A radar simulation model verification method and system

Abstract

A radar simulation model verification method and system, constructing a high-resolution simulation model; establishing a signal-to-noise ratio model, a radar detection distance model, a radar detection probability model, and a radar tracking accuracy model; performing error analysis on the signal-to-noise ratios of the high-resolution model and the signal-to-noise ratio model, and introducing factor control to improve the signal-to-noise ratio model when the error requirements are not met; substituting the values ​​of other parameters into the detection distance, detection probability, and tracking accuracy models, and verifying that when the above three meet a certain range, the signal-to-noise ratio meets the value within the error requirement range, and the intersection of the value range of the signal-to-noise ratio is determined as the reasonable range of the signal-to-noise ratio; based on the output data of the high-resolution model and the functional-level simulation model, an error optimization model is established to solve the optimal control factor value; all functional-level models and the high-resolution model are verified, and all errors meet the requirements. The present invention can solve the technical problem of adjusting and verifying the radar simulation model.

Description

Technical Field

[0001] The present invention relates to the field of radar simulation model verification technology, and in particular to a radar simulation model verification method and system. Background Art

[0002] Model verification includes two types: one is based on "causality" and the other is based on "data correlation". The "causality" model should not only reproduce the structure within the system, but also explain the generation and operation of the system behavior, as well as the possibility of changing the system behavior. The "data correlation" model ignores the causal relationship within the system structure and focuses on the consistency between the output of the model and the displayed data. In comparison, the verification of the "causality" model is more complicated, and the system dynamics model belongs to the "causality" model. The six verification methods of the model are extreme value test, sensitivity test, boundary sufficiency test, dimensional consistency test, parameter evaluation test, mental model test and authenticity test. Among them, the first five verification methods belong to model structure verification, and their purpose is to verify the accuracy of the model structure; the historical test belongs to model behavior verification, and its purpose is to verify the consistency between the output of the model and the actual data. The extreme value test is to test whether the model parameters are consistent with the behavior of the real system under extreme conditions under extreme conditions. It is also a test of the "robustness" of the model, but the overall error is not analyzed in detail; the sensitivity test is a model verification method to test whether the changes in other variables are reasonable when a parameter in the model changes. Sensitivity testing is used to analyze model parameters with low estimation accuracy and detect the changes in model simulation results when the values ​​of these parameters change; boundary adequacy testing is to test whether various parameters that are important to the research problem are already included in the model, whether the model has important variables that have not been considered, and whether the model needs to be further expanded to include more content. The model should have an appropriate boundary. If the boundary is too small, it cannot cover the important variables involved in the research problem. If the boundary is too large, the model loses its generality and reduces its universality; dimensional consistency testing ensures that the model conforms to the laws of matter and logic. At the same time, the process of correcting the model dimensions is also a process of testing the model structure and equation expressions; parameter evaluation is to compare the parameters in the model with the values ​​of the corresponding variables in the real system in theory or experience, so as to test the validity of the model parameter settings; mental model testing is a way to test whether the simulation results of the model are consistent with the actual situation. However, mental model testing is not a sufficient test of whether the model is correct or not; authenticity testing is used to judge the truth or falsity of information. In the existing technology, extreme value tests, sensitivity tests, boundary sufficiency tests and dimensional consistency tests are rarely used to analyze the overall error accuracy of the model. In addition, the real data of a certain type of radar is difficult to obtain, and parameter evaluation, mental models and authenticity tests cannot be used.

[0003] The verification referred to in the traditional simulation model VV&A includes two parts: validation and verification. Validation mainly checks the model implementation process to determine whether the model is correctly implemented, and verification mainly processes the simulation model output and the actual system output to determine whether the correct model is implemented. In the existing radar simulation technology, there is a lack of an efficient and fast-converging verification and verification method to quickly determine whether the simulation model can realize the radar model to be simulated. Summary of the invention

[0004] The present invention provides a radar simulation model verification method and system, which are used to solve the technical problem of adjusting and verifying the radar simulation model.

[0005] The present invention proposes a radar simulation model verification method, which comprises:

[0006] Based on the performance parameters of the target model radar, a high-resolution model of the target model radar is constructed;

[0007] Establish a radar simulation model, which includes a signal-to-noise ratio model, a detection distance model, a detection probability model, and a tracking accuracy model;

[0008] Assume that the signal-to-noise ratio data set of the signal-to-noise ratio model {T i} and the signal-to-noise ratio dataset of the high-resolution model {Q i}, calculate the signal-to-noise ratio standard deviation S of the signal-to-noise ratio model respectively. T and the signal-to-noise ratio standard deviation S of the high-resolution model Q ; Then establish the signal-to-noise ratio error function Δ=|S Q -S T |;

[0009] Let the control factor be θ i , the control factor function is g(θ i )=a i θ i , i=1,2,...,n, where a i is the control factor θ i The weight of all weights is 1, and each weight a i is the average of the sum of all weights;

[0010] The objective function is established with the minimum signal-to-noise ratio error function Δ as the goal, and the constraint condition is the standard deviation S of the signal-to-noise ratio in the signal-to-noise ratio model. T , the standard deviation of the signal-to-noise ratio of the high-resolution model S Q , set the control factor θ i The range of values; Based on the above conditions, g(θi ) is multiplied by the noise ratio model to establish an optimization model and solve the control factor θ i The value of is multiplied by the signal-to-noise ratio model to calculate the signal-to-noise ratio value for verifying the radar simulation model;

[0011] The verification process is that if the signal-to-noise ratio value meets the error requirement range, the detection distance, detection probability, and tracking accuracy of the radar simulation model are calculated based on the signal-to-noise ratio value and the detection distance model, detection probability model, and tracking accuracy model of the radar simulation model; if the detection distance, detection probability, and tracking accuracy meet the error requirement range, the radar simulation model is verified to be qualified;

[0012] Otherwise, the radar simulation model is verified to be unqualified, then the sum of the weights of the control factors is 1 and each weight a i It is a constraint condition of the average value of all weights, continues to introduce new control factors and retains the values ​​of the original control factors, continues to solve the values ​​of the new control factors based on the optimization model, and continues the verification until the verification is successful.

[0013] Furthermore, when the control factor is θ1, the control factor function is g(θ1)=a1θ1, a1=1, and the analytical form of the optimization model is:

[0014] minΔ=|S T -S Q |

[0015]

[0016] Among them, T i is the signal-to-noise ratio data of the signal-to-noise ratio model, T is the signal-to-noise ratio data set of the signal-to-noise ratio model {T i}, Q i is the signal-to-noise ratio data of the high-resolution model, Q is the signal-to-noise ratio dataset of the high-resolution model {Q i}, the subscript i is the number of the data, N is a natural number, P t is the radar transmission power, G t is the transmitting antenna gain in this direction, G r is the receiving antenna gain in this direction, λ is the echo signal wavelength, k is the Boltzmann constant, R is the distance between the radar and the target, T0 is the radar system temperature, B is the instantaneous bandwidth of the radar receiver, L is the radar RF front-end loss, N f is the noise coefficient, SNRθ is the radar signal-to-noise ratio corresponding to the control factor θ1, the value range of the control factor θ1 is [c1, c2], [c1, c2] depends on the specific problem; the value b1 of θ1 is determined by analytical calculation of the optimization model.

[0017] Furthermore, if the value b1 of θ1 cannot make the signal-to-noise ratio, detection distance, detection probability, and tracking accuracy of the radar simulation model meet the requirements, then the value b1 of θ1 is used to transform the control factor function to g(θ1, θ2) = 0.5b1 + 0.5θ2, and the new control factor θ2 is continued to be calculated based on the optimization model, and the radar simulation model is verified using the control factor θ2. If the radar simulation model is successfully verified, the verification ends; otherwise, new control factors are continued to be introduced, and the radar simulation model is verified using the control factors until the radar simulation model is successfully verified.

[0018] Furthermore, the signal-to-noise ratio model SNR of the radar simulation model is:

[0019]

[0020] Where P t : Radar transmitting power;

[0021] G t : The transmitting antenna gain in this direction;

[0022] G r : The receiving antenna gain in this direction;

[0023] λ: echo signal wavelength;

[0024] R: The distance between radar and target detection;

[0025] k: Boltzmann constant;

[0026] T0: radar system temperature;

[0027] B: receiver instantaneous bandwidth;

[0028] L: Radar RF front-end loss;

[0029] N f : Noise coefficient;

[0030] SNR: radar signal-to-noise ratio;

[0031] The detection distance model R of the radar simulation model is:

[0032]

[0033] Where P t : Transmit power;

[0034] G: antenna gain in this direction;

[0035] λ: echo signal wavelength;

[0036] σ: radar cross section,

[0037] k: Boltzmann constant;

[0038] B: receiver instantaneous bandwidth;

[0039] L: Radar RF front-end loss;

[0040] SNR: radar signal-to-noise ratio;

[0041] T0: radar system temperature;

[0042] The detection probability model P of the radar simulation model d for:

[0043]

[0044] Where erfc() is the complementary error function,

[0045] P f : false alarm probability;

[0046] The tracking accuracy model of the radar simulation model is:

[0047]

[0048] In the formula, S j : signal-to-noise ratio of a single pulse;

[0049] c: speed of light;

[0050] τ0: pulse width;

[0051] n: number of pulse trains.

[0052] The present invention also proposes a radar simulation model verification system, which comprises:

[0053] A high-resolution model building module builds a high-resolution model of the target model radar based on the performance parameters of the target model radar;

[0054] Radar simulation model building module, to build radar simulation model, radar simulation model includes signal-to-noise ratio model, detection distance model, detection probability model, tracking accuracy model;

[0055] The simulation model verification module performs the following processing: assuming that the signal-to-noise ratio data set of the signal-to-noise ratio model {T i} and the signal-to-noise ratio dataset of the high-resolution model {Q i}, calculate the signal-to-noise ratio standard deviation S of the signal-to-noise ratio model respectively. T and the signal-to-noise ratio standard deviation S of the high-resolution model Q ; Then establish the signal-to-noise ratio error function Δ=|S Q -S T |;

[0056] Let the control factor be θ i , the control factor function is g(θ i )=a i θ i , i=1,2,...,n, where a i is the control factor θ i The weight of all weights is 1, and each weight a i is the average of the sum of all weights;

[0057] The objective function is established with the minimum signal-to-noise ratio error function Δ as the goal, and the constraint condition is the standard deviation S of the signal-to-noise ratio in the signal-to-noise ratio model. T , the standard deviation of the signal-to-noise ratio of the high-resolution model S Q , set the control factor θ i The range of values; Based on the above conditions, g(θ i ) is multiplied by the noise ratio model to establish an optimization model and solve the control factor θ i The value of is multiplied by the signal-to-noise ratio model to calculate the signal-to-noise ratio value for verifying the radar simulation model;

[0058] The verification process is that if the signal-to-noise ratio value meets the error requirement range, the detection distance, detection probability, and tracking accuracy of the radar simulation model are calculated based on the signal-to-noise ratio value and the detection distance model, detection probability model, and tracking accuracy model of the radar simulation model; if the detection distance, detection probability, and tracking accuracy meet the error requirement range, the radar simulation model is verified to be qualified;

[0059] Otherwise, the radar simulation model is verified to be unqualified, then the sum of the weights of the control factors is 1 and each weight a i It is a constraint condition of the average value of all weights, continues to introduce new control factors and retains the values ​​of the original control factors, continues to solve the values ​​of the new control factors based on the optimization model, and continues the verification until the verification is successful.

[0060] Furthermore, the simulation model verification module further performs the following processing: when the control factor is θ1, the control factor function is g(θ1)=a1θ1, a1=1, and the analytical form of the optimization model is:

[0061]

[0062] Among them, T i is the signal-to-noise ratio data of the signal-to-noise ratio model, T is the signal-to-noise ratio data set of the signal-to-noise ratio model {T i}, Q i is the signal-to-noise ratio data of the high-resolution model, Q is the signal-to-noise ratio dataset of the high-resolution model {Q i}, the subscript i is the number of the data, N is a natural number, P t is the radar transmission power, G t is the transmitting antenna gain in this direction, G r is the receiving antenna gain in this direction, λ is the echo signal wavelength, k is the Boltzmann constant, R is the distance between the radar and the target, T0 is the radar system temperature, B is the instantaneous bandwidth of the radar receiver, L is the radar RF front-end loss, N f is the noise figure, SNR θ is the radar signal-to-noise ratio corresponding to the control factor θ1. The value range of the control factor θ1 is [c1, c2], and [c1, c2] depends on the specific problem. The value b1 of θ1 is determined by analytical calculation of the optimization model.

[0063] Furthermore, the simulation model verification module further performs the following processing: if the value b1 of θ1 cannot make the signal-to-noise ratio, detection distance, detection probability, and tracking accuracy of the radar simulation model meet the requirements, then the value b1 of θ1 is used to transform the control factor function to g(θ1, θ2) = 0.5b1 + 0.5θ2, and the new control factor θ2 is continued to be calculated based on the optimization model, and the radar simulation model is verified using the control factor θ2. If the radar simulation model is successfully verified, the verification ends; otherwise, new control factors are continued to be introduced, and the radar simulation model is verified using the control factors until the radar simulation model is successfully verified.

[0064] Furthermore, the signal-to-noise ratio model SNR of the radar simulation model is:

[0065]

[0066] Where P t : Radar transmitting power;

[0067] G t : The transmitting antenna gain in this direction;

[0068] G r : The receiving antenna gain in this direction;

[0069] λ: echo signal wavelength;

[0070] R: The distance between radar and target detection;

[0071] k: Boltzmann constant;

[0072] T0: radar system temperature;

[0073] B: receiver instantaneous bandwidth;

[0074] L: Radar RF front-end loss;

[0075] Nf : Noise coefficient,

[0076] SNR: radar signal-to-noise ratio;

[0077] The detection distance model R of the radar simulation model is:

[0078]

[0079] Where P t : Transmit power;

[0080] G: antenna gain in this direction;

[0081] λ: echo signal wavelength;

[0082] σ: radar cross section,

[0083] k: Boltzmann constant;

[0084] B: receiver instantaneous bandwidth;

[0085] L: Radar RF front-end loss;

[0086] SNR: radar signal-to-noise ratio;

[0087] T0: radar system temperature;

[0088] The detection probability model P of the radar simulation model d for:

[0089]

[0090] Where erfc() is the complementary error function,

[0091] P f : false alarm probability;

[0092] The tracking accuracy model of the radar simulation model is:

[0093]

[0094] In the formula, S j : signal-to-noise ratio of a single pulse;

[0095] c: speed of light;

[0096] τ0: pulse width;

[0097] n: number of pulse trains.

[0098] The present invention relies on a certain type of radar simulation model to conduct air interference experiments and establish a radar functional level model. Assuming that the error is within a reasonable range, the radar functional level model is verified using simulation data and reasonable parameter values ​​are obtained by optimizing the model calculation. BRIEF DESCRIPTION OF THE DRAWINGS

[0099] Figure 1 To improve the signal-to-noise ratio model bias;

[0100] Figure 2 0.5m 2 Improved signal-to-noise ratio model data output;

[0101] Figure 3 2m 2 Improved signal-to-noise ratio model data output;

[0102] Figure 4 To modify the wavelength output difference;

[0103] Figure 5 To modify the power output. DETAILED DESCRIPTION

[0104] The present invention provides a radar simulation model verification method and system. The specific implementation of the present invention is described in detail below.

[0105] The radar simulation model verification method includes:

[0106] Based on the performance parameters of the target model radar, a high-resolution model of the target model radar is constructed.

[0107] A radar simulation model is established, which includes a signal-to-noise ratio model, a detection distance model, a detection probability model, and a tracking accuracy model.

[0108] Assume that the signal-to-noise ratio data set of the signal-to-noise ratio model {T i} and the signal-to-noise ratio dataset of the high-resolution model {Q i}, calculate the signal-to-noise ratio standard deviation S of the signal-to-noise ratio model respectively. T and the signal-to-noise ratio standard deviation S of the high-resolution model Q ; Then establish the signal-to-noise ratio error function Δ=|S Q -S T |.

[0109] Let the control factor be θ i , the control factor function is g(θ i )=a i θ i , i=1,2,...,n, where a i is the control factor θ i The weight of all weights is 1, and each weight a i is the average of the sum of all weights.

[0110] The objective function is established with the minimum signal-to-noise ratio error function Δ as the goal, and the constraint condition is the standard deviation S of the signal-to-noise ratio in the signal-to-noise ratio model.T , the standard deviation of the signal-to-noise ratio of the high-resolution model S Q , set the control factor θ i The range of values; Based on the above conditions, g(θ i ) is multiplied by the noise ratio model to establish an optimization model and solve the control factor θ i The value is multiplied by the signal-to-noise ratio model to calculate the signal-to-noise ratio value for verifying the radar simulation model.

[0111] The verification process is that if the signal-to-noise ratio value meets the error requirement range, the detection distance, detection probability, and tracking accuracy of the radar simulation model are calculated based on the signal-to-noise ratio value and the detection distance model, detection probability model, and tracking accuracy model of the radar simulation model. If the detection distance, detection probability, and tracking accuracy meet the error requirement range, the radar simulation model is verified to be qualified.

[0112] Otherwise, the radar simulation model is verified to be unqualified, then the sum of the weights of the control factors is 1 and each weight a i It is a constraint condition of the average value of all weights, continues to introduce new control factors and retains the values ​​of the original control factors, continues to solve the values ​​of the new control factors based on the optimization model, and continues the verification until the verification is successful.

[0113] The specific steps of the method of the present invention are:

[0114] 1. Consult the relevant literature of a certain type of radar, determine the performance parameters of a certain type of radar, and use the ArkSIM platform to build a high-resolution simulation model.

[0115] 2. Establish radar functional level model

[0116] (1) Signal-to-noise ratio model

[0117]

[0118] Where P t : Radar transmission power, unit: W;

[0119] G t : The transmitting antenna gain in this direction;

[0120] G r : The receiving antenna gain in this direction;

[0121] λ: wavelength of echo signal, in meters;

[0122] R is the distance between the radar and the target detection; k: Boltzmann constant, 1.38×10-23J / K;

[0123] T0: radar system temperature;

[0124] B: receiver instantaneous bandwidth, in Hz;

[0125] L: Radar RF front-end loss (times);

[0126] N f : Noise coefficient;

[0127] SNR: Radar signal-to-noise ratio.

[0128] (2) Detection distance model

[0129] The formula for radar's target detection distance R is as follows:

[0130]

[0131] Where P t : Transmit power, unit: W;

[0132] G: antenna gain in this direction;

[0133] λ: wavelength of echo signal, in meters;

[0134] σ: radar cross section,

[0135] k: Boltzmann constant, 1.38×10-23J / K;

[0136] B: receiver instantaneous bandwidth, in Hz;

[0137] L: Radar RF front-end loss (times);

[0138] SNR: radar signal-to-noise ratio;

[0139] T0: radar system temperature, usually 290K is used as the system temperature;

[0140] (3) Discovery Probabilistic Model

[0141] The average detection probability P of the radar to the target d Calculation formula:

[0142]

[0143] In the formula, is the complementary error function,

[0144] P f : False alarm probability,

[0145] U T : Threshold level.

[0146] (4) Tracking accuracy model

[0147] Using pulse trains to measure the distance to a target is a common ranging method used by radar. When there is an appropriately high single pulse signal-to-noise ratio, the radar's tracking accuracy is:

[0148]

[0149] In the formula, S j : signal-to-noise ratio of a single pulse;

[0150] c: speed of light;

[0151] τ0: pulse width;

[0152] n: number of pulse trains;

[0153] T0: System temperature, usually 290K is used as the system temperature.

[0154] 3. Determine the optimization goal and establish the signal-to-noise ratio error function

[0155] Both the radar functional level model and the high-resolution model are based on the radar equation and the radar interference equation. It can be said that the principles and processing procedures are basically the same. In the radar theoretical model, the comprehensive signal-to-noise / signal-to-interference ratio of the radar is the core factor of the radar model, and the signal-to-noise ratio model is the basis. At the same time, the signal-to-noise ratio is also a variable of detection distance, detection probability and tracking accuracy. Ultimately, the determination of these three functions is determined by the signal-to-noise ratio. Therefore, the absolute value of the difference between the standard deviation of the signal-to-noise ratio model and the high-resolution model is determined as the objective function.

[0156] Combining mathematical statistics knowledge, using the signal-to-noise ratio data output by the simulation model and the signal-to-noise ratio model, the absolute value of the difference in standard deviation between the two is calculated as the error. The error is set within a%, indicating that the simulation model has passed the verification.

[0157] Assume that the signal-to-noise ratio data set of the signal-to-noise ratio model is {T i}, i = 1, ..., N, the signal-to-noise ratio data set of the high-resolution model under the same conditions is {Q i}, and calculate the standard deviation of the signal-to-noise ratio as follows.

[0158] {T i The mean of} is Standard Deviation

[0159] {Q i The mean of} is Standard Deviation

[0160] Then the signal-to-noise ratio error function is

[0161] Δ=|S Q -S T |

[0162] When the model error is within a%, it can be considered that the three (detection distance, detection probability and tracking accuracy) functional level models are also consistent with the high-resolution model; when the error is greater than a%, the calibrated signal-to-noise ratio model needs to be corrected to reduce its error to within a%.

[0163] 4. Propose a factor control method and formulate constraints. Since the radar simulation model and the signal-to-noise ratio model are established by different mechanisms, the errors generated are also different, often large or small, and may not all meet the requirements. If the signal-to-noise ratio error can be controlled within a certain range, then its detection distance, detection probability and tracking accuracy can also ensure that the error is within a reasonable range. In order to control the error, a control factor is added to the radar signal-to-noise ratio model, and a method for determining the number and expression of the control factor is proposed.

[0164] The factor control method is to add several factors to the signal-to-noise ratio calculation, and control the output of the signal-to-noise ratio by adjusting the factor value to meet the error requirements. The main content of introducing the control factor is to first add a factor to the signal-to-noise ratio model, and optimize the error to be within the set range. If the calculation result still cannot meet the error requirements, then on this basis, add another factor to make the result closer to the error requirements. Therefore, this method can achieve the purpose of gradually approaching a reasonable error.

[0165] Method for determining the number of control factors and expressions:

[0166] (1) Determination of the number of control factors: First, add a factor to the signal-to-noise ratio model. If the error requirement is not met, use two factors. If the error requirement is still not met, use three factors, and so on, until the error requirement is met.

[0167] (2) Method for determining the control factor expression

[0168] ① Based on the high-resolution model, detection distance, detection probability and tracking accuracy model, the signal-to-noise ratio value of each model can be calculated when the error is within a certain range, and the intersection of the signal-to-noise ratio values ​​is taken as the signal-to-noise ratio value of the signal-to-noise ratio model.

[0169] ② Assume the control factor is θ i , the control factor function is g(θ i )=a i θ i , (i=1,2,...,n), where a i is the control factor θ i The sum of all weights is 1, and each weight is the average of all weights, that is,

[0170] ③ When the control factor is θ1, the control factor function is g(θ1)=a1θ1, a1=1, and the optimization model is established: the objective function Δ1 is the absolute value of the difference between the signal-to-noise ratio model and the high-resolution model signal-to-noise ratio standard deviation, and the constraint condition is the standard deviation of the signal-to-noise ratio in the signal-to-noise ratio model S T , the standard deviation of the signal-to-noise ratio of the high-resolution model S Q , a signal-to-noise ratio model of the control factor is introduced (hereinafter referred to as the improved signal-to-noise ratio model), and the value range of the control factor θ1 is [c1, c2], where [c1, c2] depends on the specific problem.

[0171] The optimization model is

[0172]

[0173] Through calculation, the model can obtain the value of θ1 as b1. If the error requirement is met, b1 is substituted into SNR θ ,Right now Keep the error within a reasonable range.

[0174] ④ If the error requirement is not met, the calculated value b1 can be assigned to θ1 within an error accuracy. The specific method is: first set the error accuracy ε, the error is c%, calculate the difference between the absolute value of the difference between the signal-to-noise ratio standard deviation and the error requirement c%, when this difference is less than or equal to a given value ε, it is considered that the accuracy requirement of ε is met, when |Δ1-c%|≤ε, calculate θ1 as b1, increase the factor θ2, and then set the control factor function to g(θ1, θ2)=0.5b1+0.5θ2, and the optimization model is updated as

[0175]

[0176] [c3,c4] depends on the specific problem. The obtained θ2 is b2. If the error requirements are met, θ2 is substituted into the improved signal-to-noise ratio model.

[0177] ⑤ If the error requirement is not met, when the error accuracy satisfies |Δ2-c%|≤ε, θ2 is b2, and the factor θ3 is increased, g(θ1, θ 2， θ3)=0.34b1+0.33b2+0.33θ3, establish the optimization model,

[0178]

[0179] If the solved θ3 meets the error requirements, substitute the θ3 value into the improved radar signal-to-noise ratio model, and continue in this way until the error meets the requirements.

[0180] 5. Verify the above optimization results, change the parameters such as radar cross-section, wavelength and transmission power in the improved signal-to-noise ratio model, and perform error analysis on the obtained signal-to-noise ratio and the signal-to-noise ratio of the high-resolution model. Then substitute the signal-to-noise ratio solved by the improved signal-to-noise ratio model into the detection distance, detection probability and tracking accuracy model, and perform error analysis with the high-resolution model. If the error requirements are met, the verification is completed.

[0181] Example

[0182] Model verification when there is no radar interference and no atmospheric attenuation

[0183] 1. The high-resolution simulation model was run using the ArkSIM platform to obtain the signal-to-noise ratio at different distances without interference and atmospheric absorption loss, as shown in Table 1.

[0184] Table 1 High-resolution model output

[0185] time sensor target dis snr pd 1s mpq64 fighter1 155191 5.24889 0.0191333 2.875s mpq64 fighter1 154409 5.33664 0.0207089 4.75s mpq64 fighter1 153627 5.42483 0.0223876 6.625s mpq64 fighter1 152845 5.51348 0.0241737 8.5s mpq64 fighter1 152063 5.60257 0.0260718 10.375s mpq64 fighter1 151281 5.69213 0.0280865 12.25s mpq64 fighter1 150499 5.78215 0.0302222 14.125s mpq64 fighter1 149717 5.87264 0.0324835 16s mpq64 fighter1 148936 5.96361 0.0348748 17.875s mpq64 fighter1 148154 6.05505 0.0374006 19.75s mpq64 fighter1 147372 6.14697 0.0400654 21.625s mpq64 fighter1 146590 6.23939 0.0428733 23.5s mpq64 fighter1 145808 6.3323 0.0458288 25.375s mpq64 fighter1 145026 6.42571 0.0489359 27.25s mpq64 fighter1 144244 6.51962 0.0521987 29.125s mpq64 fighter1 143462 6.61405 0.055621 31s mpq64 fighter1 142680 6.70899 0.0592068 32.875s mpq64 fighter1 141898 6.80445 0.0629596 34.75s mpq64 fighter1 141116 6.90044 0.0668828 36.625s mpq64 fighter1 140335 6.99696 0.0709796 38.5s mpq64 fighter1 139553 7.09402 0.0752533 40.375s mpq64 fighter1 138771 7.19163 0.0797065 42.25s mpq64 fighter1 137989 7.28979 0.084342 44.125s mpq64 fighter1 137207 7.38851 0.089162 46s mpq64 fighter1 136425 7.48779 0.0941688 47.875s mpq64 fighter1 135643 7.58764 0.0993641 49.75s mpq64 fighter1 134861 7.68807 0.104749 51.625s mpq64 fighter1 134079 7.78908 0.110326 53.5s mpq64 fighter1 133297 7.89068 0.116096

[0186] 2. First, calculate the signal-to-noise ratio from the detection distance, detection probability and tracking accuracy as the signal-to-noise ratio function value of the signal-to-noise ratio model.

[0187] (1) Detection distance

[0188] If the detection distance error is to be less than 20%, with other conditions unchanged, the signal-to-noise ratio should be in the range of [0.482, 2.44].

[0189] (2) Probability of discovery

[0190] When the false alarm probability is set to 10 -6 When the radar detection probability error is to be less than 20%, the signal-to-noise ratio should be in the range of [0.5,1.08].

[0191] (3) Tracking accuracy

[0192] When the signal-to-noise ratio is lower than 4, the tracking accuracy of the radar is

[0193]

[0194] If the pulse width is 1 microsecond and the number of pulse trains is 10, in order to make the radar tracking accuracy less than 20%, the signal-to-noise ratio should be between [0.7,1.57].

[0195] In summary, when the signal-to-noise ratio is in the range of [0.7, 1.08], the errors of detection distance, detection probability, tracking accuracy, etc. can be guaranteed to be within 20%.

[0196] 3. Introduce control factors, improve the signal-to-noise ratio model, and establish an optimization model

[0197] A control factor is added to the signal-to-noise ratio model, and the output of the signal-to-noise ratio is controlled by optimizing the factor value to meet the error requirements.

[0198]

[0199] Use genetic algorithm to solve the optimal θ1. (There are many optimization algorithms here, and genetic algorithm is selected.) The solution steps are as follows:

[0200] 1) Initialization: subgroup 20, maximum number of iterations 50,

[0201] 2) Calculate the fitness of the group Δ1 = |S T -S Q |,

[0202] 3) Crossover operation: The crossover probability is 0.8,

[0203] 4) Mutation operation: the mutation probability is 0.1,

[0204] 5) The optimal solution is output, the error is set at 20%, and θ1 is 16.52.

[0205] 4. Substitute the θ1 value into the improved signal-to-noise ratio model

[0206]

[0207] Verify whether the detection distance, detection probability and tracking accuracy all meet the error requirements. Input the detection distance data and perform error analysis on the improved signal-to-noise ratio model. All are within 20%. Figure 1 is to improve the signal-to-noise ratio model bias.

[0208] In order to verify the adaptability of the improved signal-to-noise ratio model, the improved signal-to-noise ratio model was verified by modifying the target RCS, wavelength, and transmission power. 2 Modified to 0.5m 2 , the SNR at different distances is shown in Table 2. Table 2 is 0.5m 2 High resolution model output table. Improved signal-to-noise ratio model data output such as Figure 2 , Figure 2 0.5m 2 Improved signal-to-noise ratio model output results.

[0209] Table 2

[0210]

[0211] Modify RCS to 2m 2 The calculated high-resolution model output SNR is shown in Table 3. The improved signal-to-noise ratio model data results are shown in Figure 3 ,Figure 3 2m 2 Improved signal-to-noise ratio model output results.

[0212] Table 3

[0213]

[0214] Conclusion: By modifying the RCS value, the output results of the improved signal-to-noise ratio model meet the error requirements.

[0215] Modify the wavelength to 9.5GHz, calculate the SNR for the 2m2 target as shown in Table 4, modify the frequency target high-resolution model output. Improve the input parameters of the signal-to-noise ratio model, and the calculation results are as follows Figure 4 , Figure 4 Modify the wavelength output difference.

[0216] Table 4

[0217] time sensor target dis snr pd 1s mpq64 fighter1 155191 8.25919 0.138323 2.875s mpq64 fighter1 154409 8.34694 0.143906 4.75s mpq64 fighter1 153627 8.43513 0.149625 6.625s mpq64 fighter1 152845 8.52378 0.155479 8.5s mpq64 fighter1 152063 8.61287 0.161468 10.375s mpq64 fighterl 151281 8.70243 0.16759 12.25s mpq64 fighter1 150499 8.79245 0.173845 14.125s mpq64 fighter1 149717 8.88294 0.180232 16s mpq64 fighter1 148936 8.97391 0.186748 17.875s mpq64 fighter1 148154 9.06535 0.193392 19.75s mpq64 fighter1 147372 9.15727 0.200163 21.625s mpq64 fighter1 146590 9.24969 0.207059 23.5s mpq64 fighter1 145808 9.3426 0.214077 25.375s mpq64 fighter1 145026 9.43601 0.221215 27.25s mpq64 fighter1 144244 9.52992 0.22847 29.125s mpq64 fighter1 143462 9.62435 0.235841 31s mpq64 fighter1 142680 9.71929 0.243324 32.875s mpq64 fighter1 141898 9.81475 0.250917 34.75s mpq64 fighter1 141116 9.91074 0.258617 36.625s mpq64 fighter1 140335 10.0073 0.26642 38.5s mpq64 fighter1 139553 10.1043 0.274323 40.375s mpq64 fighter1 138771 10.2019 0.282323 42.25s mpq64 fighter1 137989 10.3001 0.290417 44.125s mpq64 fighter1 137207 10.3988 0.2986 46s mpq64 fighter1 136425 10.4981 0.306869 47.875s mpq64 fighter1 135643 10.5979 0.315221 49.75s mpq64 fighter1 134861 10.6984 0.32365 51.625s mpq64 fighter1 134079 10.7994 0.332154 53.5s mpq64 fighter1 133297 10.901 0.340728

[0218] Modify the power and change the radar radiation power to half power, that is, 40 kilowatts. Other conditions are the same as before. The calculation results are shown in Table 5. Table 5 is the output of the modified power high-resolution model and the output of the improved signal-to-noise ratio model. Figure 5 .

[0219] Table 5

[0220]

[0221] Through the analysis of the above experimental results, it can be seen that by increasing the control factor, the signal-to-noise ratio can be controlled within the specified range.

[0222] Based on the ArkSIM simulation platform, using a high-resolution model, 100 Monte Carlo simulations were performed to conduct non-interference experiments. The improved signal-to-noise ratio was substituted into the detection distance, discovery probability and tracking accuracy models for verification. All of them met the requirement of within 20% error. The comparison results are shown in Table 6. Table 6 is a comparison of the errors of the two models.

[0223] Table 6

[0224] High-resolution model Functional-level model Error Detection range 55.117km 53.71km 2.55％ Probability of detection 0.9 0.917 1.4％ Tracking accuracy 3.975m 4.296m 8.08％

[0225] The present invention firstly collects relevant data of a certain type of radar, comprehensively studies and judges the relevant parameters of the radar, and forms the performance parameter intelligence of the radar; based on this, a high-resolution simulation model is constructed using the ArkSIM platform; thirdly, according to the theoretical knowledge of radar, a signal-to-noise ratio model, a radar detection distance model, a radar detection probability model and a radar tracking accuracy model are established, wherein the detection distance, the detection probability and the tracking accuracy are functions of the signal-to-noise ratio; fourthly, since the signal-to-noise ratio is the most important one among many indicators, the verification of the high-resolution model is transformed into an error test of the signal-to-noise ratio. When the error meets certain requirements, it is considered to have passed the verification. The specific method is to convert the high-resolution model into an error test of the signal-to-noise ratio. The signal-to-noise ratio of the high-resolution model and the signal-to-noise ratio model is analyzed for error. When the error requirements are not met, the factor control method is introduced to improve the signal-to-noise ratio model. The values ​​of other parameters are substituted into the detection distance, detection probability and tracking accuracy models, and it is verified that when the above three meet a certain range, the signal-to-noise ratio meets the value within the error requirement range, and the intersection of the signal-to-noise ratio value range is determined as the reasonable range of the signal-to-noise ratio. Fifth, based on the output data of the high-resolution model and the functional-level simulation model, an error optimization model is established, and the model is solved to obtain the optimal control factor value so that the error meets the requirements. Finally, all functional-level models and high-resolution models are verified, and all errors meet the requirements.

Claims

1. A radar simulation model verification method, characterized in that: It includes: Based on the performance parameters of the target model radar, a high-resolution model of the target model radar is constructed; Establish a radar simulation model, which includes a signal-to-noise ratio model, a detection distance model, a detection probability model, and a tracking accuracy model; The signal-to-noise ratio model SNR of the radar simulation model is: Where P t : Radar transmitting power; G t : The transmitting antenna gain in this direction; G r : The receiving antenna gain in this direction; λ: echo signal wavelength; R: The distance between radar and target detection; k: Boltzmann constant; T0: radar system temperature; B: receiver instantaneous bandwidth; L: Radar RF front-end loss; N f : Noise coefficient, SNR: radar signal-to-noise ratio; The detection distance model R of the radar simulation model is: Where P t : Transmit power; G: antenna gain in this direction; λ: echo signal wavelength; σ: radar cross section, k: Boltzmann constant; B: receiver instantaneous bandwidth; L: Radar RF front-end loss; SNR: radar signal-to-noise ratio; T0: radar system temperature; The detection probability model P of the radar simulation model d for: Where erfc() is the complementary error function, P f : false alarm probability; The tracking accuracy model of the radar simulation model is: In the formula, S j : signal-to-noise ratio of a single pulse; c: speed of light; τ0: pulse width; n: number of pulse trains; Assume that the signal-to-noise ratio data set of the signal-to-noise ratio model {T i } and the signal-to-noise ratio dataset of the high-resolution model {Q i }, calculate the signal-to-noise ratio standard deviation S of the signal-to-noise ratio model respectively. T and the signal-to-noise ratio standard deviation S of the high-resolution model Q ; Then establish the signal-to-noise ratio error function Δ=|S Q -S T |; Let the control factor be θ i , the control factor function is g(θ i )=a i θ i , i=1,2,...,n, where a i is the control factor θ i The weight of all weights is 1, and each weight a i is the average of all weight sums; The objective function is established with the minimum signal-to-noise ratio error function Δ as the goal, and the constraint condition is the standard deviation S of the signal-to-noise ratio in the signal-to-noise ratio model. T , the standard deviation of the signal-to-noise ratio of the high-resolution model S Q , set the control factor θ i The range of values; Based on the above conditions, g(θ i ) is multiplied by the noise ratio model to establish an optimization model and solve the control factor θ i The value of is multiplied by the signal-to-noise ratio model to calculate the signal-to-noise ratio value for verifying the radar simulation model; The verification process is that if the signal-to-noise ratio value meets the error requirement range, the detection distance, detection probability, and tracking accuracy of the radar simulation model are calculated based on the signal-to-noise ratio value and the detection distance model, detection probability model, and tracking accuracy model of the radar simulation model; if the detection distance, detection probability, and tracking accuracy meet the error requirement range, the radar simulation model is verified to be qualified; Otherwise, the radar simulation model is verified to be unqualified, then the sum of the weights of the control factors is 1 and each weight a i It is a constraint condition of the average value of all weights, continues to introduce new control factors and retains the values ​​of the original control factors, continues to solve the values ​​of the new control factors based on the optimization model, and continues the verification until the verification is successful.

2. The radar simulation model verification method according to claim 1, characterized in that: When the control factor is θ1, the control factor function is g(θ1)=a1θ1, a1=1, and the analytical form of the optimization model is: minΔ=|S T -S Q | Among them, T i is the signal-to-noise ratio data of the signal-to-noise ratio model, T is the mean of the signal-to-noise ratio data set {Ti} of the signal-to-noise ratio model, Qi is the signal-to-noise ratio data of the high-resolution model, Q is the mean of the signal-to-noise ratio data set {Qi} of the high-resolution model, the subscript i is the number of the data, N is a natural number, P t is the radar transmission power, G t is the transmitting antenna gain in this direction, G r is the receiving antenna gain in this direction, λ is the echo signal wavelength, k is the Boltzmann constant, R is the distance between the radar and the target, T0 is the radar system temperature, B is the instantaneous bandwidth of the radar receiver, L is the radar RF front-end loss, N f is the noise figure, SNR θ is the radar signal-to-noise ratio corresponding to the control factor θ1, and the value range of the control factor θ1 is [c1, c2], [c1, c2] depends on the specific problem; The value b1 of θ1 is determined by analytical calculation of the optimization model.

3. The radar simulation model verification method according to claim 2, characterized in that: If the value b1 of θ1 cannot make the signal-to-noise ratio, detection distance, detection probability and tracking accuracy of the radar simulation model meet the requirements, then the value b1 of θ1 is used to transform the control factor function to g(θ1, θ2)=0.5b1+0.5θ2, and the new control factor θ2 is continued to be calculated based on the optimization model, and the radar simulation model is verified using the control factor θ2. If the radar simulation model is successfully verified, the verification ends; otherwise, new control factors are continued to be introduced, and the radar simulation model is verified using the control factors until the radar simulation model is successfully verified.

4. A radar simulation model verification system, characterized in that: It includes: A high-resolution model building module builds a high-resolution model of the target model radar based on the performance parameters of the target model radar; The radar simulation model building module is used to build a radar simulation model. The radar simulation model includes a signal-to-noise ratio model, a detection distance model, a detection probability model, and a tracking accuracy model. The signal-to-noise ratio model SNR of the radar simulation model is: Where P t : Radar transmitting power; G t : The transmitting antenna gain in this direction; G r : The receiving antenna gain in this direction; λ: echo signal wavelength; R: The distance between radar and target detection; k: Boltzmann constant; T0: radar system temperature; B: receiver instantaneous bandwidth; L: Radar RF front-end loss; N f : Noise coefficient, SNR: radar signal-to-noise ratio; The detection distance model R of the radar simulation model is: Where P t : Transmit power; G: antenna gain in this direction; λ: echo signal wavelength; σ: radar cross section, k: Boltzmann constant; B: receiver instantaneous bandwidth; L: Radar RF front-end loss; SNR: radar signal-to-noise ratio; T0: radar system temperature; The detection probability model P of the radar simulation model d for: Where erfc() is the complementary error function, P f : false alarm probability; The tracking accuracy model of the radar simulation model is: In the formula, S j : signal-to-noise ratio of a single pulse; c: speed of light; τ0: pulse width; n: number of pulse trains; The simulation model verification module performs the following processing: assuming that the signal-to-noise ratio data set of the signal-to-noise ratio model {T i } and the signal-to-noise ratio dataset of the high-resolution model {Q i }, calculate the signal-to-noise ratio standard deviation S of the signal-to-noise ratio model respectively. T and the signal-to-noise ratio standard deviation S of the high-resolution model Q ; Then establish the signal-to-noise ratio error function Δ=|S Q -S T |; Let the control factor be θ i , the control factor function is g(θ i )=a i θ i , i=1,2,...,n, where a i is the control factor θ i The weight of all weights is 1, and each weight a i is the average of all weight sums; The objective function is established with the minimum signal-to-noise ratio error function Δ as the goal, and the constraint condition is the standard deviation S of the signal-to-noise ratio in the signal-to-noise ratio model. T , the standard deviation of the signal-to-noise ratio of the high-resolution model S Q , set the control factor θ i The range of values; Based on the above conditions, g(θ i ) is multiplied by the noise ratio model to establish an optimization model and solve the control factor θ i The value of is multiplied by the signal-to-noise ratio model to calculate the signal-to-noise ratio value for verifying the radar simulation model; The verification process is that if the signal-to-noise ratio value meets the error requirement range, the detection distance, detection probability, and tracking accuracy of the radar simulation model are calculated based on the signal-to-noise ratio value and the detection distance model, detection probability model, and tracking accuracy model of the radar simulation model; if the detection distance, detection probability, and tracking accuracy meet the error requirement range, the radar simulation model is verified to be qualified; Otherwise, the radar simulation model is verified to be unqualified, then the sum of the weights of the control factors is 1 and each weight a i It is a constraint condition of the average value of all weights, continues to introduce new control factors and retains the values ​​of the original control factors, continues to solve the values ​​of the new control factors based on the optimization model, and continues the verification until the verification is successful.

5. The radar simulation model verification system according to claim 4, characterized in that: The simulation model verification module further performs the following processing: when the control factor is θ1, the control factor function is g(θ1)=a1θ1, a1=1, and the analytical form of the optimization model is: minΔ=|S T -S Q | Among them, T i is the signal-to-noise ratio data of the signal-to-noise ratio model, T is the signal-to-noise ratio data set of the signal-to-noise ratio model {T i }, Q i is the signal-to-noise ratio data of the high-resolution model, Q is the signal-to-noise ratio dataset of the high-resolution model {Q i }, the subscript i is the number of the data, N is a natural number, P t is the radar transmission power, G t is the transmitting antenna gain in this direction, G r is the receiving antenna gain in this direction, λ is the echo signal wavelength, k is the Boltzmann constant, R is the distance between the radar and the target, T0 is the radar system temperature, B is the instantaneous bandwidth of the radar receiver, L is the radar RF front-end loss, N f is the noise figure, SNR θ is the radar signal-to-noise ratio corresponding to the control factor θ1, and the value range of the control factor θ1 is [c1, c2], [c1, c2] depends on the specific problem; The value b1 of θ1 is determined by analytical calculation of the optimization model.

6. The radar simulation model verification system according to claim 5, characterized in that: The simulation model verification module further performs the following processing: if the value b1 of θ1 cannot make the signal-to-noise ratio, detection distance, detection probability, and tracking accuracy of the radar simulation model meet the requirements, then the value b1 of θ1 is used to transform the control factor function to g(θ1, θ2) = 0.5b1 + 0.5θ2, and the new control factor θ2 is continued to be calculated based on the optimization model, and the radar simulation model is verified using the control factor θ2. If the radar simulation model is successfully verified, the verification ends; otherwise, new control factors are continued to be introduced, and the radar simulation model is verified using the control factors until the radar simulation model is successfully verified.

Images