070] The fusion calculation of the regression model of the simulation test result value sequence and the regression model of the actual test result value sequence to obtain a fusion regression model comprises:

[0071] extracting the coefficient vectors of the first regression model and the second regression model, calculating the cross-correlation matrix thereof, calculating the eigenvalues of the cross-correlation matrix, and constructing an eigenvalue vector using all the eigenvalues;

[0072] using the eigenvalue vector as a weight vector to multiply the elements of the mean vector of the coefficient vectors of the first regression model and the second regression model item by item to obtain a fusion coefficient vector;

[0073] using the fusion coefficient vector as the coefficient vector of the polynomial corresponding to the fusion regression model, using the polynomial order of the first regression model as the order of the polynomial corresponding to the fusion regression model to construct the fusion regression model;

[0074] The multiplication of the elements of the weight vector and the first element of the mean vector obtained by averaging the elements of the coefficient vectors of the first regression model and the second regression model item by item to obtain the first element of the fusion coefficient vector is performed, and the corresponding elements of the fusion coefficient vector are obtained by sequentially multiplying each element.

[0075] The order of the numerator polynomial and the denominator polynomial of the corresponding polynomial of the fusion regression model is the same as that of the first regression model.

[0076] The performance comprehensive evaluation processing on the preprocessed information set obtains a comprehensive performance evaluation value of the unmanned cluster, and comprises:

[0077] S31, obtaining an index standard value set; the index standard value set comprises a standard value of each performance index;

[0078] S32, performing actual measurement performance evaluation processing on the performance actual test result information set in the preprocessed information set and the index standard value set to obtain an actual measurement performance evaluation value and a statistical fluctuation value;

[0079] S33, performing simulation performance evaluation processing on the performance simulation test result information set in the preprocessed information set and the index standard value set based on the statistical fluctuation value to obtain a simulation performance evaluation value;

[0080] S34, performing weighted summation processing on the actual measurement performance evaluation value and the simulation performance evaluation value to obtain a comprehensive performance evaluation value of the unmanned cluster.

[0081] The actual measurement performance evaluation processing on the performance actual test result information set in the preprocessed information set and the index standard value set to obtain an actual measurement performance evaluation value and a statistical fluctuation value comprises:

[0082] S321, subtracting a standard value of a corresponding performance index in the index standard value set from an actual test result value sequence of each performance index in the performance actual test result information set in the preprocessed information set to obtain a corresponding difference value sequence;

[0083] S322, constructing a difference matrix by using the difference value sequence of each performance index as a row vector;

[0084] S323, performing performance evaluation processing on the difference matrix to obtain an actual measurement performance evaluation value;

[0085] S324, performing statistical fluctuation estimation processing on the difference matrix to obtain a statistical fluctuation value.

[0086] The performance evaluation processing on the difference matrix to obtain an actual measurement performance evaluation value comprises:

[0087] performing eigenvalue solving processing on the difference matrix to obtain an average eigenvalue λ0 and a maximum eigenvalue λ max ;

[0088] respectively, to obtain a transformation matrix Y; a row vector of the transformation matrix Y is a variational modal decomposition result of each row vector of the difference matrix;

[0089] performing fusion feature calculation on the transformation matrix and the difference matrix to obtain a fusion matrix A;

[0090] performing logarithmic calculation processing on diagonal elements of the fusion matrix A to obtain a measured performance evaluation value.

[0091] An expression of the logarithmic calculation processing is:

[0092]

[0093] wherein, A i is an i th diagonal element of the fusion matrix A, M is a number of diagonal elements, and sc is the measured performance evaluation value.

[0094] An expression of the fusion feature calculation is:

[0095] A=YΔ T R0,

[0096] wherein, a singular value decomposition expression of the difference matrix C is C=WΔV, W, Δ, and V represent a left matrix, a middle matrix, and a right matrix of singular value decomposition of the difference matrix C respectively, and R0 is a cross-correlation matrix of the transformation matrix, and is obtained by performing cross-correlation calculation on row vectors of the transformation matrix;

[0097] An i th row, j th column element of the cross-correlation matrix is a cross-correlation value of an i th row vector and a j th row vector of the transformation matrix.

[0098] In the calculation of the measured performance evaluation value, average eigenvalue λ0 and maximum eigenvalue λmax are obtained by performing eigenvalue solving on the difference matrix, so that the feature structure of the difference matrix can be grasped as a whole, and the internal change trend of data can be reflected. The variational modal decomposition decomposes the row vectors of the difference matrix into different modes, deeply mines the multi-scale features of data, and provides more abundant information for subsequent analysis. Through the fusion feature calculation expression, the transformation matrix Y is combined with the singular value decomposition result of the difference matrix and the cross-correlation matrix R0 of the transformation matrix, so that the information of different dimensions of the matrices is fully utilized, and the fusion matrix A can comprehensively reflect the multiple characteristics of the difference between the actual test result and the standard value. The logarithmic calculation processing expression introduces the eigenvalue ratio and the sine function, enhances the sensitivity to the change of data features, and through the logarithmic operation, the diagonal elements of the fusion matrix A are converted, the difference between data is effectively amplified, and the measured performance evaluation value can more accurately and carefully reflect the difference degree between the actual performance of the unmanned cluster and the standard performance.

[0099] The statistical fluctuation estimation processing on the difference matrix obtains a statistical fluctuation value, comprising:

[0100] The mean, variance and median values of each row vector of the difference matrix are statistically obtained;

[0101] The mean, variance and median values of all row vectors are subjected to fusion statistical fluctuation processing to obtain a statistical fluctuation value bd;

[0102] The expression of the fusion statistical fluctuation processing is:

[0103]

[0104] Wherein, α i , β i , λ i are the mean, variance and median values of the i-th row vector of the difference matrix, and M1 is the row dimension of the difference matrix.

[0105] In terms of statistical fluctuation estimation, the mean, variance and median values of each row vector of the difference matrix are first statistically obtained, which respectively describes the data from the three dimensions of central tendency, dispersion degree and data distribution position, and comprehensively captures the fluctuation characteristics of the data. The fusion statistical fluctuation processing expression combines trigonometric functions and exponential functions to nonlinearly fuse the mean, variance and median values of different row vectors, which can effectively amplify the differences in data fluctuations and more accurately quantify the statistical fluctuation of the performance of the unmanned cluster in different indicators, providing a reliable basis for subsequent simulation performance evaluation based on statistical fluctuation values, and further improving the stability and accuracy of the overall performance evaluation of the unmanned cluster.

[0106] The data cleaning processing includes filling missing values, smoothing noise data, smoothing or deleting outlier points; the smoothing of noise data is to first identify noise data, and then smooth the noise data according to the data before and after the noise data; the noise data is a value less than the detection sensitivity of the sensor of the observation data or greater than the measurement upper limit of the sensor of the observation data. The identification of the outlier point can use the Kalman filter method. For the determination of the filling value of the missing value, the measurement values within a certain sampling interval before and after the missing value can be averaged.

[0107] The category detection processing is to determine whether the data category of each element in the test data sequence of each performance indicator of the first information set is consistent with the preset data category, and to delete the inconsistent data from the test data sequence of the performance indicator to obtain a second information set.

[0108] The simulation performance evaluation value is obtained by performing simulation performance evaluation processing on the performance simulation test result information set and the index standard value set in the pretreatment information set based on the statistical fluctuation value, and the simulation performance evaluation value comprises:

[0109] The simulation test result value sequence of each performance index in the performance simulation test result information set in the pretreatment information set is subtracted from the standard value of the corresponding performance index in the index standard value set, so that a corresponding simulation difference value sequence is obtained.

[0110] Difference estimation calculation is performed on all simulation difference value sequences, so that a simulation performance evaluation value cy is obtained.

[0111] The expression of the difference estimation calculation is as follows:

[0112]

[0113] Wherein, a ij is the jth element of the simulation difference value sequence of the ith performance index, b ij is the jth element of the simulation test result value sequence of the ith performance index in the performance simulation test result information set in the pretreatment information set, and N1 and M2 are the element number of the simulation test result value sequence and the total number of performance indexes, respectively.

[0114] In a second aspect of the embodiment of the present application, a simulation evaluation device for the performance of an unmanned cluster is disclosed, and the device comprises:

[0115] A memory in which executable program codes are stored;

[0116] A processor coupled with the memory;

[0117] The processor invokes the executable program codes stored in the memory to execute the simulation evaluation method for the performance of an unmanned cluster.

[0118] In a third aspect of the embodiment of the present application, a computer storage medium is disclosed, and the computer storage medium stores computer instructions, which are invoked by a computer to execute the simulation evaluation method for the performance of an unmanned cluster.

[0119] In a fourth aspect of the embodiment of the present application, an information data processing terminal is disclosed, and the information data processing terminal is used to implement the simulation evaluation method for the performance of an unmanned cluster.

[0120] The above merely illustrates the embodiments of the present application but should not be taken as limitations. Various modifications and variations can be made to the present application based on the skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the scope of claims of the present application.

Claims

1. A method for evaluating the performance of unmanned swarms by combining numerical and real simulations, characterized in that, The method comprises the following steps: S1, collecting performance actual test result information set and performance simulation test result information set of unmanned cluster; The performance actual test result information set comprises actual test result value sequence of each performance index of the unmanned cluster; The performance simulation test result information set comprises simulation test result value sequence of each performance index of the unmanned cluster; S2, preprocessing the performance actual test result information set and the performance simulation test result information set to obtain a preprocessing information set, comprising: S21, performing data cleaning processing on the performance actual test result information set and the performance simulation test result information set to obtain a first information set; S22, performing category detection processing on the first information set to obtain a second information set; S23, performing numerical-real joint discrimination processing on the second information set to obtain the preprocessing information set, comprising: S231, calculating the average value of the simulation test result value sequence and the actual test result value sequence of each performance index in the second information set respectively; S232, subtracting the average value from the simulation test result value sequence and the actual test result value sequence of each performance index in the second information set respectively to obtain the difference sequence; S233, performing deviation upper limit estimation processing on the difference sequence of each performance index to obtain the deviation upper limit value; based on the deviation upper limit value, the deviation range information of each performance index is constructed; The deviation upper limit value of each performance index is obtained by performing deviation upper limit estimation processing on the difference sequence, comprising: The difference sequence corresponding to the simulation test result value sequence of each performance index is modeled by taking the collection time of the elements of the simulation test result value sequence corresponding to the difference sequence as the independent variable and the elements of the difference sequence as the dependent variable to obtain the regression model of the simulation test result value sequence; The difference sequence corresponding to the actual test result value sequence of each performance index is modeled by taking the collection time of the elements of the actual test result value sequence corresponding to the difference sequence as the independent variable and the elements of the difference sequence as the dependent variable to obtain the regression model of the actual test result value sequence; The regression models of the simulation test result value sequence and the actual test result value sequence are fused to obtain a fusion regression model; The collection time of all elements of the simulation test result value sequence is calculated by using the fusion regression model to obtain a deviation value sequence; The maximum value in the deviation value sequence is determined as the deviation upper limit value of the performance index; S234, discriminating whether the elements of the simulation test result value sequence and the actual test result value sequence of each performance index in the second information set are within the corresponding deviation range information, deleting the elements not within the deviation range information from the second information set to obtain the preprocessing information set; S3, performing performance comprehensive evaluation processing on the preprocessing information set to obtain the comprehensive performance evaluation value of the unmanned cluster.

2. The method of claim 1, wherein the performance of the unmanned cluster is evaluated by combining the simulation results with the real data. The performance comprehensive evaluation processing on the preprocessed information set obtains a comprehensive performance evaluation value of the unmanned cluster, and comprises: S31, obtaining an index standard value set; the index standard value set comprises a standard value of each performance index; S32, performing actual measurement performance evaluation processing on the performance actual test result information set in the preprocessed information set and the index standard value set to obtain an actual measurement performance evaluation value and a statistical fluctuation value; S33, based on the statistical fluctuation value, performing simulation performance evaluation processing on the performance simulation test result information set in the preprocessed information set and the index standard value set to obtain a simulation performance evaluation value; S34, performing weighted summation processing on the actual measurement performance evaluation value and the simulation performance evaluation value to obtain a comprehensive performance evaluation value of the unmanned cluster.

3. The method of claim 2, wherein the performance of the unmanned cluster is evaluated by combining the simulation results with the real data. The actual measurement performance evaluation processing on the performance actual test result information set in the preprocessed information set and the index standard value set to obtain an actual measurement performance evaluation value and a statistical fluctuation value comprises: S321, subtracting a standard value of a corresponding performance index in the index standard value set from an actual test result value sequence of each performance index in the performance actual test result information set in the preprocessed information set to obtain a corresponding difference value sequence; S322, using each performance index difference value sequence as a row vector to construct a difference matrix; S323, performing performance evaluation processing on the difference matrix to obtain an actual measurement performance evaluation value; S324, performing statistical fluctuation estimation processing on the difference matrix to obtain a statistical fluctuation value.

4. The method of claim 3, wherein the performance of the unmanned cluster is evaluated by combining the simulation results with the real data. The performance evaluation processing on the difference matrix to obtain an actual measurement performance evaluation value comprises: performing eigenvalue solving processing on the difference matrix to obtain an average eigenvalue λ0and a maximum eigenvalue λ max ; performing variational mode decomposition on each row vector of the difference matrix to obtain a transformation matrix Y; a row vector of the transformation matrix Y is a variational mode decomposition result of each row vector of the difference matrix; performing fusion feature calculation on the transformation matrix and the difference matrix to obtain a fusion matrix A; performing logarithmic calculation processing on diagonal elements of the fusion matrix A to obtain an actual measurement performance evaluation value.

5. A data-real combined simulation evaluation device for the performance of unmanned swarms, characterized in that, The device comprises: a memory storing executable program codes; a processor coupled with the memory; the processor invokes the executable program codes stored in the memory to execute the unmanned cluster performance numerical-realistic combined simulation evaluation method in any one of claims 1 to 4.

6. A computer storable medium, characterized by The computer storage medium stores computer instructions, which are invoked by a computer to execute the unmanned cluster performance numerical-realistic combined simulation evaluation method in any one of claims 1 to 4.

7. An information data processing terminal, characterized by The information data processing terminal is used to implement the unmanned cluster performance numerical-realistic combined simulation evaluation method in any one of claims 1 to 4.

Images