Multifunctional laser weapon simulation training terminal correction method

Abstract

The invention relates to the technical field of laser weapon simulation training, in particular to a multifunctional laser weapon simulation training terminal correction method which is used for improving the performance and accuracy of a laser weapon in simulation training. Comprising the following steps: comprehensively monitoring environmental factors influencing the performance of the laser weapon; integrating multi-source data through an unscented Kalman filtering algorithm to form consistent environment feature description; based on the obtained environment feature description, a physical model and a data driving model are constructed respectively, and a comprehensive environment influence model is formed through weighted fusion; calculating correction parameters based on the obtained environmental influence model, and applying the calculated correction parameters to the laser weapon simulation training terminal in real time to realize dynamic correction of the laser weapon performance; and evaluating the effect of the dynamic correction, and establishing a feedback loop to optimize a correction strategy. The performance and the accuracy of the multifunctional laser weapon simulation training terminal are improved, and the authenticity, the effectiveness and the reliability of simulation training are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of laser weapon simulation training, and in particular to a multifunctional laser weapon simulation training terminal correction method for improving the performance and accuracy of laser weapons in simulation training. Background Art

[0002] With the continuous development of laser weapon technology, simulation training of laser weapons has become increasingly important. In laser weapon simulation training, environmental factors have a significant impact on the performance of laser weapons, such as temperature, humidity, air pressure, atmospheric turbulence, etc. In order to improve the authenticity and effectiveness of simulation training, it is necessary to calibrate the laser weapon simulation training terminal to compensate for the impact of environmental factors on the performance of laser weapons.

[0003] At present, some calibration methods for laser weapon simulation training terminals have problems such as low accuracy, poor real-time performance, and weak adaptability, which cannot meet the needs of practical applications. Therefore, a more accurate, real-time, and effective calibration method for laser weapon simulation training terminals is needed. Summary of the invention

[0004] In order to achieve the above objectives, the present invention adopts the following technical solutions:

[0005] The purpose of the present invention is to provide a multifunctional laser weapon simulation training terminal calibration method, which effectively improves the performance and accuracy of the multifunctional laser weapon simulation training terminal and enhances the authenticity, effectiveness and reliability of the simulation training.

[0006] The multifunctional laser weapon simulation training terminal calibration method comprises the following steps:

[0007] S1. Deploy sensors in the training environment to comprehensively monitor environmental factors that affect the performance of laser weapons;

[0008] S2. Data extraction and fusion, using the sensor data collected by S1, integrates multi-source data through the unscented Kalman filter algorithm to form a consistent description of environmental characteristics and provide input data for model construction;

[0009] S3. Model construction and parameter calculation: Based on the environmental characteristics description obtained in S2, a physical model and a data-driven model are constructed respectively, and a comprehensive environmental impact model is formed through weighted fusion;

[0010] S4. Calculate correction parameters based on the environmental impact model obtained in S3, and apply the calculated correction parameters to the laser weapon simulation training terminal in real time to achieve dynamic correction of laser weapon performance;

[0011] S5. Evaluate the effectiveness of the dynamic correction in S4 and establish a feedback loop to optimize the correction strategy.

[0012] Further, in S2, the unscented Kalman filter algorithm comprises the following steps:

[0013] S21. Initialization, determine the initial state of the system and covariance matrix P0;

[0014] Assume the initial state is: Among them, u0 is the mean vector of sensor data, covariance matrix P0 = Σ0, Σ0 is the covariance matrix of sensor data;

[0015] S22. Prediction step: predicting the state and covariance matrix at the next moment according to the dynamic model of the system;

[0016] Assume the dynamic model of the system is:

[0017] f represents the state transfer function, v k-1 is the process noise;

[0018] The prediction formula of the covariance matrix is:

[0019] Fk-1 is the state transfer matrix (unit matrix), Qk-1 is the process noise v k-1 The covariance matrix of

[0020] S23. Update step, combining the sensor's measurement value, updates the prediction result to obtain a more accurate state estimate and covariance matrix:

[0021] Assume the sensor's measurement value is: Among them, ω k is the measurement noise; the measurement matrix

[0022] Kalman gain:

[0023] The update formula is:

[0024] P k =(IK k H k ) kk-1 ;

[0025] It is the predicted state at time k based on the information at time k-1; is a measurement function that will predict the state Mapped to the measurement space, in the laser weapon system, if the predicted state includes the internal parameters of the laser weapon and environmental factors, then this measurement function can calculate the corresponding measurable performance indicators based on these parameters;

[0026] It represents the correction amount of the difference between the measured value and the predicted measured value after the Kalman gain adjustment. If the difference between the measured value and the predicted measured value is large, then the correction amount will be large, and the state estimation will be adjusted more toward the measured value; if the difference is small, the correction amount will be small, and the state estimation will be kept closer to the predicted state. k represents the actual observation value obtained by the sensor measuring the system state at time k. For the laser weapon simulation training system, z k It is the measured value of certain performance indicators of laser weapons, such as laser energy intensity, range, accuracy and other specific physical quantities; R k is the measurement noise ω k The covariance matrix of the measurement noise, which represents the uncertainty of the measurement noise; I is the identity matrix; K k H k represents the product of Kalman gain and measurement matrix, which is used to adjust the prediction covariance matrix; P kk-1 is the prediction covariance matrix obtained at time based on the information at time;

[0027] S24. Weighted fusion: Determine the weight of each sensor based on factors such as the accuracy, reliability, and timeliness of the sensor, perform weighted fusion on multi-source data, and form a consistent description of environmental characteristics.

[0028] Furthermore, in S24, the weight calculation formula is:

[0029]

[0030] is the measurement error of sensor i; z i is the measurement value of sensor i; is the mean value of the sensor i; e i is the reliability index, e i =1-f i , f i Historical failure rate of sensor i; t j is the time lag of the data; represents the number of sensors;

[0031] The fused environmental characteristics are described as follows:

[0032] Furthermore, in S3, the physical model is established based on the physical principles of the laser weapon and the influence mechanism of environmental factors. The physical model should be able to accurately describe the performance changes of the laser weapon under different environmental conditions; the physical model is expressed as:

[0033]

[0034] P ph The power is predicted for the physical model of the laser weapon; T is the ambient temperature; H is the ambient humidity; P is the ambient air pressure; Turb is the atmospheric turbulence sensor; k1 is a parameter related to the square of the temperature; k2 is a parameter related to the cube of the humidity; k3 is a parameter related to the air pressure; k4 is a parameter related to the square root of the atmospheric turbulence; k5 is a parameter related to the product of temperature and humidity; k6 is a parameter related to the product of air pressure and humidity; k7 is a parameter related to the product of air pressure and atmospheric turbulence.

[0035] Furthermore, in S3, the data-driven model is trained using a large amount of monitoring data:

[0036] Assume that the input environment data is E data = (T, H, P, Turb), which is usually the original sensor data, including the specific values ​​of various environmental factors such as temperature, humidity, air pressure, atmospheric turbulence, etc. After training with a neural network with l layers, the output of the jth layer is y i =σ(W j y j-1 +b j ), where W j and b j is the weight and bias of the jth layer, σ is the activation function, σ(x) = max(0, x), and the laser weapon performance prediction output by the data-driven model is P data =y l =σ(W l y k-1 +b l ); During the training process, the weights and biases are continuously adjusted to make the output of the neural network as close as possible to the actual laser weapon performance data.

[0037] Furthermore, in S3, the outputs of the physical model and the data-driven model are integrated to form a comprehensive environmental impact model:

[0038] Assume that the weight of the physical model is α and the weight of the data-driven model is β, then:

[0039] M com =αP ph +βP data ;

[0040] Among them, Mcom is the output of the environmental impact model; α+β=1, α and β are determined by minimizing the mean square error, P ture For the real laser weapon performance, P data The final output of the data-driven model is the laser weapon performance prediction, P ph The physical model of the laser weapon predicts the power; then the mean square error is:

[0041]

[0042] Where n is the number of samples, P ph,i The power predicted by the laser weapon physical model for the i-th sample; P date,i is the laser weapon performance prediction output by the data-driven model of the i-th sample; P ture,i is the actual laser weapon performance of the sample;

[0043] By taking the partial derivatives of MSE with respect to α and β and setting them to zero, we can obtain the following system of equations:

[0044]

[0045] By solving the system of equations using matrix operations, we can obtain the optimal α and β:

[0046] make

[0047] Then the system of equations is expressed as: Solve this system of equations using the matrix inversion method to obtain the values ​​of α and β:

[0048] Further, in S4, the correction parameter is calculated by the environmental impact model, and the correction parameter δ=h(M com ), where h is a linear function, M com is the output of the environmental impact model, h(x) = kx + d; then the correction parameter δ = k(αP ph +βP date )+d, where k and d are parameters determined according to actual conditions;

[0049] k and d are used to adjust the performance of laser weapons so that they can maintain the best state under different environmental conditions. k represents the slope of the linear function, which determines the degree of influence of the output of the environmental impact model on the correction parameters. Specifically, the larger the value of k, the greater the influence of the change in the output of the environmental impact model on the correction parameters; conversely, the smaller the value of k, the smaller the influence of the change in the output of the environmental impact model on the correction parameters.

[0050] d represents the intercept of the linear function, which is a constant term used to adjust the baseline value of the correction parameter; even if the output of the environmental impact model is zero, the correction parameter will have a basic value d.

[0051] Furthermore, in step S4, the calculated correction parameters are applied in real time to the laser weapon simulation training terminal to achieve a method for dynamic correction of laser weapon performance:

[0052] S41. Parameter transmission: The calculated correction parameters are transmitted to the laser weapon simulation training terminal in real time through a high-speed data transmission channel; the transmission channel should have low latency and high bandwidth to ensure that the parameters can be applied to the terminal device in a timely manner;

[0053] S42. Dynamic correction: In the terminal device, a special correction module is set up, which can quickly receive and execute correction parameters to achieve dynamic correction of laser weapon performance;

[0054] Assume the corrected laser weapon performance is P co =P or +δ, where P or is the performance of the laser weapon before correction, and δ is the correction parameter; the correction module should have high real-time performance and accuracy to ensure that the performance of the laser weapon can respond to environmental changes in a timely manner; for example, when environmental factors change, the correction module can quickly receive new correction parameters and adjust the relevant parameters of the laser weapon to keep the performance of the laser weapon at the best state; if the performance indicator of the laser weapon is energy output, the unit is joule, then the energy output after correction is P co =P or +δ Joule; in this way, real-time dynamic adjustment of laser weapon performance can be achieved, improving the authenticity and effectiveness of simulation training.

[0055] Further, the S51 includes:

[0056] S51. Effect evaluation: establish an evaluation index system to evaluate the effect of correction execution; the comprehensive evaluation index is expressed as:

[0057] E ev =k1ΔI r +k2ΔI a +k3ΔI e ;

[0058] Among them, k1, k2, k3 are weight coefficients, and satisfy k1+k2+k3=1; ΔI r is the change in range; ΔI a is the change in accuracy; ΔI e is the change in energy output;

[0059] S52. Establish a feedback loop to optimize the correction strategy. According to the results of the evaluation in S51, establish a feedback loop to optimize the correction strategy;

[0060] Assume that the sensor's measurement range is R and the original sensor resolution is R es , adjust the measurement range and resolution according to the change in the evaluation index; the adjustment formula of the sensor parameters after adjustment is:

[0061] R n =R+λΔI r ;

[0062]

[0063] Among them, R n To adjust the sensor measurement range; is the adjusted sensor resolution; λ is the measurement range adjustment amount determined according to the evaluation results. If the range changes greatly and the measurement range needs to be expanded, λ takes a larger positive value; if the range increases greatly and the measurement range needs to be appropriately reduced, λ takes a smaller negative value; η is the resolution adjustment amount determined according to the evaluation results. If the accuracy decreases greatly and the resolution needs to be increased, η takes a larger positive value; if the accuracy increases greatly and the resolution needs to be appropriately reduced to balance other performance, η takes a smaller negative value; ΔI r is the range change; ΔI a is the change in accuracy.

[0064] In summary, due to the adoption of the above technical solution, the beneficial technical effects of the invention are:

[0065] Improve the accuracy of laser weapon performance by deploying sensors to comprehensively monitor environmental factors that affect laser weapon performance, and use the unscented Kalman filter algorithm to integrate multi-source data to form a consistent description of environmental characteristics and provide accurate input data for model construction. The physical model and data-driven model built on this basis can more accurately describe the performance changes of laser weapons under different environmental conditions, thereby calculating more accurate correction parameters and improving the accuracy of laser weapon performance.

[0066] The laser weapon performance is dynamically corrected, and the calculated correction parameters are applied to the laser weapon simulation training terminal in real time. Through parameter transmission and dynamic correction steps, the correction module is ensured to quickly receive and execute the correction parameters, and the real-time dynamic adjustment of the laser weapon performance is realized. When environmental factors change, the correction module can respond quickly, so that the performance of the laser weapon is always kept in the best state, which improves the authenticity and effectiveness of the simulation training.

[0067] Optimize the correction strategy, evaluate the effect of dynamic correction, establish a feedback loop, adjust the sensor's measurement range and resolution based on the evaluation results, optimize the weight calculation method of the fusion algorithm, and improve the model's structure and parameters. This can continuously improve the system's performance and adaptability to better meet the needs of actual applications.

[0068] Taking into account the influence of various environmental factors, the physical model takes into account multiple environmental factors such as the influence of temperature on the refractive index of the laser medium, the influence of humidity on laser energy absorption, the influence of air pressure on laser propagation speed, and the influence of atmospheric turbulence on laser beam quality. It can comprehensively and accurately describe the performance changes of laser weapons under different environmental conditions.

[0069] Improve the reliability of simulation training, use a large amount of monitoring data for training through data-driven models, and make the output of the neural network as close as possible to the actual laser weapon performance data. At the same time, through weighted fusion to form a comprehensive environmental impact model, it can more reliably predict the performance of laser weapons and improve the reliability of simulation training.

[0070] Enhance the stability and real-time performance of the system. The parameter transmission channel has the characteristics of low latency and high bandwidth, ensuring that the parameters can be applied to the terminal device in a timely manner. The correction module has high real-time and accuracy, and can respond to environmental changes in a timely manner, enhancing the stability and real-time performance of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 A flow chart of the calibration method for the multifunctional laser weapon simulation training terminal;

[0072] Figure 2 It is the flow chart of the unscented Kalman filter algorithm;

[0073] Figure 3 A flow chart of a method for realizing dynamic correction of laser weapon performance;

[0074] Figure 4 Flowchart of the method for establishing feedback loops to optimize correction strategies; DETAILED DESCRIPTION

[0075] In order to make the purpose, technical solution and advantages of the invention more clear, the invention is further described in detail below in combination with the embodiments. It should be understood that the specific embodiments described here are only used to explain the invention and are not used to limit the invention.

[0076] like Figure 1 As shown, the multifunctional laser weapon simulation training terminal calibration method includes the following steps:

[0077] S1. Deploy sensors in the training environment to comprehensively monitor environmental factors that affect the performance of laser weapons;

[0078] S2. Data extraction and fusion, using the sensor data collected by S1, integrates multi-source data through the unscented Kalman filter algorithm to form a consistent description of environmental characteristics and provide input data for model construction;

[0079] S3. Model construction and parameter calculation: Based on the environmental characteristics description obtained in S2, a physical model and a data-driven model are constructed respectively, and a comprehensive environmental impact model is formed through weighted fusion;

[0080] S4. Calculate correction parameters based on the environmental impact model obtained in S3, and apply the calculated correction parameters to the laser weapon simulation training terminal in real time to achieve dynamic correction of laser weapon performance;

[0081] S5. Evaluate the effectiveness of the dynamic correction in S4 and establish a feedback loop to optimize the correction strategy.

[0082] like Figure 2 As shown, in S2, the unscented Kalman filter algorithm includes the following steps:

[0083] S21. Initialization, determine the initial state of the system and covariance matrix P0;

[0084] Assume the initial state is: Among them, u0 is the mean vector of sensor data, covariance matrix P0 = Σ0, Σ0 is the covariance matrix of sensor data;

[0085] S22. Prediction step: predicting the state and covariance matrix at the next moment according to the dynamic model of the system;

[0086] Assume the dynamic model of the system is:

[0087] f represents the state transfer function, v k-1 is the process noise;

[0088] The prediction formula of the covariance matrix is:

[0089] Fk-1 is the state transfer matrix (unit matrix), Qk-1 is the process noise v k-1 The covariance matrix of

[0090] S23. Update step, combining the sensor's measurement value, updates the prediction result to obtain a more accurate state estimate and covariance matrix:

[0091] Assume the sensor's measurement value is: Among them, ωk is the measurement noise; the measurement matrix

[0092] Kalman gain:

[0093] The update formula is:

[0094] P k =(IK k H k ) kk-1 ;

[0095] It is the predicted state at time k based on the information at time k-1; is a measurement function that will predict the state Mapped to the measurement space, in the laser weapon system, if the predicted state includes the internal parameters of the laser weapon and environmental factors, then this measurement function can calculate the corresponding measurable performance indicators based on these parameters;

[0096] It represents the correction amount of the difference between the measured value and the predicted measured value after the Kalman gain adjustment. If the difference between the measured value and the predicted measured value is large, then the correction amount will be large, and the state estimation will be adjusted more toward the measured value; if the difference is small, the correction amount will be small, and the state estimation will be kept closer to the predicted state. k represents the actual observation value obtained by the sensor measuring the system state at time k. For the laser weapon simulation training system, z k It is the measured value of certain performance indicators of laser weapons, such as laser energy intensity, range, accuracy and other specific physical quantities; R k is the measurement noise ω k The covariance matrix of the measurement noise, which represents the uncertainty of the measurement noise; I is the identity matrix; K k H k represents the product of Kalman gain and measurement matrix, which is used to adjust the prediction covariance matrix; P kk-1 is the prediction covariance matrix obtained at time based on the information at time;

[0097] S24. Weighted fusion: Determine the weight of each sensor based on factors such as the accuracy, reliability, and timeliness of the sensor, perform weighted fusion on multi-source data, and form a consistent description of environmental characteristics.

[0098] In the S24,

[0099] The weight calculation formula is:

[0100] is the measurement error of sensor i; is the mean value of the sensor i; e i is the reliability index, e i =1-f i , f i Historical failure rate of sensor i; t i is the time lag of the data; m represents the number of sensors;

[0101] The fused environmental characteristics are described as follows:

[0102] S3. Model construction and parameter calculation: Based on the environmental characteristics description obtained in S2, a physical model and a data-driven model are constructed respectively, and a comprehensive environmental impact model is formed through weighted fusion;

[0103] In S3, the physical model is a model established based on the physical principles of laser weapons and the influence mechanism of environmental factors. The model should be able to accurately describe the performance changes of laser weapons under different environmental conditions; for example, considering the influence of temperature on the refractive index of the laser medium, the influence of humidity on laser energy absorption, the influence of air pressure on laser propagation speed, and the influence of atmospheric turbulence on laser beam quality. The physical model is expressed as:

[0104]

[0105] P ph The power is predicted for the physical model of the laser weapon; T is the ambient temperature; H is the ambient humidity; P is the ambient air pressure; Turb is the atmospheric turbulence sensor; k1 is a parameter related to the square of the temperature, which determines the degree of influence of the square of the temperature on the performance of the laser weapon; k2 is a parameter related to the cube of the humidity, which is used to describe the influence of the cube of the humidity on the performance of the laser weapon; k3 is a parameter related to the air pressure, which reflects the direct influence of the air pressure on the performance of the laser weapon; k4 is a parameter related to the square root of the atmospheric turbulence, which indicates that the influence of the square root of the atmospheric turbulence on the performance of the laser weapon is determined by; k5 is a parameter related to the product of temperature and humidity, which reflects the influence of the interaction between temperature and humidity on the performance of the laser weapon; k6 is a parameter related to the product of air pressure and humidity, which is used to describe the influence of the combined effect of air pressure and humidity on the performance of the laser weapon; k7 is a parameter related to the product of air pressure and atmospheric turbulence, which reflects the influence of the interaction between air pressure and atmospheric turbulence on the performance of the laser weapon.

[0106] In S3, the data-driven model is trained using a large amount of monitoring data;

[0107] Assume that the input environment data is E data= (T, H, P, Turb), which is usually the original sensor data, including the specific values ​​of various environmental factors such as temperature, humidity, air pressure, atmospheric turbulence, etc. After training with a neural network with l layers, the output of the jth layer is y i =σ(W j y j-1 +b j ), where W j and b j is the weight and bias of the jth layer, σ is the activation function, σ(x) = max(0, x), and the laser weapon performance prediction output by the data-driven model is P data =y l =σ(W l y k-1 +b l ); During the training process, the weights and biases are continuously adjusted to make the output of the neural network as close as possible to the actual laser weapon performance data.

[0108] In S3, the outputs of the physical model and the data-driven model are integrated to form a comprehensive environmental impact model;

[0109] Assume that the weight of the physical model is α and the weight of the data-driven model is β, then:

[0110] M com =αP ph +βP data ;

[0111] Among them, α+β=1, α and β are determined by minimizing the mean square error, P ture For the real laser weapon performance, P data The final output of the data-driven model is the laser weapon performance prediction, P ph The physical model of the laser weapon predicts the power; then the mean square error is:

[0112]

[0113] Where n is the number of samples, P ph,i The power predicted by the laser weapon physical model for the i-th sample; P date,i is the laser weapon performance prediction output by the data-driven model of the i-th sample; P ture,i is the actual laser weapon performance of the sample;

[0114] By taking the partial derivatives of MSE with respect to α and β and setting them to zero, we can obtain the following system of equations:

[0115]

[0116] By solving the system of equations using matrix operations, we can obtain the optimal α and β:

[0117] make

[0118] Then the system of equations is expressed as: Solve this system of equations using the matrix inversion method to obtain the values ​​of α and β:

[0119] S4. Calculate correction parameters based on the environmental impact model obtained in S3, and apply the calculated correction parameters to the laser weapon simulation training terminal in real time to achieve dynamic correction of laser weapon performance;

[0120] In S4, the correction parameter is calculated by the environmental impact model, that is, δ = h(M com ), where h is a linear function, h(x) = kx + d, then the correction parameter δ = k(αP ph +βP date )+d, where k and d are parameters determined according to actual conditions;

[0121] k and d are used to adjust the performance of laser weapons so that they can maintain the best state under different environmental conditions. k represents the slope of the linear function, which determines the degree of influence of the output of the environmental impact model on the correction parameters. Specifically, the larger the value of k, the greater the influence of the change in the output of the environmental impact model on the correction parameters; conversely, the smaller the value of k, the smaller the influence of the change in the output of the environmental impact model on the correction parameters.

[0122] d represents the intercept of the linear function, which is a constant term used to adjust the baseline value of the correction parameter; even if the output of the environmental impact model is zero, the correction parameter will have a basic value d.

[0123] like Figure 3 As shown in S4, the dynamic correction method is:

[0124] S41. Parameter transmission: The calculated correction parameters are transmitted to the laser weapon simulation training terminal in real time through a high-speed data transmission channel; the transmission channel should have low latency and high bandwidth to ensure that the parameters can be applied to the terminal device in a timely manner;

[0125] S42. Dynamic correction: In the terminal device, a special correction module is set up, which can quickly receive and execute correction parameters to achieve dynamic correction of laser weapon performance;

[0126] Assume the corrected laser weapon performance is P co =P or +δ, where P oris the performance of the laser weapon before correction, and δ is the correction parameter; the correction module should have high real-time performance and accuracy to ensure that the performance of the laser weapon can respond to environmental changes in a timely manner; for example, when environmental factors change, the correction module can quickly receive new correction parameters and adjust the relevant parameters of the laser weapon to keep the performance of the laser weapon at the best state; if the performance indicator of the laser weapon is energy output, the unit is joule, then the energy output after correction is P co =P or +δ Joule; in this way, real-time dynamic adjustment of laser weapon performance can be achieved, improving the authenticity and effectiveness of simulation training.

[0127] S5. Evaluate the effectiveness of the dynamic correction in S4 and establish a feedback loop to optimize the correction strategy;

[0128] like Figure 4 As shown, the S5 includes:

[0129] S51. Effect evaluation: Establish an evaluation index system to evaluate the effect of correction execution; the evaluation index includes performance indicators such as the range, accuracy, and energy output of laser weapons, as well as the real-time, accuracy, and stability of the correction strategy; the comprehensive evaluation index is expressed as:

[0130] E ev =k1ΔI r +k2ΔI a +k3ΔI e ;

[0131] Among them, k1, k2, k3 are weight coefficients, and satisfy k1+k2+k3=1; ΔI r It is the change in range, indicating the degree of change in the range of the laser weapon after correction relative to that before correction; ΔI a It is the change in accuracy, reflecting the impact of correction on the accuracy performance of laser weapons, expressed in percentage; ΔI e It is the change in energy output, expressed as the energy output after correction minus the energy output before correction, in joules, which reflects the effect of correction on the energy output performance of laser weapons;

[0132] S52. Establish a feedback loop to optimize the correction strategy. According to the results of the evaluation in S51, establish a feedback loop to optimize the correction strategy;

[0133] Assume that the sensor's measurement range is R and its resolution is R es , adjust the measurement range and resolution according to the change in the evaluation index; for example, if the change in range is large, the measurement range should be appropriately expanded; if the change in accuracy is large, the resolution needs to be increased; the adjustment formula of the sensor parameters after adjustment is:

[0134] R n =R+λΔI r ;

[0135]

[0136] Among them, R n To adjust the sensor measurement range; is the adjusted sensor resolution; λ is the measurement range adjustment amount determined according to the evaluation results. If the range changes greatly and the measurement range needs to be expanded, λ takes a larger positive value; if the range increases greatly and the measurement range needs to be appropriately reduced, λ takes a smaller negative value; η is the resolution adjustment amount determined according to the evaluation results. If the accuracy decreases greatly and the resolution needs to be increased, η takes a larger positive value; if the accuracy increases greatly and the resolution needs to be appropriately reduced to balance other performance, η takes a smaller negative value; ΔI r is the range change; ΔI a is the change in accuracy.

[0137] The above description is a preferred embodiment of the invention and is not intended to limit the invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the invention should be included in the protection scope of the invention.

Claims

1. A multifunctional laser weapon simulation training terminal calibration method, characterized in that: include: S1. Deploy sensors in the training environment to comprehensively monitor environmental factors that affect the performance of laser weapons; S2. Using the sensor data collected by S1, the multi-source data is integrated through the unscented Kalman filter algorithm to form a consistent description of the environmental characteristics; S3. Based on the environmental feature description obtained in S2, a physical model and a data-driven model are constructed respectively, and a comprehensive environmental impact model is formed through weighted fusion; S4. Calculate correction parameters based on the environmental impact model obtained in S3, and apply the calculated correction parameters to the laser weapon simulation training terminal in real time to achieve dynamic correction of laser weapon performance; S5. Evaluate the effectiveness of the dynamic correction in S4 and establish a feedback loop to optimize the correction strategy.

2. According to the multifunctional laser weapon simulation training terminal calibration method of claim 1, said S2 comprises the following steps: S21. Initialization, determine the initial state of the system and covariance matrix P0; Assume the initial state is: Among them, u0 is the mean vector of sensor data, covariance matrix P0 = Σ0, Σ0 is the covariance matrix of sensor data; S22. Prediction step: predicting the state and covariance matrix at the next moment according to the dynamic model of the system; Assume the dynamic model of the system is: f represents the state transfer function, v k-1 is the process noise; The prediction formula of the covariance matrix is: F k-1 is the state transfer matrix, Q k-1 is the process noise v k-1 The covariance matrix of S23. Update step, combining the sensor's measurement value, updates the prediction result to obtain a more accurate state estimate and covariance matrix: Assume the sensor's measurement value is: Among them, ω k is the measurement noise; the measurement matrix Kalman gain: The update formula is: P k =(I-K k H k )P k|k-1 ; It is the predicted state at time k based on the information at time k-1; is a measurement function that will predict the state Mapping to measurement space; It represents the correction amount of the difference between the measured value and the predicted measured value after the Kalman gain adjustment; k represents the actual observation value obtained by the sensor measuring the system state at time k. For the laser weapon simulation training system, z k is the measured value of some performance indicators of laser weapons; R k is the measurement noise ω k The covariance matrix of k H k represents the product of Kalman gain and measurement matrix; P kk-1 is the prediction covariance matrix obtained at time based on the information at time; S24. Weighted fusion: Calculate the weight of each sensor based on the accuracy, reliability and timeliness of the sensor, perform weighted fusion on multi-source data, and form a consistent description of environmental features; Environmental characteristics description: Among them, z i is the measurement value of sensor i; ω i is the weight; The weight calculation formula is: in, is the measurement error of sensor i; z i is the measurement value of sensor i; is the mean value of the sensor i; e i is the reliability index, e i =1-f i , f i Historical failure rate of sensor i; t j is the time lag of the data; represents the number of sensors.

3. The multifunctional laser weapon simulation training terminal calibration method according to claim 1 is characterized in that: In S3, the physical model is expressed as: P ph The power is predicted for the physical model of the laser weapon; T is the ambient temperature; H is the ambient humidity; P is the ambient air pressure; Turb is the atmospheric turbulence sensor; k1 is a parameter related to the square of the temperature; k2 is a parameter related to the cube of the humidity; k3 is a parameter related to the air pressure; k4 is a parameter related to the square root of the atmospheric turbulence; k5 is a parameter related to the product of temperature and humidity; k6 is a parameter related to the product of air pressure and humidity; k7 is a parameter related to the product of air pressure and atmospheric turbulence.

4. The multifunctional laser weapon simulation training terminal calibration method according to claim 1 is characterized in that: The S3 data-driven model: P data =y l =σ(W l y l-1 +b l ): Assume that the input environment data is E data =(T,H,P,Turb), trained by a neural network with l layers, the output of the lth layer is y l =σ(W l y l-1 +b l ), where W l and b l is the weight and bias of the lth layer, σ is the activation function, σ(x) = max(0,x), and the laser weapon performance prediction output by the data-driven model is P data =y l =σ(W l y l-1 +b l ).

5. The multifunctional laser weapon simulation training terminal calibration method according to claim 1 is characterized in that: In S3, the step of fusing the outputs of the physical model and the data-driven model to form a comprehensive environmental impact model includes: Assume that the weight of the physical model is α and the weight of the data-driven model is β, then: M com =αP ph +βP data ; Among them, α+β=1, α and β are determined by minimizing the mean square error, P ture For the real laser weapon performance, P data The final output of the data-driven model is the laser weapon performance prediction, P ph The physical model of the laser weapon predicts the power; then the mean square error is: Where n is the number of samples, P ph,i The power predicted by the laser weapon physical model for the i-th sample; P date,i is the laser weapon performance prediction finally output by the data-driven model of the i-th sample; P ture,i is the actual laser weapon performance of the sample; By taking the partial derivatives of MSE with respect to α and β and setting them to zero, we can obtain the following system of equations: By solving the system of equations using matrix operations, we can obtain the optimal α and β: make Then the system of equations is expressed as: Solve this system of equations using the matrix inversion method to obtain the values ​​of α and β:

6. The multifunctional laser weapon simulation training terminal calibration method according to claim 1 is characterized in that: In S4, the correction parameter is calculated by the environmental impact model, that is, δ = h(M com ), where h is a linear function, h(x) = kx + d, then the correction parameter δ = k(αP ph +βP date )+d, where k and d are parameters determined according to actual conditions.

7. The multifunctional laser weapon simulation training terminal calibration method according to claim 1 is characterized in that: In S4, applying the calculated correction parameters to the laser weapon simulation training terminal in real time to achieve dynamic correction of the laser weapon performance includes: S41. Parameter transmission: The calculated correction parameters are transmitted to the laser weapon simulation training terminal in real time through a high-speed data transmission channel; the transmission channel should have low latency and high bandwidth to ensure that the parameters can be applied to the terminal device in a timely manner; S42. Dynamic correction: In the terminal device, a special correction module is set up, which can quickly receive and execute correction parameters to achieve dynamic correction of laser weapon performance; Assume the corrected laser weapon performance is P co =P or +δ, where P or is the performance of the laser weapon before correction, and δ is the correction parameter; the correction module should have high real-time performance and accuracy to ensure that the performance of the laser weapon can respond to environmental changes in a timely manner; for example, when environmental factors change, the correction module can quickly receive new correction parameters and adjust the relevant parameters of the laser weapon to keep the performance of the laser weapon at the best state; if the performance indicator of the laser weapon is energy output, the unit is joule, then the energy output after correction is P co =P or +δ Joule; in this way, real-time dynamic adjustment of laser weapon performance can be achieved, improving the authenticity and effectiveness of simulation training.

8. The multifunctional laser weapon simulation training terminal calibration method according to claim 1 is characterized in that: The S5 includes: S51. Effect evaluation: establish an evaluation index system to evaluate the effect of correction execution; the comprehensive evaluation index is expressed as: E ev =k1ΔI r +k2ΔI a +k3ΔI e ; Among them, k1, k2, k3 are weight coefficients, and satisfy k1+k2+k3=1; ΔI r is the change in range; ΔI a is the change in accuracy; ΔI e is the change in energy output; S52. Establish a feedback loop to optimize the correction strategy. According to the results of the evaluation in S51, establish a feedback loop to optimize the correction strategy; Assume that the sensor's measurement range is R and the original sensor resolution is R es , adjust the measurement range and resolution according to the change in the evaluation index; the adjustment formula of the sensor parameters after adjustment is: R n =R+λΔI r ; Among them, R n To adjust the sensor measurement range; is the adjusted sensor resolution; λ is the measurement range adjustment amount determined according to the evaluation results. If the range changes greatly and the measurement range needs to be expanded, λ takes a larger positive value; if the range increases greatly and the measurement range needs to be appropriately reduced, λ takes a smaller negative value; η is the resolution adjustment amount determined according to the evaluation results. If the accuracy decreases greatly and the resolution needs to be increased, η takes a larger positive value; if the accuracy increases greatly and the resolution needs to be appropriately reduced to balance other performance, η takes a smaller negative value; ΔI r is the range change; ΔI a is the change in accuracy.

Images