Target measurement and control station distribution method based on BP neural network

Through BP neural network analysis and selection of target measurement and control site locations, the problems of instability and inefficiency selection in traditional methods are solved, and more efficient site selection and measurement and control performance are achieved, and the quality and efficiency of flight tests are improved.

CN119004016APending Publication Date: 2024-11-22XIAN AEROSPACE PROPULSION INST
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411017031.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-26
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

The traditional target measurement and control station layout method lacks engineering methods, which leads to unstable and inefficient selection of station layout locations, which requires long-term iteration, and repeated performance comparisons are time-consuming and labor-intensive.

Method used

By using a BP neural network method, the training sample data set and the sample data set to be predicted are established, and multiple influencing factors are combined, such as the maximum distance between the site location and the target flight trajectory, the minimum tracking speed, the coverage range of the ground measurement and control antenna, etc. , use BP neural network for training and prediction, and automatically select the optimal site location.

Benefits of technology

The stability and consistent selection of the site locations are achieved, the efficiency and performance of target measurement and control are significantly improved, manual operations are reduced, and the quality and efficiency of flight test identification and evaluation are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119004016A_ABST
    Figure CN119004016A_ABST
Patent Text Reader

Abstract

The invention belongs to the field of aircraft measurement and control, and particularly relates to a target measurement and control distribution method based on a BP neural network, which is used for solving the problems of unstable effect caused by lack of an engineering method for distribution site selection in a traditional target measurement and control distribution method, time and labor consumption and long-time iteration required by repeated comparison of performance of a plurality of distribution sites, and high efficiency of target measurement and control distribution. And the efficiency is low. According to the target measurement and control station distribution method based on the BP neural network, a standardized training sample data set and a unified evaluation standard are established, so that station distribution sites can be systematically analyzed and selected, and the stability and consistency of station distribution site selection are ensured; through automatic data processing and model training, a large amount of data can be rapidly processed and analyzed, manual operation is greatly reduced, and efficiency is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of aircraft measurement and control, and in particular relates to a target measurement and control station layout method based on BP neural network. Background Art

[0002] The remote control and telemetry system is an important part of the range flight measurement and control system, used to capture and locate flight targets, measure and capture test data, etc. It consists of missile-ground remote control and telemetry equipment and supporting software and hardware, and is an important data source for flight test identification and evaluation. In target flight tests, selecting appropriate station locations is crucial to achieving the best measurement and control effect.

[0003] The traditional target measurement and control station layout method is affected by indicators such as the functional performance of the missile and ground measurement and control equipment, the measurement and control distance, the tracking performance of the ground measurement and control antenna, and the coverage of the ground measurement and control antenna. The existing technology has the following problems:

[0004] (1) The site selection lacks an engineering approach, resulting in unstable results;

[0005] (2) Repeated comparison of the performance of multiple sites is time-consuming and labor-intensive, requiring long iterations and low efficiency. Summary of the invention

[0006] The purpose of the present invention is to solve the shortcomings of traditional target measurement and control station layout methods, such as lack of engineering methods for station location selection, resulting in unstable results, and time-consuming and labor-intensive repeated comparison of the performance of multiple station locations, requiring long-term iterations and low efficiency, and to provide a target measurement and control station layout method based on BP neural network.

[0007] In order to solve the deficiencies of the above-mentioned prior art, the present invention provides the following technical solutions:

[0008] A target measurement and control station layout method based on BP neural network, which is special in that it includes the following steps:

[0009] Step 1: Establish a training sample data set, which includes multiple training samples. Each training sample includes input data and output data. The input data is the actual value I1~I2 of the site location and its n influencing factors. n , the output data is the total score S of the site location, and the value of n is 9;

[0010] The n influencing factors are respectively the maximum distance between the station location and the target flight trajectory, the minimum tracking speed of the station location for the target flight trajectory, the minimum coverage range of the ground measurement and control antenna, the minimum value of the telemetry channel margin, the minimum value of the remote control channel margin, the moment when the ground measurement and control antenna pitch angle exceeds the 1° threshold for the first time, the proportion of stable tracking time of the measurement and control signal, the maximum angular rate of the ground measurement and control antenna pitch angle in the first 30 seconds, and the distance between the station location and the starting point of the mission personnel;

[0011] The total score S is expressed as follows:

[0012]

[0013] Among them, i is the i-th influencing factor, K i , P i are the weight coefficient and initial score of the i-th influencing factor respectively; P i According to the actual value I of the i-th influencing factor i Compare with the corresponding threshold value;

[0014] Step 2: Based on the training sample data set obtained in step 1, establish and train the BP neural network;

[0015] Step 3: Establish a sample data set to be predicted and perform normalization processing. The sample data set to be predicted includes multiple samples to be predicted, each sample to be predicted includes the actual values ​​I1~I2 of the site location and its n influencing factors. n ;

[0016] Step 4: Use the BP neural network obtained in step 2 to perform target measurement and control station deployment on the multiple samples to be predicted obtained in step 3, select the station deployment position with the highest total score as the optimal station deployment position, and complete the target measurement and control station deployment.

[0017] Furthermore, the step 1 is specifically as follows:

[0018] Step 1.1, set corresponding thresholds and scoring criteria for n influencing factors of each site location;

[0019] Step 1.2: According to the actual values ​​of n influencing factors I1~I2 of each site location, n , and the threshold and scoring criteria set in step 1.1, the initial scores P1~P of n influencing factors at each site are obtained. n ;

[0020] Step 1.3: According to the initial scores P1~P of n influencing factors at each site n , get the total score S of each site location, and then get the training sample data set.

[0021] Furthermore, the step 1.1 is specifically as follows:

[0022] Set the thresholds of n influencing factors to be \(I'_1\) to \(I'_n\) respectively. n ;

[0023] Set the scoring criterion for the maximum distance between the station - setting point and the target flight trajectory as: if \(I_1>I'_1\), then \(P_1 = 0\); if \(I_1\leq I'_1\), then \(P_1\propto(I'_1 - I_1)\);

[0024] Set the scoring criterion for the minimum tracking speed of the station - setting point for the target flight trajectory as: if \(I_2<I'_2\), then \(P_2 = 0\); if \(I_2\geq I'_2\), then \(P_2\propto(I_2 - I'_2)\);

[0025] Set the scoring criterion for the minimum coverage range of the ground TT&C antenna as: if \(I_3<I'_3\), then \(P_3 = 0\); if \(I_3\geq I'_3\), then \(P_3\propto(I_3 - I'_3)\);

[0026] Set the scoring criterion for the minimum value of the telemetry channel margin as: if \(I_4<I'_4\), then \(P_4 = 0\); if \(I_4\geq I'_4\), then \(P_4\propto(I_4 - I'_4)\);

[0027] Set the scoring criterion for the minimum value of the remote - control channel margin as: if \(I_5<I'_5\), then \(P_5 = 0\); if \(I_5\geq I'_5\), then \(P_5\propto(I_5 - I'_5)\);

[0028] Set the scoring criterion for the ground TT&C antenna's elevation angle exceeding \(1^{\circ}\) for the first time as: if \(I_6>I'_6\), then \(P_6 = 0\); if \(I_6\leq I'_6\), then \(P_6\propto(I'_6 - I_6)\);

[0029] Set the scoring criterion for the proportion of the stable tracking time of the TT&C signal as: if \(I_7<I'_7\), then \(P_7 = 0\); if \(I_7\geq I'_7\), then \(P_7\propto(I'_7 - I_7)\);

[0030] Set the scoring criterion for the maximum value of the elevation - angle rate of the ground TT&C antenna in the first 30 s as: if \(I_8>I'_8\), then \(P_8 = 0\); if \(I_8\leq I'_8\), then \(P_8\propto(I'_8 - I_8)\);

[0031] Set the scoring criterion for the distance between the station - setting point and the starting point of the mission personnel as: if \(I_9>I'_9\), then \(P_9 = 0\); if \(I_9\leq I'_9\), then \(P_9\propto(I'_9 - I_9)\).

[0032] Furthermore, the threshold \(I'_1\) of the maximum distance between the station - setting point and the target flight trajectory is the smaller one of the telemetry operating distance and the remote - control operating distance of the ground station;

[0033] The minimum tracking speed \(v\) of the station - setting point for the target flight trajectory t is expressed as follows:

[0034] \(v\) t \(=\) min \(\cdot\omega\)max

[0035] Among them, r min is the minimum distance between the target flight trajectory and the location of the station, ω max is the maximum angular velocity of the ground measurement and control antenna;

[0036] The minimum tracking speed v of the target flight trajectory for the station location t The threshold value I′2 is the maximum flight speed of the target;

[0037] The stable tracking time proportion TP of the measurement and control signal is expressed as follows:

[0038] TP=(T2-T1) / T

[0039] Among them, T2 is the time when the pitch angle in the descent phase is less than the pitch angle threshold of the ground measurement and control antenna tracking, T1 is the time when the pitch angle of the target after climbing is greater than the pitch angle threshold of the ground measurement and control antenna tracking, and T is the total time.

[0040] Furthermore, the step 2 is specifically as follows:

[0041] Step 2.1, set the number of nodes in the input layer, hidden layer and output layer;

[0042] Step 2.2, define the activation function and training algorithm;

[0043] Step 2.3, configure training parameters;

[0044] Step 2.4, initialize the weight matrix W1 from the input layer to the hidden layer, the weight matrix W2 from the hidden layer to the output layer, the bias vector B1 of the hidden layer, and the bias vector B2 of the output layer;

[0045] Step 2.5: Use the training sample data set obtained in step 1 to train the BP neural network.

[0046] Furthermore, the step 2.2 is specifically as follows:

[0047] The activation function of the hidden layer is defined as tansig, the activation function of the output layer is purelin, and the training algorithm is Levenberg-Marquardt algorithm.

[0048] Furthermore, in step 2.1, the number of nodes in the input layer, hidden layer, and output layer are n, 6, and 1, respectively;

[0049] In step 2.3, the training parameters include the number of training times, learning rate and training target error, which are 1000 times, 0.01, 10 -5 .

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] (1) The target measurement and control station layout method based on the BP neural network of the present invention can systematically analyze and select the station layout locations by establishing a standardized training sample data set and a unified evaluation standard, thereby ensuring the stability and consistency of the station layout location selection; through automated data processing and model training, it can quickly process and analyze large amounts of data, greatly reducing manual operations and improving efficiency.

[0052] (2) The present invention can comprehensively consider a variety of influencing factors, including the maximum distance between the station location and the target flight trajectory, the minimum tracking speed, the coverage of the ground measurement and control antenna, the channel margin, the pitch angle change, etc. Through the comprehensive analysis and weighted calculation of these influencing factors, it is ensured that the selected station location can achieve the best measurement and control effect.

[0053] (3) The present invention has high flexibility and can adjust the weights and thresholds of influencing factors according to actual needs to adapt to different measurement and control conditions and needs. By adjusting the structure and parameters of the neural network, the prediction model can be continuously optimized and improved to adapt to different application scenarios.

[0054] (4) The present invention can significantly improve the overall performance of the remote control and telemetry system by optimizing the selection of station locations, including the accuracy of capturing and locating flight targets, the reliability of measuring and capturing test data, etc. This helps to improve the quality and efficiency of flight test identification and evaluation, and provides strong support for the optimization of the flight measurement and control system at the range. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a flow chart of an embodiment of a target measurement and control station deployment method based on BP neural network of the present invention;

[0056] Figure 2 Schematic diagram of the distribution of the station locations 1# to 5# in an embodiment of the present invention. DETAILED DESCRIPTION

[0057] The present invention will be further described below in conjunction with the accompanying drawings and exemplary embodiments.

[0058] Reference Figure 1 A target measurement and control station layout method based on BP neural network includes the following steps:

[0059] Step 1: Establish a training sample data set, which includes 50 training samples. Each training sample has input data and output data. The input data is the actual value I1~I2 of the site location and its n influencing factors. n , the output data is the total score S of the site location, and the value of n is 9;

[0060] The n influencing factors are the maximum distance between the station location and the target flight trajectory, the minimum tracking speed of the station location to the target flight trajectory, the minimum coverage range of the ground measurement and control antenna, the minimum value of the telemetry channel margin, the minimum value of the remote control channel margin, the first time the ground measurement and control antenna elevation angle exceeds the 1° threshold, the proportion of the stable tracking time of the measurement and control signal, the maximum angular velocity of the ground measurement and control antenna elevation angle in the first 30 seconds, and the distance between the station location and the starting point of the mission personnel;

[0061] The minimum tracking speed v of the target flight trajectory for the station location t It is expressed as follows:

[0062] v t =r min ·ω max

[0063] Among them, r min is the minimum distance between the target and the station location on the flight trajectory, ω max is the maximum angular velocity of the ground measurement and control antenna;

[0064] The stable tracking time proportion TP of the measurement and control signal is expressed as follows:

[0065] TP=(T2-T1) / T

[0066] Among them, T2 is the time when the pitch angle in the descending stage is less than the pitch angle threshold of the ground measurement and control antenna tracking, T1 is the time when the pitch angle of the target after climbing is greater than the pitch angle threshold of the ground measurement and control antenna tracking, and T is the total time;

[0067] Step 1.1: Set the corresponding thresholds and scoring criteria for the n influencing factors of each site location, as shown in Table 1:

[0068] Table 1

[0069]

[0070] Step 1.2: According to the actual values ​​of n influencing factors I1~I2 of each site location, n ,, and the threshold and scoring criteria set in step 1.1, the initial scores P1~P of n influencing factors at each site are obtained. n ;

[0071] Step 1.3: According to the initial scores P1~P of n influencing factors at each site n , get the total score S of each site location, and then get the training sample data set;

[0072] The total score S is as follows

[0073]

[0074] Among them, i is the i-th influencing factor, K i is the weight coefficient of the i-th influencing factor;

[0075] Step 2: Based on the training sample data set obtained in step 1, establish and train the BP neural network;

[0076] Step 2.1, set the number of input layer nodes to n, the number of hidden layer nodes to 6, and the number of output layer nodes to 1;

[0077] Step 2.2, define the activation function of the hidden layer as tansig (Hyperbolic Tangent Sigmoid Transfer Function), the activation function of the output layer as purelin (Linear Transfer Function), and the training algorithm as Levenberg-Marquardt algorithm;

[0078] Step 2.3, configure training parameters: set the number of training times to 1000, set the learning rate to 0.01, and set the training target error to 10 -5 ;

[0079] Step 2.4, initialize the weight matrix W1 from the input layer to the hidden layer, the weight matrix W2 from the hidden layer to the output layer, the bias vector B1 of the hidden layer, and the bias vector B2 of the output layer;

[0080] Step 2.5, use the training sample data set obtained in step 1.4 to train the BP neural network;

[0081] Step 3: Establish a sample data set to be predicted and perform normalization processing. The sample data set to be predicted includes 5 samples to be predicted. Each sample to be predicted includes the actual values ​​I1~I2 of the site location and its n influencing factors. n , specifically:

[0082] by Figure 2 The target flight trajectory measurement and control station location selection is taken as an example. The target flight trajectory is shown as the arrow trajectory in the figure. L1 and L2 are the starting point and the end point respectively. A is the starting point of the mission personnel. 1#~5# are the station locations. The actual values ​​of n influencing factors of each station location are I1~I n As shown in Table 2:

[0083] Table 2

[0084]

[0085]

[0086] Step 4: Use the BP neural network obtained in step 2 to perform target measurement and control station layout to predict the sample data set to be predicted obtained in step 3. The total scores of station locations 1# to 5# are 90.0, 82.8, 82.4, 92.9 and 86 respectively. Therefore, station location 4# with the highest total score is selected as the optimal station location to complete the target measurement and control station layout.

Claims

1. A target measurement and control station layout method based on BP neural network, characterized in that: It includes the following steps: Step 1: Establish a training sample data set, which includes multiple training samples. Each training sample includes input data and output data. The input data is the actual value I1~I2 of the site location and its n influencing factors. n , the output data is the total score S of the site location, and the value of n is 9; The n influencing factors are respectively the maximum distance between the station deployment point and the target flight trajectory, the minimum tracking speed of the station deployment point for the target flight trajectory, the minimum coverage range of the ground TT&C antenna, the minimum value of the telemetry channel margin, the minimum value of the remote control channel margin, the moment when the elevation angle of the ground TT&C antenna first exceeds the 1° threshold, the proportion of the stable tracking time of the TT&C signal, the maximum value of the elevation angle rate of the ground TT&C antenna in the first 30 s, and the distance between the station deployment point and the starting point of the mission personnel; The total score S is expressed as follows: Among them, i is the i-th influencing factor, K i , P i are the weight coefficient and initial score of the i-th influencing factor respectively; P i According to the actual value I of the i-th influencing factor i Compare with the corresponding threshold value; Step 2: Based on the training sample data set obtained in Step 1, establish and train a BP neural network; Step 3: Establish a sample data set to be predicted and perform normalization processing. The sample data set to be predicted includes multiple samples to be predicted, each sample to be predicted includes the actual values ​​I1~I2 of the site location and its n influencing factors. n ; Step 4: Use the BP neural network obtained in Step 2 to perform target TT&C station deployment for the multiple samples to be predicted obtained in Step 3, and select the station deployment point with the highest total score as the optimal station deployment point to complete the target TT&C station deployment.

2. According to the target measurement and control station layout method based on BP neural network in claim 1, it is characterized in that: The specific content of Step 1 is as follows: Step 1.1: Set corresponding thresholds and scoring criteria for the n influencing factors of each station deployment point; Step 1.2: According to the actual values ​​of n influencing factors I1~I2 of each site location, n , and the threshold and scoring criteria set in step 1.1, the initial scores P1~P of n influencing factors at each site are obtained. n ; Step 1.3: According to the initial scores P1~P of n influencing factors at each site n , get the total score S of each site location, and then get the training sample data set.

3. The target measurement and control station layout method based on BP neural network according to claim 2 is characterized in that: The specific content of Step 1.1 is: Set the thresholds of n influencing factors to be I′1~I′ n ; Set the scoring criterion for the maximum distance between the station deployment point and the target flight trajectory as if I1 > I′1, then P1 = 0; if I1 ≤ I′1, then P1 ∝ (I′1 - I1); Set the scoring criterion for the minimum tracking speed of the station deployment point for the target flight trajectory as if I2 < I′2, then P2 = 0; if I2 ≥ I′2, then P2 ∝ (I2 - I′2); Set the scoring criterion for the minimum coverage range of the ground TT&C antenna as if I3 < I′3, then P3 = 0; if I3 ≥ I′3, then P3 ∝ (I3 - I′3); Set the scoring criterion for the minimum value of the telemetry channel margin as if I4 < I′4, then P4 = 0; if I4 ≥ I′4, then P4 ∝ (I4 - I′4); Set the scoring criterion for the minimum value of the remote control channel margin as if I5 < I′5, then P5 = 0; if I5 ≥ I′5, then P5 ∝ (I5 - I′5); Set the scoring criterion for the elevation angle of the ground TT&C antenna when it first exceeds 1° as if I6 > I′6, then P6 = 0; if I6 ≤ I′6, then P6 ∝ (I′6 - I6); Set the scoring criterion for the proportion of the stable tracking time of the TT&C signal as if I7 < I′7, then P7 = 0; if I7 ≥ I′7, then P7 ∝ (I′7 - I7); Set the scoring criterion for the maximum value of the elevation angle rate of the ground TT&C antenna in the first 30 s as if I8 > I′8, then P8 = 0; if I8 ≤ I′8, then P8 ∝ (I′8 - I8); Set the scoring criterion for the distance between the station deployment point and the starting point of the mission personnel as if I9 > I′9, then P9 = 0; if I9 ≤ I′9, then P9 ∝ (I′9 - I9).

4. A method for target TT&C station deployment based on a BP neural network according to claim 3, characterized in that: The threshold I′1 for the maximum distance between the station deployment point and the target flight trajectory is the smaller one of the telemetry operating distance of the ground station and the remote control operating distance of the ground station; The minimum tracking speed v of the target flight trajectory for the station location t It is expressed as follows: v t =r min ·ω max Among them, r min is the minimum distance between the target flight trajectory and the location of the station, ω max is the maximum angular velocity of the ground measurement and control antenna; The minimum tracking speed v of the target flight trajectory for the station location t The threshold value I′2 is the maximum flight speed of the target; The proportion TP of the stable tracking time of the TT&C signal is expressed as follows: TP = (T2 - T1) / T Among them, T2 is the time when the pitch angle in the descent phase is less than the pitch angle threshold of the ground measurement and control antenna tracking, T1 is the time when the pitch angle of the target after climbing is greater than the pitch angle threshold of the ground measurement and control antenna tracking, and T is the total time.

5. A target measurement and control station layout method based on BP neural network according to any one of claims 1 to 4, characterized in that: The step 2 is specifically as follows: Step 2.1, set the number of nodes in the input layer, hidden layer and output layer; Step 2.2, define the activation function and training algorithm; Step 2.3, configure training parameters; Step 2.4, initialize the weight matrix W1 from the input layer to the hidden layer, the weight matrix W2 from the hidden layer to the output layer, the bias vector B1 of the hidden layer, and the bias vector B2 of the output layer; Step 2.5: Use the training sample data set obtained in step 1 to train the BP neural network.

6. A target measurement and control station layout method based on BP neural network according to claim 5, characterized in that: The step 2.2 is specifically as follows: The activation function of the hidden layer is defined as tansig, the activation function of the output layer is purelin, and the training algorithm is Levenberg-Marquardt algorithm.

7. The target measurement and control station layout method based on BP neural network according to claim 5 is characterized in that: In step 2.1, the number of nodes in the input layer, hidden layer, and output layer are n, 6, and 1 respectively; In step 2.3, the training parameters include the number of training times, the learning rate, and the training target error, which are 1000 times, 0.01, and 10 respectively.