A method for selecting features for high-speed aircraft confrontation based on distance correlation coefficient

Abstract

The invention relates to a method for selecting features of high-speed aircraft confrontation based on distance correlation coefficient, comprising the following steps: S1: collecting original data, using a simulation environment to perform a high-speed aircraft confrontation simulation experiment, and collecting data affecting high-speed aircraft confrontation variables; S2: removing abnormal values ​​and repeated values ​​from the data affecting high-speed aircraft confrontation variables obtained in step S1, retaining normal data and performing range normalization processing; S3: dividing the data after range normalization processing into two groups, and constructing a linear combination of all variables in the independent variable group; S4: transmitting the linear combination weight coefficient of the independent variable in the linear combination constructed in step S2 as a parameter to be optimized to a particle swarm optimization algorithm, and using the distance correlation coefficient between the independent variable linear combination and the dependent variable as the fitness function of the particle swarm optimization algorithm; the invention has the advantages of improving the overall design capability and refined design level of high-speed aircraft confrontation.

Description

Technical Field

[0001] The invention belongs to the technical field of high-speed aircraft, and in particular relates to a high-speed aircraft confrontation feature selection method based on distance correlation coefficient. Background Art

[0002] Today, as high-speed aircraft defense weapons and technologies continue to develop, developing effective countermeasures and methods to deal with high-speed aircraft defense weapons, thereby maintaining and improving the countermeasures of high-speed aircraft, is the only way to ensure that high-speed aircraft continue to play their due military role, and has very important strategic significance. Efficient countermeasures can easily avoid the enemy's air defense system, reduce the risk of being intercepted by the enemy, improve the strike capability and operational flexibility of high-speed aircraft, enable them to more effectively strike enemy targets, and increase strategic deterrence capabilities.

[0003] In order to improve the overall design capability and refined design level of high-speed aircraft confrontation, ensure that qualitative requirements can be put forward for various index factors affecting high-speed aircraft confrontation, extract the key factors affecting high-speed aircraft confrontation under the premise of ensuring economy, rationality and effectiveness, and give the correlation degree of each index to guide which indicators need to be focused on in the overall design of high-speed aircraft confrontation, so as to better put forward requirements for the influencing parameters of each system.

[0004] The current feature selection methods assume that the relationship between features and target variables is linear and cannot capture nonlinear relationships. In this case, the feature selection methods cannot select the best feature subset. In addition, the existing feature selection methods are sensitive to the distribution of data. Inappropriate data distribution will make it impossible to find the best feature subset.

[0005] Based on the problem requirements in this technical background and the shortcomings of existing feature selection methods, the present invention is particularly proposed. Summary of the invention

[0006] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a high-speed aircraft confrontation feature selection method based on distance correlation coefficient, which can guide which indicators need to be focused on in the overall design of high-speed aircraft confrontation, better put forward requirements for the influencing parameters of each system, and improve the overall design capability and refined design level of high-speed aircraft confrontation.

[0007] The technical solution of the present invention is as follows:

[0008] A high-speed aircraft confrontation feature selection method based on distance correlation coefficient includes the following steps:

[0009] Step S1: collecting original data, using a simulation environment to conduct a high-speed aircraft confrontation simulation experiment, and collecting data on variables that affect the high-speed aircraft confrontation;

[0010] Step S2: remove abnormal values ​​and duplicate values ​​from the data of the high-speed aircraft confrontation variables obtained in step S1, retain the normal data and perform range normalization processing;

[0011] Step S3: Divide the data after range normalization into two groups, namely, the independent variable group and the dependent variable group, and construct linear combinations of all variables in the independent variable group;

[0012] Step S4: transmitting the linear combination weight coefficient of the independent variables in the linear combination constructed in step S2 as the parameter to be optimized to the particle swarm optimization algorithm, and taking the distance correlation coefficient between the linear combination of the independent variables and the dependent variable as the fitness function of the particle swarm optimization algorithm;

[0013] Step S5: After multiple rounds of iterations of the particle swarm optimization algorithm, a set of linear combination weight coefficients that can maximize the fitness function is finally obtained;

[0014] Step S6: Analyze and screen the linear combination weight coefficients obtained in step S5, calculate the contribution rate of each linear combination weight coefficient, sort the contribution rates of all independent variable group variables from large to small, and add them up one by one from large to small. If the contribution rate exceeds 90%, the variables that have been added up will be used as the key elements of high-speed aircraft confrontation to complete the feature selection.

[0015] Furthermore, the range normalization process in step S2 is specifically to make a difference between the normal data and the minimum value in the normal data, and divide it by the difference between the maximum value and the minimum value in the normal data.

[0016] Furthermore, in step S3, the independent variable with a large contribution to the correlation is found by constructing its linear combination, wherein the method of constructing the linear combination is specifically to select all the independent variable indicators of the independent variable group as a group of vectors, determine the corresponding weight for each vector in the group, multiply each vector with the corresponding weight and then add them to obtain the final linear combination vector, that is, the independent variable with a large contribution to the correlation.

[0017] Furthermore, the distance correlation coefficient in step S4 is a standardized form of the distance covariance, in the form of a ratio, wherein the numerator is the square root of the distance covariance between the two variables, and the denominator is the square root of the product of the square roots of the distance covariance of the two variables.

[0018] Furthermore, the distance covariance of the distance correlation coefficient in step S4 needs to calculate the Euclidean distance between the linear combination of the independent variable and each pair of observations of the dependent variable in the corresponding space, perform combined product, and finally calculate the average value.

[0019] Furthermore, the particle swarm optimization algorithm in step S5 specifically includes the following contents:

[0020] The mathematical model abstracted by the particle swarm optimization algorithm is:

[0021]

[0022] Where h(α) represents the function to be optimized, and α represents the variable to be optimized in the function. It represents the value of the variable α corresponding to the maximum value of the optimization function h(α). Here, the linear combination weight variable of the independent variable is used as the variable to be optimized;

[0023] The constraint equation is:

[0024]

[0025] dcorr(α 1 x 1 +α 2 x 2 +…+α m x m ,y) as the constraint equation represents the distance correlation coefficient between the linear combination of the independent variable group variables and the dependent variable group variables, α 1 x 1 +α 2 x 2 +…+α m x m Indicates that the constructed linear combination of independent variables is used as the first parameter of the distance correlation coefficient operation, x 1 ,x 2 …,x m represents each independent variable, α 1 ,α 2 ...,α m represents the linear combination weight coefficient of each independent variable, and y represents the dependent variable as the second parameter of the distance correlation coefficient operation;

[0026] The constraints are:

[0027] α i ∈(-1,1),1≤i≤m

[0028] Where m represents the number of variables that need to be optimized, that is, constraint α i ∈(-1,1),1≤i≤m, the condition is that the weight coefficient of the linear combination of each independent variable takes a value range between ±1.

[0029] Furthermore, the steps of multiple rounds of iterations of the particle swarm optimization algorithm in step S5 are as follows:

[0030] S601: Initialize various parameters of the particle swarm;

[0031] S602: Calculate the fitness function value of each particle;

[0032] S603: Update the optimal position of each particle and the global optimal position of all particles according to the fitness;

[0033] S604: If the set maximum number of iterations is reached or the global optimal position meets the minimum limit, the algorithm ends, otherwise returns to step S603.

[0034] Furthermore, the calculation formula for the correlation degree of the linear combination weight coefficient in step S6 is the absolute value of the linear combination weight coefficient corresponding to the linear combination weight coefficient divided by the sum of the absolute values ​​of all indicator linear combination weight coefficients.

[0035] Furthermore, the countermeasure variables affecting the high-speed aircraft include the high-speed aircraft maneuvering time, the high-speed aircraft maneuvering moment, the high-speed aircraft maneuvering direction, the high-speed aircraft maneuvering overload, the high-speed aircraft guidance law guidance coefficient and the distance deviation.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] 1. The range normalization process of the present invention only involves linear transformation and scaling of data, without the need for complex mathematical operations or model assumptions, and can process data quickly and effectively in large-scale data sets.

[0038] 2. The range normalization of the present invention can preserve the distribution shape and relative relationship of the original data, that is, it does not change the relative size and order between the data. This helps to maintain the inherent structure and information of the data, so that the normalized data can still reflect the characteristics and trends of the original data.

[0039] 3. The distance correlation coefficient of the present invention can not only detect linear relationships, but also capture nonlinear relationships. This means that even if there is a complex nonlinear relationship between two variables, the distance correlation coefficient can reflect this relationship.

[0040] 4. The distance correlation coefficient of the present invention does not depend on the distribution assumption of the data and is more robust.

[0041] 5. The particle swarm optimization algorithm of the present invention has a strong global search capability and can effectively explore and find potential optimal solutions in the solution space.

[0042] 6. The particle swarm optimization algorithm of the present invention has an adaptive search strategy and parameter adjustment mechanism, which can be dynamically adjusted and optimized according to the characteristics of the problem and the structure of the solution space, thereby improving the robustness and efficiency of the algorithm.

[0043] 7. The goal of the particle swarm optimization algorithm of the present invention is to search and find the optimal solution in the solution space. By processing the search and exploration process of multiple particles in parallel, the convergence speed of the algorithm can be accelerated and the global search capability can be improved. This enables the algorithm to more effectively explore potential optimal solutions in the solution space.

[0044] 8. By calculating the contribution rate, we can identify which variables have the greatest contribution to the correlation of the dependent variable, so as to select variables, and capture most of the information in the original data through fewer variables, which is the key factor. It can guide which indicators need to be improved in the overall design of high-speed aircraft confrontation, better put forward requirements for the influencing parameters of each system, and improve the overall design capability and refined design level of high-speed aircraft confrontation. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the process of the present invention;

[0046] Figure 2 It is the fitness function iteration curve diagram of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] A high-speed aircraft confrontation feature selection method based on distance correlation coefficient, such as Figure 1-2 As shown, the following steps are included:

[0049] Step S1: Collecting raw data, using Python language to write code according to the high-speed aircraft and arresting aircraft confrontation logic to build a simulation environment for simulation experiments, and collecting data that affects high-speed aircraft confrontation variables;

[0050] In this example, the variables affecting the high-speed aircraft confrontation as indicators for analysis include high-speed aircraft maneuvering time, high-speed aircraft maneuvering moment, high-speed aircraft maneuvering direction, high-speed aircraft maneuvering overload, high-speed aircraft guidance law guidance coefficient and distance deviation. The original data of the simulation experiment are shown in Table 1 below:

[0051]

[0052] Table 1

[0053] Step S2: Data processing: for the data obtained in step S1, outliers and duplicate values ​​are removed, normal data are retained, and range normalization is performed;

[0054] Due to the unreasonable setting of variable value range, some data that violates objective laws may be generated during the simulation experiment, which are called outliers. In order to ensure the reliability of the analysis results, it is necessary to clean these original data, delete outliers and duplicate values, and perform range normalization to eliminate the dimensional differences between different indicators. The specific implementation method is as follows:

[0055] Let X represent the attribute variable, x represent the sample value of attribute X, minx and maxx represent the minimum and maximum values ​​of attribute X, then the value of the attribute after range normalization is:

[0056]

[0057] That is, the sample value of attribute X is subtracted from the minimum value in the sample, and then divided by the difference between the maximum value and the minimum value in the sample, where attribute X is the original data after deleting outliers and duplicate values;

[0058] The original data of the simulation experiment after data cleaning and range normalization are shown in the following table:

[0059] Table 2

[0060]

[0061] Step S3: Divide the data into two groups, namely, the independent variable group and the dependent variable group, and construct linear combinations of all variables in the independent variable group;

[0062] The correlation between two sets of variables is captured by constructing a linear combination of the independent variables and exploring the relationship between the linear combination and the dependent variable. The specific construction method is as follows:

[0063]

[0064] X 1 ,X 2 ,…X m It represents all the indicator elements in the independent variable, that is, the attribute variable; α 1 ,α 2 ,…α m represents the linear combination weight coefficient, For the newly constructed linear combination variable, that is, first take all the independent variable indicators as a group of vectors, determine the corresponding weight for each vector in the group, multiply each vector with the corresponding weight, and then add them to get the final linear combination vector.

[0065] Step S4: transmitting the linear combination weight coefficient of the independent variables in the linear combination constructed in step S2 as the parameter to be optimized to the particle swarm optimization algorithm, and taking the distance correlation coefficient between the linear combination of the independent variables and the dependent variable as the fitness function of the particle swarm optimization algorithm;

[0066] The particle swarm optimization algorithm searches for the optimal solution to a problem by simulating the cooperation and information sharing between individuals in a group. The basic idea is to make the entire particle swarm gradually tend to the optimal solution by constantly adjusting the position and speed of particles in the search space. The fitness function is an indicator used to quantify the quality of the solution in the particle swarm optimization algorithm. It can guide the movement direction of particles and update the individual optimal solution and the global optimal solution of each particle. It plays a key role in the performance and convergence of the algorithm.

[0067] In order to better describe the correlation between two sets of variables, the present invention uses the distance correlation coefficient as the fitness function. The larger the distance correlation coefficient, the stronger the correlation between the two sets of variables, and vice versa. The distance correlation coefficient is defined as:

[0068] Assume {(ξ i ,θ i ), i = 1, 2, ..., n} is a sample of the population (ξ, θ), where ξ and θ represent any two variables, and i represents the index value of each sample. The distance correlation coefficient between the two groups of sample variables is:

[0069]

[0070] The part of the formula can be further rewritten as

[0071]

[0072] where dcov 2 (ξ,θ) represents the distance covariance between the two variables ξ and θ, n represents the total number of samples, and i and j represent the index value of each sample. ||ξ i -ξ j || represents the Euclidean distance between the sample value of the i-th index of the variable ξ and the sample value of the j-th index, ||θ i -θ j ‖ represents the Euclidean distance between the sample value of the i-th index and the sample value of the j-th index of the variable θ, ‖θ j -θ l ‖ represents the Euclidean distance between the sample value of the jth index and the sample value of the lth index of the variable θ, ||ξ i -ξ l || represents the Euclidean distance between the sample value of the i-th index and the sample value of the l-th index of the variable ξ

[0073]

[0074] where dcov 2 (θ,θ) represents the distance covariance of the single θ variable, n represents the total number of samples, and i, j represent the index value of each sample. ‖θ i -θ j || represents the Euclidean distance between the sample value of the i-th index and the sample value of the j-th index of the variable θ, ||θ j -θ l || represents the Euclidean distance between the sample value of the jth index and the sample value of the lth index of the variable θ, ‖θ i -θ l || represents the Euclidean distance between the sample value of the i-th index and the sample value of the l-th index of the variable θ

[0075]

[0076] where dcov 2 (ξ,ξ) represents the distance covariance of the single ξ variable, n represents the total number of samples, and i, j represent the index value of each sample. ||ξ i -ξ j || represents the Euclidean distance between the sample value of the i-th index and the sample value of the j-th index of the variable ξ, ||ξ i -ξ l || represents the Euclidean distance between the sample value of the i-th index and the sample value of the l-th index of the variable ξ, ||ξ j -ξ l || represents the Euclidean distance between the sample value of the jth index and the sample value of the lth index of the variable ξ. Here, the newly constructed linear combination variable The distance correlation coefficient between As the fitness function of the particle swarm optimization algorithm;

[0077] Step S5: After multiple rounds of iterations of the particle swarm optimization algorithm, a set of linear combination weight coefficients that can maximize the fitness function is finally obtained;

[0078] The mathematical model abstracted by the particle swarm optimization algorithm is:

[0079]

[0080] Where h(α) represents the function to be optimized, and α represents the variable to be optimized in the function. It represents the value of variable α when the optimization function h(α) reaches the maximum. Here, the linear combination weight variable of the independent variable is used as the variable to be optimized;

[0081] The constraint equation is:

[0082]

[0083] dcorr(α 1 x 1 +α 2 x 2 +…+α m x m ,y) as the constraint equation represents the distance correlation coefficient between the linear combination of independent variables and the dependent variable, α 1 x 1 +α 2 x 2 +…+α m x m Indicates that the constructed linear combination of independent variables is used as the first parameter of the distance correlation coefficient operation, x 1 ,x 2 …,x m represents each independent variable, α 1 ,α 2 …,α m represents the linear combination weight coefficient of each independent variable, and y represents the dependent variable as the second parameter of the distance correlation coefficient operation;

[0084] The constraints are:

[0085] α i ∈(-1,1),1≤i≤m

[0086] Where m represents the number of variables that need to be optimized, that is, the constraint condition is that the value range of the linear combination weight coefficient of each independent variable is between ±1;

[0087] The steps of the particle swarm optimization algorithm are as follows:

[0088] Step S601: Initialize various parameters of the particle swarm;

[0089] These include population learning factor, individual learning factor, inertia weight factor, number of iterations, population size, minimum particle speed limit, maximum particle speed limit, minimum particle position limit, maximum particle position limit, number of indicator elements and minimum fitness limit;

[0090] Step S602: Calculate the fitness function value of each particle;

[0091] That is, the distance correlation coefficient value of each particle is calculated as the fitness function;

[0092] Step S603: updating the optimal position of each particle and the global optimal position of all particles according to the fitness;

[0093] Step S604: If the algorithm reaches the set maximum number of iterations or the global optimal position meets the minimum limit, the algorithm ends; otherwise, it returns to step S603.

[0094] The set of linear combinations that maximize the correlation between two sets of variables is called the key linear combination. The key linear combination describes which independent variables contribute most to the correlation between the two sets of variables, and gives quantitative results in the form of linear combination weight coefficients. The larger the absolute value of the linear combination weight coefficient, the greater the contribution, and vice versa.

[0095] After multiple rounds of iterations, the key linear combination weight coefficients finally obtained are shown in the following table:

[0096]

[0097] Table 3

[0098] Step S6: Analyze and screen the linear combination weight coefficients obtained in step S5, calculate the contribution rate of each linear combination weight coefficient, sort the contribution rates of all independent variable group variables from large to small, and add them up one by one from large to small. If the contribution rate exceeds 90%, the variables that have been added up are used as the key elements of high-speed aircraft confrontation to complete feature selection;

[0099] Among them, the contribution rate calculation formula of each variable affecting the high-speed aircraft confrontation is:

[0100]

[0101] where |α i | represents the absolute value of the linear combination weight coefficient of the i-th element;

[0102] That is, the contribution rate calculation formula of each indicator is the absolute value of the linear combination weight coefficient corresponding to the indicator divided by the sum of the absolute values ​​of the linear combination weight coefficients of all indicators.

[0103] Indicator name Maneuvering time Moment of mobility Maneuvering overload Maneuvering direction Guidance law coefficient Relevance 35.3％ 4.99％ 56.12％ 0 3.59％

[0104] Table 4

[0105] The results of sorting the linear combination weight coefficients in descending order are shown in the following table:

[0106] Indicator name Contribution rate Maneuvering overload 56.12％ Maneuvering time 35.3％ Moment of mobility 4.99％ Guidance law coefficient 3.59％ Maneuvering direction 0

[0107] Table 5

[0108] It can be seen from the above table that the contribution rate of maneuvering overload is the largest, the correlation degree of maneuvering direction is the smallest, and the contribution rate of maneuvering overload and maneuvering time is one order of magnitude larger than that of maneuvering moment and guidance law guidance coefficient;

[0109] By adding up the contribution rates of the linear combination weight coefficients of the independent variables from large to small, it is found that when the sum of the contribution rate of maneuvering overload and the correlation degree of maneuvering time is greater than 90%, maneuvering overload and maneuvering time are the key elements of high-speed aircraft confrontation, which is of great significance for guiding the overall design of high-speed aircraft confrontation.

[0110] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for selecting high-speed aircraft confrontation features based on distance correlation coefficient, characterized in that: The steps include: Step S1: collecting original data, using a simulation environment to conduct a high-speed aircraft confrontation simulation experiment, and collecting data on variables that affect the high-speed aircraft confrontation; Step S2: remove abnormal values ​​and duplicate values ​​from the data of the high-speed aircraft confrontation variables obtained in step S1, retain the normal data and perform range normalization processing; Step S3: Divide the data after range normalization into two groups, namely, the independent variable group and the dependent variable group, and construct linear combinations of all variables in the independent variable group; Step S4: transmitting the linear combination weight coefficient of the independent variables in the linear combination constructed in step S2 as the parameter to be optimized to the particle swarm optimization algorithm, and taking the distance correlation coefficient between the linear combination of the independent variables and the dependent variable as the fitness function of the particle swarm optimization algorithm; The distance correlation coefficient is a standardized form of the distance covariance, using the ratio method, where the numerator is the square root of the distance covariance between the two variables, and the denominator is the square root of the product of the square roots of the distance covariance of the two variables. The distance covariance of the distance correlation coefficient requires calculating the Euclidean distance between the linear combination of the independent variable and each pair of observations of the dependent variable in the corresponding space, performing the combined product, and finally calculating the average value; Step S5: After multiple rounds of iterations of the particle swarm optimization algorithm, a set of linear combination weight coefficients that can maximize the fitness function is finally obtained; The particle swarm optimization algorithm specifically includes the following contents: The mathematical model abstracted by the particle swarm optimization algorithm is: Where h(α) represents the function to be optimized, and α represents the variable to be optimized in the function. It represents the value of the variable α corresponding to the maximum value of the optimization function h(α). Here, the linear combination weight variable of the independent variable is used as the variable to be optimized; The constraint equation is: dcorr(α1x1+α2x2+…+α m x m ,y) as the constraint equation represents the distance correlation coefficient between the linear combination of the independent variable group variables and the dependent variable group variables, α1x1+α2x2+…+α m x m Indicates that the constructed linear combination of independent variables is used as the first parameter of the distance correlation coefficient operation, x1, x2…, x m Represents each independent variable, α1, α2…, α m represents the linear combination weight coefficient of each independent variable, and y represents the dependent variable as the second parameter of the distance correlation coefficient operation; The constraints are: a i ∈(-1,1),1≤i≤m Where m represents the number of variables that need to be optimized, that is, constraint α i ∈(-1,1),1≤i≤mThe condition is that the weight coefficient of each independent variable linear combination ranges between ±1; Step S6: Analyze and screen the linear combination weight coefficients obtained in step S5, calculate the contribution rate of each linear combination weight coefficient, sort the contribution rates of all independent variable group variables from large to small, and add them up one by one from large to small. If the contribution rate exceeds 90%, the variables that have been added up will be used as the key elements of high-speed aircraft confrontation to complete the feature selection.

2. The method for selecting high-speed aircraft countermeasure features based on distance correlation coefficient according to claim 1, characterized in that: The range normalization process in step S2 is specifically to make a difference between the normal data and the minimum value in the normal data, and divide it by the difference between the maximum value and the minimum value in the normal data.

3. The method for selecting high-speed aircraft countermeasure features based on distance correlation coefficient according to claim 1, characterized in that: In step S3, the independent variable with a large contribution to the correlation is found by constructing its linear combination, wherein the method of constructing the linear combination is specifically to select all the independent variable indicators of the independent variable group as a group of vectors, determine the corresponding weight for each vector in the group, multiply each vector with the corresponding weight and then add them to obtain the final linear combination vector, that is, the independent variable with a large contribution to the correlation.

4. The method for selecting high-speed aircraft countermeasure features based on distance correlation coefficient according to claim 1, characterized in that: The steps of multiple rounds of iterations of the particle swarm optimization algorithm in step S5 are as follows: S601: Initialize various parameters of the particle swarm; S602: Calculate the fitness function value of each particle; S603: Update the optimal position of each particle and the global optimal position of all particles according to the fitness; S604: If the set maximum number of iterations is reached or the global optimal position meets the minimum limit, the algorithm ends, otherwise returns to step S603.

5. The method for selecting high-speed aircraft confrontation features based on distance correlation coefficient according to claim 1, characterized in that: The calculation formula for the correlation degree of the linear combination weight coefficient in step S6 is the absolute value of the linear combination weight coefficient corresponding to the linear combination weight coefficient divided by the sum of the absolute values ​​of all indicator linear combination weight coefficients.

6. The method for selecting high-speed aircraft countermeasure features based on distance correlation coefficient according to claim 1, characterized in that: The countermeasure variables that affect the high-speed aircraft include the high-speed aircraft maneuvering time, the high-speed aircraft maneuvering moment, the high-speed aircraft maneuvering direction, the high-speed aircraft maneuvering overload, the high-speed aircraft guidance law guidance coefficient and the distance deviation.

Images