Antenna polarization simulation method based on phase and amplitude calibration

Abstract

The present invention discloses an antenna polarization simulation method based on phase and amplitude calibration, including collecting raw data, performing preprocessing, and obtaining preprocessed raw data; constructing an initial calibration model, generating a post-calibration lookup table; obtaining target polarization parameters, calculating Stokes parameters; constructing a Mueller matrix, determining an ideal polarization state; constructing a closed-loop control system, and obtaining monitoring results; constructing a polarization parameter adjustment strategy, performing multi-objective optimization, and generating optimized polarization parameters; obtaining the position information of each antenna, and constructing a multi-antenna system model; using a distributed optimization algorithm, obtaining a coordination result, realizing a spatial diversity strategy, and generating polarization control parameters; and using beamforming technology to generate final polarization control parameters. The present invention not only increases the versatility of the system, but also improves the accuracy and reliability of polarization simulation, and also enhances the adaptability and stability of the system, reduces the complexity of real-time calculation, and improves the response speed of the system.

Description

Technical Field

[0001] The invention belongs to the field of antenna communication, and in particular to an antenna polarization simulation method based on phase and amplitude calibration. Background Art

[0002] Antenna polarization characteristics play a vital role in modern wireless communication systems, directly affecting the quality, reliability and spectrum utilization efficiency of communication links. With the rapid development of 5G, satellite communications and advanced radar systems, unprecedented high requirements are placed on the accuracy and flexibility of antenna polarization control. Accurate polarization control can not only improve the quality of signal reception, but also effectively suppress interference and enhance the system's anti-multipath capability, thereby significantly improving the overall performance of the communication system. In complex electromagnetic environments, such as densely populated urban areas, accurate polarization simulation and dynamic adjustment capabilities have become key determinants of system performance. Therefore, the study of antenna polarization simulation methods based on phase and amplitude calibration has great theoretical significance and practical application value, and is crucial to promoting the advancement of wireless communication technology and meeting the growing demand for high-quality communications.

[0003] At present, antenna polarization simulation and control technology has made certain progress. Traditional methods mainly rely on mechanical adjustment or simple electronic control to change the polarization state of the antenna. In recent years, with the development of digital signal processing technology and adaptive algorithms, some software-defined polarization control methods have emerged. These methods achieve fine control of the polarization state by adjusting the phase and amplitude of the antenna feeding network. At the same time, some studies have begun to explore the use of machine learning algorithms to optimize polarization parameters to adapt to complex and changing communication environments. In terms of calibration technology, vector network analyzer (VNA) has been widely used to measure and calibrate antenna characteristics. Some advanced calibration methods, such as polynomial fitting and lookup table technology, are used to compensate for errors caused by factors such as hardware nonlinearity and temperature drift. In addition, in multi-antenna systems, some studies have begun to focus on cooperative polarization control strategies, trying to achieve higher system performance through the collaborative work of multiple antennas.

[0004] However, there are still some technical problems that need to be solved in current research. First, in terms of accurate description and control of polarization state, traditional simple amplitude and phase adjustment methods are difficult to accurately characterize complex polarization states, especially in broadband systems, where changes in polarization characteristics at different frequencies are difficult to handle uniformly. Secondly, existing calibration methods are often static and difficult to adapt to performance drift caused by factors such as hardware aging and temperature changes during long-term use, affecting the long-term stability of the system. Furthermore, in terms of real-time dynamic polarization adjustment, current methods often have slow response speeds and are difficult to cope with rapidly changing communication environments, especially in high-speed mobile scenarios. In addition, the cooperative polarization control strategy of multi-antenna systems is still in its early stages, and how to achieve global optimal polarization control while reducing computational complexity remains a challenge. Finally, existing polarization control methods often separate polarization adjustment from spatial beam control, lack a unified framework to simultaneously optimize these two dimensions, and limit the system's ability to jointly optimize in the spatial and polarization domains. The existence of these technical problems seriously restricts the application of antenna polarization simulation and control technology in high-performance communication systems, and new methods need to be developed to address these challenges. Summary of the invention

[0005] The purpose of the invention is to provide an antenna polarization simulation method based on phase and amplitude calibration to solve the above problems existing in the prior art.

[0006] The technical solution is an antenna polarization simulation method based on phase and amplitude calibration, comprising the following steps:

[0007] S1. Collecting raw data using a vector network analyzer, and preprocessing the raw data to obtain preprocessed raw data; wherein the raw data includes amplitude data and phase data;

[0008] S2. Based on the preprocessed raw data, an initial calibration model is constructed, and the initial calibration model is iteratively optimized to minimize the error between the measured value and the theoretical value to obtain the final calibration parameters; using the final calibration parameters, a post-calibration lookup table is generated, including an "amplitude value-attenuation code" lookup table and a "phase difference value-phase shift code" lookup table;

[0009] S3, obtain the target polarization parameters and calculate the corresponding Stokes parameters; based on the final calibration parameters, construct the Mueller matrix of the system; use the Mueller matrix to transform the Stokes parameters to obtain the output polarization state; according to the output polarization state, query the calibration lookup table to determine the required attenuation code and the required phase shift code as the ideal polarization state;

[0010] S4. Construct a closed-loop control system to monitor the quality of received signals in real time and obtain monitoring results; based on the monitoring results, use a reinforcement learning algorithm to construct a polarization parameter adjustment strategy; based on the polarization parameter adjustment strategy, perform multi-objective optimization to obtain optimization results; based on the optimization results and the ideal polarization state, generate optimized polarization parameters;

[0011] S5. Obtain the location information of each antenna, and build a multi-antenna system model based on the location information of each antenna and the optimized polarization parameters; based on the multi-antenna system model, use a distributed optimization algorithm to coordinate the polarization states of predetermined antennas; based on the coordination result, implement a spatial diversity strategy and generate polarization control parameters that take spatial diversity into account; based on the polarization control parameters, use beamforming technology to generate the ultimate control parameters for each antenna.

[0012] Beneficial effects: the present invention not only increases the versatility of the system, but also improves the accuracy and reliability of polarization simulation, enhances the adaptability and stability of the system, reduces the complexity of real-time calculation, and improves the response speed of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a flow chart of the present invention.

[0014] Figure 2 This is a flow chart of step S1 of the present invention.

[0015] Figure 3 This is a flow chart of step S2 of the present invention.

[0016] Figure 4 This is a flow chart of step S3 of the present invention.

[0017] Figure 5 This is a flow chart of step S4 of the present invention.

[0018] Figure 6 This is a flow chart of step S5 of the present invention.

[0019] Figure 7 Schematic diagram of device connection according to an embodiment of the present invention.

[0020] Figure 8 is a flow chart of an embodiment of the present invention. DETAILED DESCRIPTION

[0021] like Figure 1 As shown, the present application proposes an antenna polarization simulation method based on phase and amplitude calibration, comprising the following steps:

[0022] S1. Collecting raw data using a vector network analyzer, and preprocessing the raw data to obtain preprocessed raw data; wherein the raw data includes amplitude data and phase data;

[0023] S2. Based on the preprocessed raw data, an initial calibration model is constructed, and the initial calibration model is iteratively optimized to minimize the error between the measured value and the theoretical value to obtain the final calibration parameters; using the final calibration parameters, a post-calibration lookup table is generated, including an "amplitude value-attenuation code" lookup table and a "phase difference value-phase shift code" lookup table;

[0024] S3, obtain the target polarization parameters and calculate the corresponding Stokes parameters; based on the final calibration parameters, construct the Mueller matrix of the system; use the Mueller matrix to transform the Stokes parameters to obtain the output polarization state; according to the output polarization state, query the calibration lookup table to determine the required attenuation code and the required phase shift code as the ideal polarization state;

[0025] S4. Construct a closed-loop control system to monitor the quality of received signals in real time and obtain monitoring results; based on the monitoring results, use a reinforcement learning algorithm to construct a polarization parameter adjustment strategy; based on the polarization parameter adjustment strategy, perform multi-objective optimization to obtain optimization results; based on the optimization results and the ideal polarization state, generate optimized polarization parameters;

[0026] S5. Obtain the location information of each antenna, and build a multi-antenna system model based on the location information of each antenna and the optimized polarization parameters; based on the multi-antenna system model, use a distributed optimization algorithm to coordinate the polarization states of predetermined antennas; based on the coordination result, implement a spatial diversity strategy and generate polarization control parameters that take spatial diversity into account; based on the polarization control parameters, use beamforming technology to generate the ultimate control parameters for each antenna.

[0027] like Figure 2 As shown, according to one aspect of the present application, step S1 is further:

[0028] S11. According to a predefined frequency range and step value, the vector network analyzer is configured, including frequency scanning settings, output power, intermediate frequency bandwidth and average factor of the vector network analyzer, to obtain a configured vector network analyzer; and the configured vector network analyzer is calibrated to obtain a calibrated vector network analyzer.

[0029] S12. At each frequency point, cyclically set the attenuation code and phase shift code of the polarization simulation device; for each attenuation code, use a calibrated vector network analyzer to measure the amplitude data of the horizontal channel and the vertical channel; for each phase shift code, use a calibrated vector network analyzer to measure the phase data of the horizontal channel and the vertical channel; save the amplitude data and the phase data as original data;

[0030] S13. Based on the original data, use the moving window method to calculate the local mean and standard deviation of each data point, and mark the data points that deviate from the local mean by more than a predetermined number of standard deviations as outliers; use the interpolation method to replace the outliers to obtain the original data without the outliers; based on the original data without the outliers, apply a low-pass filter to remove high-frequency noise to obtain the original data without the noise; perform data standardization on the original data without the noise to obtain the preprocessed original data.

[0031] In one embodiment of the present application, the operating parameters of the vector network analyzer are set. The frequency scanning settings of the vector network analyzer are configured according to the predefined frequency range and step value. The output power of the vector network analyzer is adjusted to match the input range of the polarization simulation device. The intermediate frequency bandwidth and averaging factor of the vector network analyzer are set to optimize the measurement accuracy and speed. The vector network analyzer is calibrated to eliminate systematic errors and ensure the accuracy of the measurement.

[0032] Perform data acquisition on the horizontal channel. At each frequency point, set the attenuation code of the polarization simulation device in a loop, ranging from the minimum value to the maximum value, with a step of a predefined value. For each attenuation code setting, record the amplitude value measured by the vector network analyzer. Set the phase shift code in a loop, ranging from 0 degrees to 360 degrees, with a step of a predefined value. For each phase shift code setting, record the measured phase value. After completing the data acquisition of the horizontal channel, switch to the vertical channel and repeat the same process. Temporarily store the acquired raw data in a structured format, including information such as frequency, channel, attenuation code, phase shift code, amplitude value, and phase value.

[0033] Perform outlier detection, using a moving window method to calculate the local mean and standard deviation of each data point, and mark data points that deviate from the local mean by more than three standard deviations as outliers. Use interpolation to replace these outliers, specifically using linear interpolation of adjacent valid data points. Apply a low-pass filter to remove high-frequency noise, and the cutoff frequency of the filter is dynamically adjusted according to the signal bandwidth. Perform data normalization. For amplitude data, calculate the mean and standard deviation of each frequency point and convert the raw amplitude values ​​to z-scores. For phase data, map it to the interval [-π, π]. Store the preprocessed data in a structured format, including information such as frequency, channel, attenuation code, phase shift code, normalized amplitude value, and normalized phase value.

[0034] In another embodiment of the present application, a vector network analyzer is used to collect raw data. Assume that the frequency point set is F={f1, f2, ..., f N}, where f1=1GHz, f N =18GHz, N=1701 (10 MHz step). For each frequency point f i, attenuation code set A = {a1, a2, ..., a M}, phase-shift code set P = {p1, p2, ..., p K}. The collected raw data is in the form of: D(f i , a j , p k )=[A H ,φ H , A V ,φ V ] where f i Indicates the current frequency point; a j represents the attenuation code; p k Represents phase shift code; A H , A V Indicates the amplitude value of the horizontal and vertical channels (unit: dB); φ H , φ V Indicates the phase value of the horizontal and vertical channels (unit: degree).

[0035] Data preprocessing includes outlier detection and removal. Using the moving window method, calculate the local mean μ and standard deviation σ: μ(f i )=(1 / W) *ΣD(f i +w), w∈[-W / 2, W / 2];σ(f i ) = sqrt((1 / W)*Σ(D(f i +w)-μ(f i )) 2 ), w∈[-W / 2, W / 2]; where W is the window size, usually W=21. Data points that deviate from the mean by more than 3σ are marked as outliers and replaced by linear interpolation: D'(f i ) = (D(f i-1 ) + D(f i+1 )) / 2, if |D(f i ) - μ(f i )|>3σ(f i ).

[0036] This embodiment realizes high-precision, wide-band antenna polarization characteristic measurement by using a vector network analyzer to collect raw amplitude and phase data and preprocessing the collected data. Specifically, by setting the frequency scanning range and step of the vector network analyzer, it is ensured that the measurement covers the entire target frequency band, providing comprehensive basic data for subsequent polarization simulation. Data collection is performed for the horizontal and vertical channels respectively to ensure a complete description of the polarization state. The attenuation code and phase shift code of the polarization simulation device are set cyclically at each frequency point, and the corresponding amplitude value and phase value are recorded. This fine-grained measurement method makes the subsequent calibration and simulation process more accurate. After the acquisition is completed, the raw data is cleaned to remove outliers and noise, thereby improving the data quality. By using statistical methods to identify and eliminate outliers, the impact of measurement errors on subsequent processing is effectively reduced. The subsequent data standardization process ensures that data of different channels and frequencies are comparable, laying the foundation for subsequent analysis and processing. This standardization process solves the problem of data inconsistency caused by hardware differences and frequency characteristics, making cross-band and cross-channel data analysis possible. Through data acquisition and preprocessing, not only the quality and reliability of the original data are improved, but also high-quality input data is provided for subsequent calibration algorithms and polarization simulations. Compared with the traditional single frequency point or rough scanning method, this embodiment can capture the polarization characteristics of the antenna more comprehensively, especially in wide-band applications, and can accurately characterize the polarization response of the antenna at different frequencies. At the same time, by processing the horizontal and vertical channels separately, sufficient flexibility is provided for the subsequent polarization state synthesis. In addition, it also has strong adaptability, and can adjust the frequency scanning range and step, as well as the setting range of the attenuation code and phase shift code according to actual needs, so as to adapt to different types of antennas and application scenarios. This embodiment not only improves the accuracy and reliability of the data, but also provides a solid data foundation for subsequent polarization simulation and dynamic adjustment.

[0037] like Figure 3 As shown, according to one aspect of the present application, step S2 is further:

[0038] S21, based on the preprocessed raw data, extracting the original amplitude value, phase value, attenuation code and phase shift code; based on the original amplitude value, phase value, attenuation code and phase shift code, constructing an initial calibration model to obtain the calibrated amplitude and phase values;

[0039] S22, calculating the mean square error between the calibrated amplitude and phase values ​​and the preset theoretical values ​​as the objective function; based on the objective function, using the Levenberg-Marquardt algorithm to iteratively optimize the parameters until the preset convergence condition is reached, and obtaining the final optimized parameter values ​​as the final parameters of the calibration model;

[0040] S23, based on the final parameters of the calibration model, calibrating the original amplitude value and phase value to obtain the final calibrated amplitude and phase value;

[0041] S24, based on the final calibrated amplitude, within a predefined amplitude range, generate a target amplitude value with a preset fixed step size; for each target amplitude value, use the inverse function of the calibration model to calculate the corresponding attenuation code, and generate an "amplitude value-attenuation code" lookup table;

[0042] S25. Based on the final calibrated phase value, a target phase difference value is generated within a predefined phase range with a preset fixed step size; for each target phase difference value, the corresponding phase shift code is calculated using the inverse function of the calibration model to generate a "phase difference value-phase shift code" lookup table.

[0043] In one embodiment of the present application, an initial calibration model is constructed. Amplitude values, phase values, attenuation codes, and phase shift code data are extracted from preprocessed data. For amplitude calibration, a quadratic function relationship is assumed to exist, and the function parameters are initialized. For phase calibration, a linear relationship is assumed to exist, and the slope and intercept are initialized. The initial model parameters are fitted using the least squares method. The initial model parameters are stored in a vector form for subsequent optimization.

[0044] Perform an iterative optimization process. Construct the objective function and calculate the mean square error between the calibrated amplitude and phase values ​​and the theoretical values. Use the Levenberg-Marquardt algorithm for parameter optimization. In each iteration, calculate the Jacobian matrix and update the parameter estimates. Check the convergence condition and stop the iteration if it is met, otherwise continue with the next round of iteration. Record the final optimized parameter values ​​as the final parameters of the calibration model.

[0045] Generate a calibrated lookup table. Use the optimized calibration model parameters to calibrate the original amplitude and phase values. For amplitude calibration, generate a series of target amplitude values ​​with a fixed step size within the predefined amplitude range. For each target amplitude value, use the inverse function of the calibration model to calculate the corresponding attenuation code and generate an "amplitude value-attenuation code" mapping. For phase calibration, generate a series of target phase difference values ​​with a fixed step size within the range of 0 to 360 degrees. For each target phase difference value, use the inverse function of the calibration model to calculate the corresponding phase shift code and generate a "phase difference value-phase shift code" mapping. Store the generated lookup table as structured data, including information such as frequency, amplitude value, attenuation code, phase difference value, and phase shift code.

[0046] In another embodiment of the present application, an initial calibration model is constructed: the amplitude calibration model is a quadratic function: A cal =α*A raw 2 +β*A raw+γ; the phase calibration model is a linear function: φ cal =a*φ raw +b; where A cal ,φ cal Represents the calibrated amplitude and phase; A raw ,φ raw represents the amplitude and phase of the original measurement; α, β, γ, a, b represent the calibration parameters to be optimized.

[0047] Use the Levenberg-Marquardt algorithm to minimize the objective function: J(θ)=Σ[(A cal _ i -A true _ i ) 2 +λ*(φ cal _ i -φ true _ i ) 2 ], where θ=[α, β, γ, a, b] represents the parameter vector; λ represents the weight factor, balancing the amplitude and phase errors; i represents the data point index; A cal _ i represents the amplitude of the ith data point after calibration, A true _ i represents the true amplitude of the ith data point, φ cal _ i represents the phase of the ith data point after calibration, φ true _ i represents the true phase of the i-th data point. The parameter update rule is: θ new =θ old -(J'J+μI) -1 *J'r; where J' represents the Jacobian matrix, μ represents the damping factor, I represents the identity matrix, and r represents the residual vector. For each frequency point f, a lookup table of "amplitude value-attenuation code" and "phase difference value-phase shift code" is generated.

[0048] In another embodiment of the present application, an adaptive calibration algorithm is performed to improve calibration accuracy and robustness. An initial calibration model is established, including amplitude and phase calibration functions. Calibration parameters are iteratively updated using real-time measurement data. An adaptive filtering algorithm is applied to suppress measurement noise and interference. An updated calibration parameter lookup table is generated for subsequent polarization simulation. Specifically:

[0049] Establish the initial calibration model. Define the amplitude calibration function A cal =f(A raw ,θ A ) and the phase calibration function φ cal =g(φraw ,θ φ ), where A raw and φ raw is the original measurement value, θ A and θ φ is the parameter vector to be optimized, f( ) represents the amplitude calibration function, and g( ) represents the phase calibration function. Use historical data to estimate the initial parameter values ​​by the least squares method. Calculate the initial calibration error and obtain the error statistical characteristics. Store the initial parameters and error characteristics as structured data for subsequent optimization.

[0050] Iteratively update the calibration parameters. Get a new measurement data set {A raw _ i ,φ raw _ i , A true _ i ,φ true _ i}. Define the objective function J(θ)=Σ[(A cal _ i -A true _ i ) 2 +λ(φ cal _ i -φ true _ i ) 2 ], where λ is the weight factor. The Levenberg-Marquardt algorithm is used to solve the optimal parameter: θ new = θ old - (H + μI) -1 * ▽J, where H is the Hessian matrix of the objective function, μ is the damping factor, and ▽ represents the gradient of the objective function. Repeat this process until convergence or the maximum number of iterations is reached. Update the parameter storage and record the parameter change trend during the optimization process.

[0051] Apply adaptive filtering algorithm. Construct Kalman filter model, the state vector is calibration parameter θ, and the observation vector is calibration error e. Define state transfer equation θ k =θ k-1 +w k and the observation equation e k =h(θ k )+v k , where w k and v k are process noise and observation noise respectively. The prediction and update steps of Kalman filtering are performed to estimate the optimal calibration parameters in real time. At the same time, the noise covariance matrix is ​​adaptively adjusted to adapt to different measurement conditions. The filtered calibration parameter estimates are output.

[0052] Generate an updated calibration parameter lookup table. Based on the filtered calibration parameter estimates, calculate the calibrated values ​​within the predefined amplitude and phase ranges. Generate a continuous calibration surface using a multidimensional interpolation method. Sample at discrete grid points to form a lookup table. To improve query efficiency, implement a nearest neighbor search algorithm based on a kd tree. Compress the lookup table and record version information and generation timestamp for subsequent use and management.

[0053] This embodiment implements high-precision, adaptive antenna polarization characteristic calibration by executing an adaptive calibration algorithm and generating a calibrated lookup table. This solves the problem of reduced accuracy of the traditional fixed calibration method in the face of complex electromagnetic environments and long-term use. The initial calibration model is constructed using preprocessed data, which makes full use of the collected high-quality data and ensures that the starting point of the calibration has high accuracy. The calibration parameters are continuously adjusted through an iterative optimization process to minimize the error between the measured value and the theoretical value. This adaptive optimization method can effectively cope with the nonlinearity and time-varying nature of antenna characteristics. Specifically, by using advanced optimization methods such as the Levenberg-Marquardt algorithm, the calibration process can quickly converge to the optimal solution while avoiding falling into a local optimum. Compared with traditional fixed parameter calibration, this embodiment can better adapt to different working environments and antenna states. After the optimization is completed, the final calibration parameters are used to generate the "amplitude value-attenuation code" and "phase difference value-phase shift code" lookup tables. This lookup table generation method ensures that the polarization state can be adjusted quickly and accurately in actual use. The use of the lookup table reduces the complexity of real-time calculations and improves the response speed of the system. The generated lookup table is stored as structured data. This storage method not only facilitates subsequent quick access, but also facilitates data updating and maintenance, and can continuously optimize performance. By re-executing the calibration process regularly or when performance degradation is detected, the system can maintain long-term high-precision performance, which is particularly important for communication systems that require long-term stable operation. In addition, this embodiment is also highly versatile and can adapt to different types of antennas and polarization simulation devices. It only needs to adjust the parameters and objective functions of the optimization algorithm. In practical applications, this adaptive calibration method can improve the accuracy of polarization simulation, especially in the presence of temperature drift, device aging or environmental changes. For example, in a satellite communication system, the ground station antenna needs to accurately adjust the polarization state to match the satellite signal, and this adaptive calibration method can ensure long-term high-precision tracking. This embodiment not only improves the accuracy and reliability of polarization simulation, but also enhances the adaptability and stability of the system, laying the foundation for high-performance polarization control.

[0054] like Figure 4 As shown, according to one aspect of the present application, step S3 is further:

[0055] S31, obtaining target polarization parameters input by the user, including ellipticity and azimuth, and calculating a normalized Jones vector based on the target polarization parameters; calculating a Stokes parameter vector based on the Jones vector; and storing the Stokes parameter vector as a four-dimensional vector;

[0056] S32. Based on the final calibration parameters, a frequency-dependent Mueller matrix model is constructed; using the Mueller matrix model, the Mueller matrix elements of each key frequency point are calculated; based on the Mueller matrix elements, a continuous frequency-dependent Mueller matrix function is constructed using an interpolation method;

[0057] S33, obtaining and based on the current operating frequency, obtaining an applicable Mueller matrix in a continuous frequency-dependent Mueller matrix function; multiplying the Stokes parameter vector by the Mueller matrix to obtain a product result;

[0058] S34, converting the product result to obtain converted parameters; based on the converted parameters, querying the post-calibration lookup table to determine the attenuation code and phase shift code required for each antenna as the ideal polarization state of each antenna.

[0059] In one embodiment of the present application, Stokes parameters are calculated based on the input target polarization state. A user-defined target polarization state is received, which may be given in the form of ellipticity and azimuth. The Stokes parameters are calculated using the formula S0=1, S1=cos(2χ)cos(2ψ), S2 =cos(2χ)sin(2ψ), S3= sin(2χ), where χ is the ellipticity and ψ is the azimuth. The calculated Stokes parameter vector [S0, S1, S2, S3] is stored as structured data for subsequent processing.

[0060] Construct the Mueller matrix of the system. Use the obtained calibration data to establish the frequency-dependent Mueller matrix model M(f). For each key frequency point, calculate the corresponding Mueller matrix element. Use the interpolation method to construct a continuous frequency-dependent Mueller matrix function. Finally, store the complete Mueller matrix model in function form M(f), where f is the frequency.

[0061] Perform Mueller matrix transformation. Get Stokes parameter vector S in . According to the current operating frequency f, obtain the applicable Mueller matrix M(f) from the model. Perform matrix multiplication S out =M(f)*S in , and get the output Stokes parameters. out Store for further processing.

[0062] Determine the control parameters of the ideal polarization state. Convert the output Stokes parameters back to ellipticity and azimuth. According to the converted parameters, query the post-calibration lookup table to determine the attenuation code and phase shift code required for each antenna. Output the ideal polarization state (including attenuation code and phase shift code) of each antenna as structured data.

[0063] According to one aspect of the present application, the Stokes parameters are used to describe the polarization state: S = [S0, S1, S2, S3] T , where: S0=A H 2 +A V 2 ; S1=A H 2 -A V 2 ; S2=2*A H *A V *cos(φ V -φ H ); S3=2*A H *A V *sin(φ V -φ H ). The Mueller matrix M(f) can be expressed as: M(f)=[m 00 (f), m 01 (f), m 02 (f), m 03 (f); m 10 (f), m 11 (f), m 12 (f), m 13 (f); m 20 (f), m 21 (f), m 22 (f), m 23 (f); m 30 (f), m 31 (f), m 32 (f), m 33 (f)], where each element m ij (f) is a function of frequency f. Output Stokes parameter S out Calculated by: S out =M(f)*S in .

[0064] In another embodiment of the present application, a polarization simulation algorithm based on Stokes parameters and Mueller matrices is implemented. The input polarization state is converted to Stokes parameter representation. A Mueller matrix model of the system is constructed, including frequency dependence. The Mueller matrix transformation is performed to obtain the output Stokes parameters. The Stokes parameters are converted back to conventional amplitude and phase representation for controlling the hardware. Specifically:

[0065] Converts input polarization state to Stokes parameters. Accepts user-defined polarization state, which may be given as ellipticity and azimuth, Jones vector, or other form. Calculates normalized electric field components Ex and Ey. Uses the formula S0=|Ex| 2 +|Ey| 2 , S1=|Ex| 2 -|Ey| 2 , S2=2Re(ExEy), S3=2Im(ExEy) to calculate the Stokes parameters, where * represents complex conjugation, Re( ) represents taking the real part of the complex number, and Im( ) represents taking the imaginary part of the complex number. Normalize the Stokes vector to ensure that S0=1.

[0066] Construct a Mueller matrix model of the system. Consider each component of the system, such as antennas, phase shifters, attenuators, etc., and establish their Mueller matrix models respectively. Consider the frequency dependence and establish an independent Mueller matrix for each key frequency point. Use the interpolation method to construct a continuous frequency-dependent Mueller matrix function M(f). Consider the influence of environmental factors such as temperature and introduce parameterized environmental compensation terms. The complete Mueller matrix model is expressed as M(f, T, θ), where f is the frequency, T is the temperature, and θ is other environmental parameters.

[0067] Perform Mueller matrix transformation. Obtain the applicable Mueller matrix M from the model based on the current operating frequency and environmental parameters. Transform the input Stokes parameter vector S in Multiply with the Mueller matrix: S out =M*S in The output Stokes parameters are normalized to ensure physical validity. The polarization preservation of polarization transformation is calculated to evaluate system performance. out and related performance indicators are stored for subsequent processing.

[0068] Convert the Stokes parameters back to magnitude and phase representation. From the output Stokes parameters S out Extract the normalized electric field component from . Calculate the amplitude of the horizontal and vertical channels: A H=sqrt((S0+S1) / 2), A V =sqrt((S0-S1) / 2). Calculate the phase difference: Δφ=arctan(S3 / S2). To improve accuracy, use the four-quadrant inverse tangent function atan2(S3, S2). Finally, query the calibration parameter lookup table to convert the ideal amplitude and phase values ​​into actual control parameters, such as attenuation code and phase shift code.

[0069] This embodiment improves the accuracy and flexibility of antenna polarization simulation by implementing a polarization simulation algorithm based on Stokes parameters and Mueller matrix. It solves the problem that it is difficult to accurately describe and control complex polarization states in traditional polarization simulation. The corresponding Stokes parameters are calculated according to the input target polarization state, and the complex polarization state is converted into a mathematically easy-to-handle form. The use of Stokes parameters enables the system to uniformly handle any polarization state, including linear, circular polarization and elliptical polarization, increasing the versatility of the system. The Mueller matrix of the system is constructed and the frequency dependence is considered, so that the simulation process can accurately reflect the changes in the polarization characteristics of the antenna at different frequencies. By using the frequency as a parameter of the Mueller matrix, the system can accurately simulate the polarization state within a wide frequency band, which is crucial for modern broadband communication systems. The Stokes parameters are transformed using the Mueller matrix to obtain the output polarization state. This matrix transformation method is not only computationally efficient, but also can accurately simulate the changes in the polarization state during transmission, including the rotation of the polarization state, the change of the ellipticity, etc. This precise simulation capability is particularly important in complex multipath propagation environments, and can help the system better cope with the polarization distortion of the channel. According to the output polarization state, the calibration lookup table is queried to determine the required attenuation code and phase shift code, and the theoretical calculation results are seamlessly connected with the actual hardware control, ensuring that the simulation results can be accurately implemented on the hardware. This method based on Stokes parameters and Mueller matrices has many advantages over traditional simple amplitude and phase control. It can more comprehensively describe and control the polarization state, including partial polarization and depolarization effects, which is particularly important in complex wireless communication environments; at the same time, it has a strong mathematical foundation, enabling the system to perform accurate theoretical analysis and prediction. Since the Mueller matrix can describe any polarization transformation, this embodiment can be easily extended to more complex scenarios, such as multi-hop links or multi-layer medium propagation. In practical applications, the performance of the communication system can be improved, especially in scenarios where precise polarization control is required, such as satellite communications, radar systems, or high-capacity MIMO systems. For example, in satellite communications, the ground station can adjust the polarization state of the transmitted signal in real time according to the satellite's motion trajectory and atmospheric conditions to maximize the signal quality. This embodiment not only improves the accuracy and flexibility of polarization simulation, but also provides a strong theoretical and practical basis for the implementation of advanced polarization control strategies.

[0070] like Figure 5 As shown, according to one aspect of the present application, step S4 is further:

[0071] S41, based on predefined comprehensive performance indicators, using a preconfigured signal quality monitoring module, collecting performance indicator data in real time, including signal-to-noise ratio, signal-to-interference ratio, and bit error rate;

[0072] S42, constructing a state space based on the performance indicator data and the current polarization parameters; constructing an action space based on the preset adjustable attenuation code and phase shift code range; implementing a Q-learning algorithm based on the state space and the action space to obtain an updated Q table; selecting an optimal polarization parameter adjustment strategy for each antenna based on the values ​​in the Q table;

[0073] S43. Based on the performance indicator data, a predetermined objective function is constructed; based on the predetermined objective function and the polarization parameter adjustment strategy, the NSGA-II algorithm is used for optimization, and the Pareto rank and the optimized population are iteratively generated; based on the Pareto rank and the optimized population, the optimal polarization parameter is selected;

[0074] S44. Based on the optimal polarization parameter and the ideal polarization state, an adjustment amount is calculated, and the adjustment amount is ensured to be within a preset hardware limit range; based on the adjustment amount, a new attenuation code and a new phase shift code are generated for each antenna as optimized polarization parameters.

[0075] According to one aspect of the present application, step S43 is further:

[0076] S431, constructing a predetermined objective function based on the performance indicator data; generating an initial population using the NSGA-II algorithm based on the predetermined objective function, wherein each individual of the initial population represents a set of polarization parameters;

[0077] S432, based on the polarization parameter adjustment strategy and the initial population, non-dominated sorting and crowding degree calculation are performed to obtain the Pareto rank and crowding degree of each individual;

[0078] S433. Based on the Pareto rank and crowding degree of each individual, selection, crossover and mutation operations are performed, and the operations are repeated until a preset termination condition is reached to obtain an iterative optimized population; based on the Pareto rank and the optimized population, the optimal polarization parameters are selected.

[0079] In one embodiment of the present application, a closed-loop control system is established. A signal quality monitoring module is configured, and the sampling rate and measurement period are set. A comprehensive performance index J=w1*SNR+w2*SIR-w3*BER is defined, where SNR is the signal-to-noise ratio, SIR is the signal-to-interference ratio, BER is the bit error rate, and w1, w2, and w3 are weight coefficients. Performance index data is collected in real time, and the collected performance data is associated with the ideal polarization state parameters and stored as time series data for subsequent optimization.

[0080] Use reinforcement learning to optimize polarization parameters. Define state space S, which contains current polarization parameters and performance indicator data. Define action space A, which includes adjustable attenuation code and phase shift code range. Implement Q-learning algorithm and initialize Q table. Loop through the following steps: (1) Select action based on current state (ε-greedy strategy); (2) Execute action, observe new state and reward; (3) Update Q value. Based on the converged Q table, select the optimal polarization parameter adjustment strategy for each antenna.

[0081] Perform multi-objective optimization. Construct multiple objective functions f1(x)=-SNR(x), f2(x)=-SIR(x), f3(x)=BER(x), where x is the polarization parameter vector. Use NSGA-II algorithm for optimization: initialize the population, each individual represents a set of polarization parameters; perform non-dominated sorting and crowding calculation; perform selection, crossover and mutation operations. Repeat this process until convergence. Select the best polarization parameter combination from the Pareto front.

[0082] Generate optimized polarization parameters. Compare the optimal polarization parameters to the ideal polarization state. Calculate the adjustment amount to ensure it is within the hardware limitations. Generate new attenuation codes and phase shift codes for each antenna. Output the optimized polarization parameters (including attenuation codes and phase shift codes for each antenna) as structured data.

[0083] In another embodiment of the present application, a real-time dynamic polarization adjustment mechanism is implemented. A closed-loop control system is established to monitor the quality of the received signal in real time. A reinforcement learning algorithm is used to optimize the polarization parameters. Multi-objective optimization is performed to balance signal strength, signal-to-noise ratio, and interference suppression. An optimized control instruction sequence is generated to dynamically adjust the polarization state. Specifically:

[0084] Establish a closed-loop control system. Configure the signal quality monitoring module, set the sampling rate and measurement period. Define the comprehensive performance index J=w1*SNR+w2*SIR-w3*BER, where SNR is the signal-to-noise ratio, SIR is the signal-to-interference ratio, BER is the bit error rate, and w1, w2, and w3 are weight coefficients. Implement the data acquisition program to periodically read the performance indicators. At the same time, build an adaptive filtering algorithm, such as Kalman filtering, to estimate the true value and change trend of the performance indicators in real time. Store the processed performance data in a circular buffer for use by the optimization algorithm.

[0085] Use reinforcement learning algorithm to optimize polarization parameters. Define state space S, including current polarization parameters and performance indicators. Define action space A, including adjustable attenuation code and phase shift code range. Implement Q-learning algorithm and initialize Q table: Q(s, a) = 0. Define reward function R(s, a) as the improvement of performance indicator relative to the baseline value. Loop through the following steps: (1) Select action a in the current state s (ε-greedy strategy); (2) Execute the action, observe the new state s' and reward r; (3) Update Q value: Q(s, a)←Q(s, a)+α[r+γ*max a' Q(s', a') -Q(s, a)], where s, a represent the current state and action; s', a' represent the next state and action; r represents the reward; α is the learning rate; γ is the discount factor. Based on the converged Q table, the optimal polarization parameter adjustment strategy is selected.

[0086] Perform multi-objective optimization. Define multiple objective functions, f1(x)=-SNR(x), f2(x)=-SIR(x), f3(x)=BER(x), where x is the polarization parameter vector. Use the NSGA-II (Non-dominated Sorting Genetic AlgorithmII) algorithm for optimization. Initialize the population P, where each individual represents a set of polarization parameters. Perform the following steps: (1) Non-dominated sorting of the population; (2) Calculate the crowding distance; (3) Select the parent generation, perform crossover and mutation operations, and generate the offspring Q; (4) Merge P and Q, sort and select again to form a new generation of population. Repeat this process until the termination condition is reached. Select the best polarization parameter combination from the Pareto front as the optimization result.

[0087] Generate an optimized control instruction sequence. Convert the optimal polarization parameters into hardware control instructions, including attenuation codes and phase shift codes. Build a smooth transition strategy to avoid sudden changes in parameters. Use the spline interpolation method to generate a continuous parameter change curve between the current state and the target state. Uniformly sample on the time axis to form a discrete control instruction sequence. At the same time, considering the hardware response characteristics, pre-distort the control sequence to compensate for system delay and nonlinearity. Pack the generated control instruction sequence, add timestamps and verification information, and output it to the hardware control module in the form of a data stream.

[0088] This embodiment improves the adaptability and performance of the antenna system in a complex and dynamic electromagnetic environment by implementing a real-time dynamic polarization adjustment mechanism. It solves the problem that traditional static polarization control is slow to respond and has degraded performance when facing a rapidly changing communication environment. A closed-loop control system is established to monitor the quality of the received signal in real time, so that the system can perceive environmental changes and performance fluctuations in a timely manner. By continuously monitoring key performance indicators such as signal-to-noise ratio, signal-to-interference ratio, and bit error rate, the system can quickly identify performance degradation and provide real-time and accurate feedback for subsequent optimization. The use of reinforcement learning algorithms to optimize polarization parameters can continuously optimize decision-making strategies through continuous interaction with the environment, which is particularly suitable for dealing with complex and dynamic problems. Using reinforcement learning in polarization adjustment enables the system to autonomously learn the optimal adjustment strategy without pre-defining complex rules. It can not only cope with known interference patterns, but also adapt to newly emerging and unknown interference types. Perform multi-objective optimization to balance signal strength, signal-to-noise ratio, and interference suppression. This multi-objective optimization method takes into account multiple aspects of communication system performance and avoids the overall performance degradation that may be caused by optimizing a single indicator. By using multi-objective optimization algorithms such as NSGA-II, the system can find the best balance between different performance indicators to meet the diverse needs of complex application scenarios. The polarization simulation parameters are dynamically adjusted according to the optimization results, and the optimization results are applied to the hardware control in real time to ensure that the theoretical optimization can be converted into actual performance improvement. This embodiment can quickly respond to environmental changes and complete the adjustment of polarization state within milliseconds, which is crucial for high-speed mobile communications or rapidly changing interference environments; through continuous learning and optimization, the system can continuously improve its ability to cope with various interference and channel conditions and achieve long-term performance improvement; it has strong versatility and scalability, and can adapt to different types of antenna systems and communication scenarios, just by adjusting the parameters and objective functions of the optimization algorithm. In practical applications, this real-time dynamic polarization adjustment mechanism can improve the reliability and efficiency of the communication system. For example, in the application of Internet of Vehicles in urban environments, the rapid movement of vehicles causes frequent changes in channel conditions, and real-time dynamic polarization adjustment can ensure the continuous stability of the communication link. This embodiment not only improves the adaptability and performance of the system in complex environments, but also lays the foundation for intelligent and autonomous polarization control systems, opening a new era of antenna polarization control.

[0089] like Figure 6 As shown, according to one aspect of the present application, step S5 is further:

[0090] S51, obtaining the position information of each antenna and constructing the geometric structure of the antenna array; obtaining the optimized polarization parameters and initializing them to obtain the initial polarization state of each antenna; constructing the coupling matrix between the antennas based on the geometric structure of the antenna array and the initial polarization state of each antenna; integrating the position information, initial polarization state and coupling matrix of each antenna into structured data to form a multi-antenna system model;

[0091] S52, dividing the multi-antenna system model into predetermined subsystems, constructing a local objective function for each subsystem; based on the local objective function, implementing a distributed optimization algorithm based on ADMM to obtain an optimized global polarization state;

[0092] S53, based on the multi-antenna system model, calculating the spatial correlation matrix of each antenna pair; based on the spatial correlation matrix, performing eigenvalue decomposition to obtain an eigenvector matrix; selecting orthogonal eigenvectors from the eigenvector matrix as polarization directions of different antennas; combining the orthogonal eigenvectors with the optimized global polarization state to generate polarization control parameters;

[0093] S54. Based on the preset desired beam direction, calculate the beamforming weight of each antenna; combine the beamforming weight with the polarization control parameter to generate a composite weight matrix; use the MVDR algorithm to optimize the composite weight matrix to obtain an optimized composite weight matrix; convert the optimized composite weight into the ultimate control parameter of each antenna.

[0094] In one embodiment of the present application, a multi-antenna system model is established. The geometric structure of the antenna array is defined, and the position of each antenna is represented by three-dimensional coordinates (xi, yi, zi). The optimized polarization parameters of each antenna are obtained and the polarization state is initialized. A mutual coupling model between antennas is established and a coupling matrix C(f) is defined. The system model parameters are integrated into structured data, including antenna positions, initial polarization states, and coupling matrices.

[0095] Execute the distributed optimization algorithm. Divide the multi-antenna system into multiple subsystems. Construct the local objective function Ji(S i )=w1*SNRi+w2*SIRi-w3*BERi, where S i is the set of polarization states of the ith subsystem. Implement the distributed optimization algorithm based on ADMM: initialize the global variable z and the dual variable u; iteratively solve the local optimization problem, update the global variable and the dual variable. The converged z is output as the optimized global polarization state.

[0096] Implement spatial diversity strategy. Based on the system model, calculate the spatial correlation matrix R of each antenna pair. Perform eigenvalue decomposition R=U∧U H , where ∧ represents the diagonal eigenvalue matrix and U represents the eigenvector matrix. Orthogonal eigenvectors are selected as the polarization directions of different antennas. The assigned polarization state is combined with the optimization result to generate polarization parameters considering spatial diversity.

[0097] Integrated beamforming technology. Define the desired beam direction (θ0, φ0). Calculate the complex weights wi for each antenna. Combine the beamforming weights with the polarization control parameters to generate a composite weight matrix W. Use the MVDR algorithm to optimize the composite weights to suppress interference and noise. Convert the optimized composite weights to the ultimate control parameters (attenuation code and phase shift code) for each antenna as structured data output.

[0098] The specific implementation of the multi-antenna cooperative polarization control strategy is as follows: define the antenna position matrix P: P = [x1, y1, z1; x2, y2, z2; ...; x16, y16, z16]; the initial polarization state vector S: S = [S1; S2; ...; S 16 ], where S i is the Stokes parameter of the ith antenna. Using the ADMM algorithm: minimize Σ i J i (S i ) +(ρ / 2)Σ i ||S i -z i +u i || 2 , where J i (S i ) represents the local objective function of the ith antenna, z i Indicates global variables, u i represents the dual variable, and ρ represents the penalty parameter. Calculate the spatial correlation matrix R and perform eigenvalue decomposition: R=U∧U H , select orthogonal eigenvectors as the polarization directions of different antennas. Calculate the beamforming weights: w i =exp(-j2π(x i sinθcosφ+y i sinθsinφ+z i *cosθ) / λ), where θ, φ represent the desired beam directions, λ represents the operating wavelength, and exp( ) represents the exponential function.

[0099] In another embodiment of the present application, it is assumed that there are L antennas, and the polarization state of each antenna is determined by the Stokes parameter S l Represents. Define the global objective function: Jglobal =Σ l w l *J local (S l )+λ*Σ l，m C lm (S l , S m );where w l represents the antenna weight, J local represents the local objective function, C lm represents the coupling function between antennas l and m, and λ represents the coupling weight.

[0100] Use the alternating direction method of multipliers (ADMM) to solve: Minimize the global objective function J global Satisfy the constraint S l =z l , l=1,...,L; solve by iteration:

[0101] S l k+1 = argmin Sl [J local (S l ) + (ρ / 2) ||S l -z l k + u l k ||2 2 ];

[0102] z k+1 = argmin z [λ * Σ l，m C lm (z l , z m ) + (ρ / 2) Σ l ||S l k+1 -z l + u l k ||2 2 ];

[0103] u l k+1 = u l k + S l k+1 -z l k+1 ;

[0104] Where ρ is the penalty parameter, u l is the dual variable, S lk+1 represents the Stokes parameter vector of the lth antenna in the k+1th iteration, z k+1 represents the intermediate variable vector of all antennas in the k+1th iteration, u l k+1 represents the dual variable of the lth antenna in the k+1th iteration, argmin represents the variable value that minimizes the objective function, and u l k represents the dual variable of the lth antenna in the kth iteration, z l and z m Represents the lth and mth elements in the intermediate variable vector.

[0105] In another embodiment of the present application, a multi-antenna system model is established, including the position and initial polarization state of each antenna. A distributed optimization algorithm is constructed to coordinate the polarization states of multiple antennas. A spatial diversity strategy is implemented to use different polarization states at different spatial locations. Beamforming technology is integrated to combine polarization control and beam control. Specifically:

[0106] Build a multi-antenna system model. Define the geometry of the antenna array and use three-dimensional coordinates (xi, yi, zi) to represent the position of each antenna. Initialize the polarization state of each antenna and use the Stokes parameter S i =[S 0i , S 1i , S 2i , S 3i ] is represented. A mutual coupling model between antennas is established and a coupling matrix C is defined, where Cij represents the coupling coefficient between antenna i and antenna j. Considering the frequency dependence, the coupling matrix is ​​represented as C(f). At the same time, an environmental scattering model is established and statistical parameters are used to describe the multipath effect. The system model parameters are integrated into structured data, including antenna position, initial polarization state, coupling matrix, and environmental scattering parameters.

[0107] Construct a distributed optimization algorithm. Divide the multi-antenna system into multiple subsystems, each of which contains several adjacent antennas. Define the local objective function Ji(Si)=w1*SNRi+w2*SIRi-w3*BERi, where S i is the set of polarization states of the ith subsystem. Implement a distributed optimization algorithm based on the alternating direction method of multipliers (ADMM). Define global variables z and dual variables u, and construct an augmented Lagrangian function: L(S, z, u) = Σ i Ji(S i )+(ρ / 2)Σ i ||S i -z i + u i || 2. Iterate the following steps: (1) For each subsystem, solve the local optimization problem: S i k+1 =argmin Si [Ji(S i )+(ρ / 2)||S i -z i k +u i k || 2 ]; (2) Update global variables: z k+1 = (1 / N)Σ i (S i k+1 +u i k );(3) Update the dual variable: u i k+1 =u i k +S i k+1 -z i k+1 . Where ρ is the penalty parameter and k is the number of iterations. The converged z is the optimal global polarization state.

[0108] Based on the system model, calculate the spatial correlation matrix R for each antenna pair. Perform eigenvalue decomposition: R = U ∧ U H , where ∧ represents the diagonal eigenvalue matrix and U represents the eigenvector matrix. Select the eigenvector corresponding to the maximum eigenvalue as the main polarization direction and the orthogonal eigenvector as the diversity polarization direction. Assign a polarization state to each antenna so that the polarization states of adjacent antennas are as orthogonal as possible. Specifically, a greedy algorithm is used: (1) Assign the main polarization direction to the first antenna; (2) Assign polarizations to other antennas in turn, and select the direction that is most orthogonal to the polarization of the assigned antenna. Convert the assigned polarization state into a Stokes parameter representation. Convert the Stokes parameters into actual control parameters, such as attenuation codes and phase shift codes.

[0109] Define the desired beam direction (θ0, φ0) and beam width Δθ. Calculate the complex weights for each antenna: w i =exp(-j2π(x i sinθ0cosφ0+y i sinθ0sinφ0+z i*cosθ0) / λ), where λ is the operating wavelength. The beamforming weights are combined with the polarization control parameters to generate a composite weight matrix W. For each antenna i, its composite weight is Wi = wi*[cos(ψi / 2), sin(ψi / 2)*exp(jΔi)], where ψi and Δi are the ellipsoid angle and phase difference, respectively. The composite weights are optimized using an adaptive algorithm (such as MVDR, Minimum Variance Distortionless Response) to suppress interference and noise. The optimized composite weights are converted into amplitude and phase control instructions for each antenna and integrated with the polarization control instructions.

[0110] This embodiment improves the performance and flexibility of the antenna system in three dimensions of space, frequency and polarization by implementing a multi-antenna cooperative polarization control strategy. It solves the problem that the traditional single-antenna polarization control has limited capabilities when facing a complex spatial electromagnetic environment. A multi-antenna system model is established, including the position and initial polarization state of each antenna, which lays a theoretical foundation for subsequent cooperative control. By accurately modeling the spatial position and polarization characteristics of each antenna element, the system can fully grasp the electromagnetic characteristics of the entire antenna array and provide accurate input for the optimization algorithm. A distributed optimization algorithm is constructed to coordinate the polarization states of multiple antennas, which allows the system to achieve global optimal polarization control while reducing computational complexity. By decomposing the optimization problem into multiple sub-problems, each sub-problem corresponds to an antenna or antenna subarray, and then reaching the global optimum through an iterative coordination process, the scalability and computational efficiency of the system are improved. A spatial diversity strategy is implemented, and different polarization states are used at different spatial locations, making full use of the spatial freedom of the multi-antenna system. By using different polarization states at different spatial locations, the system can effectively resist spatial selective fading and improve the reliability of the communication link. At the same time, this embodiment can also improve the spatial multiplexing capability of the system, supporting multiple data streams simultaneously in the same frequency band, thereby improving spectrum utilization efficiency. Integrated beamforming technology combines polarization control and beam control. Combining polarization control with spatial beam control achieves higher-dimensional signal processing capabilities. By simultaneously optimizing the spatial pattern and polarization state of the antenna array, the system can focus on the target signal more accurately while effectively suppressing interference from other directions. This multi-dimensional signal processing capability is particularly important in complex multi-user, multipath propagation environments, and can improve the system's signal-to-interference ratio and signal-to-noise ratio.

[0111] This embodiment has many advantages over traditional single-antenna or non-cooperative multi-antenna systems. It enhances the spatial selectivity and anti-interference capability of the system. By cooperatively controlling the polarization states of multiple antennas, the system can form a complex spatial-polarization response pattern to accurately capture the desired signal while effectively suppressing interference. This can reduce co-channel interference and increase network capacity in densely deployed wireless networks; it improves the spectrum utilization efficiency of the system. Through the combination of spatial multiplexing and polarization multiplexing, the system can support more concurrent data streams on the same frequency resources, which is crucial to meet the growing demand for data transmission. This collaborative control strategy enhances the adaptability and robustness of the system. Faced with a complex and changeable electromagnetic environment, the system can flexibly adjust the polarization state of each antenna, quickly adapt to changes in channel conditions, and maintain optimal communication performance.

[0112] In practical applications, in 5G and future 6G mobile communication systems, massive MIMO technology is key. By realizing cooperative polarization control of antenna arrays, the capacity and reliability of MIMO systems can be further improved. In high-density urban environments, this embodiment can effectively reduce signal fading and interference caused by building reflections and multipath effects. In the field of satellite communications, ground station antenna arrays can more accurately track satellite signals by adopting this cooperative control strategy, while effectively suppressing ground interference and improving communication quality and reliability. In radar systems, target detection and recognition capabilities can be improved, especially in complex electromagnetic environments and low observable targets. This embodiment also opens up new possibilities for cognitive radio and smart antenna systems. By sensing and learning the electromagnetic environment in real time, the system can dynamically adjust the polarization state and beam direction of each antenna to achieve truly intelligent and adaptive communication. This not only improves communication performance, but also makes more efficient use of spectrum resources and reduces interference to other users.

[0113] This embodiment not only improves the performance and adaptability of the antenna system in complex environments, but also provides strong technical support for future intelligent communication systems. It represents an important trend of antenna technology developing towards higher dimensions and more intelligent directions, and has the potential to change the design and deployment of future wireless communication systems. By organically combining polarization control, spatial processing and intelligent algorithms, it provides a powerful tool for solving increasingly complex wireless communication challenges and paves the way for more efficient and reliable communication systems.

[0114] According to one aspect of the present application, step S3 may also be:

[0115] S3a, obtaining target polarization parameters input by the user, including ellipticity and azimuth, and calculating a normalized electric field component based on the target polarization parameters; and calculating a Stokes parameter vector based on the electric field component;

[0116] S3b, obtaining the current operating frequency, and based on the current operating frequency, extracting material property data from pre-stored frequency-related parameters; based on the material property data, calculating the birefringence effect and polarization-related loss of the system; based on the birefringence effect and polarization-related loss of the system, constructing a Mueller matrix and normalizing it; wherein the diagonal elements of the Mueller matrix represent attenuation, and the non-diagonal elements represent polarization conversion;

[0117] S3c, multiplying the Stokes parameter vector and the Mueller matrix to obtain a product result; normalizing the product result to obtain a normalized product result; extracting new ellipticity and azimuth information based on the normalized product result; converting the new ellipticity and azimuth information into the required amplitude ratio and phase difference of the horizontal and vertical components;

[0118] S3d. Based on the required amplitude ratio of the horizontal and vertical components, the target amplitude values ​​required for the horizontal and vertical channels are calculated respectively; based on the target amplitude values ​​required for the horizontal and vertical channels, in the "amplitude value-attenuation code" lookup table, a binary search algorithm is used to find the attenuation code closest to the target amplitude value as the required attenuation code; based on the required phase difference between the horizontal and vertical components, in the "phase difference value-phase shift code" lookup table, a binary search algorithm is used to find the phase shift code closest to the target phase difference as the required phase shift code.

[0119] In one embodiment of the present application, the Stokes parameters of the target polarization state are calculated. The target polarization parameters input by the user are received, including ellipticity and azimuth. The normalized Jones vector is calculated using the polarization parameters. The complete Stokes parameter vector is calculated based on the Jones vector, including four components S0, S1, S2 and S3. The calculated Stokes parameter vector is stored as a four-dimensional vector for subsequent processing.

[0120] Construct the Mueller matrix of the system. Extract material property data from pre-stored frequency-dependent parameters based on the current operating frequency. Calculate the birefringence effect and polarization-dependent loss of the system. Use these parameters to construct a 4x4 Mueller matrix, where the diagonal elements represent attenuation and the off-diagonal elements represent polarization conversion. Normalize the constructed Mueller matrix to ensure that it meets the principle of energy conservation.

[0121] Perform a Mueller matrix transformation. Multiply the Stokes parameter vector with the Mueller matrix. Normalize the product to ensure that the total intensity remains unchanged. Extract new ellipticity and azimuth information from the transformed Stokes parameters. Convert this information into the required amplitude ratio and phase difference of the horizontal and vertical components.

[0122] Determine the attenuation code and phase shift code. According to the obtained amplitude ratio, calculate the target amplitude values ​​required for the horizontal and vertical channels respectively. In the "amplitude value-attenuation code" lookup table, use the binary search algorithm to find the attenuation code closest to the target amplitude value. According to the phase difference obtained in the step, in the "phase difference value-phase shift code" lookup table, also use the binary search algorithm to find the phase shift code closest to the target phase difference. Combine the determined attenuation code and phase shift code into a control instruction and store it as structured data for subsequent polarization control.

[0123] According to one aspect of the present application, the method further includes S6, converting the final polarization control parameter into a hardware control command, and generating a performance evaluation report, specifically:

[0124] S61, obtaining hardware specifications, and based on the hardware specifications, converting the final polarization control parameters into voltage or current values; based on the voltage or current values, generating a standardized control instruction sequence; packaging the control instruction sequence, and adding a timestamp and verification information;

[0125] S62, establishing a communication link with the polarization simulation hardware, sending the control instruction sequence to the polarization simulation hardware and executing it according to a predefined protocol; monitoring the execution status of the control instruction sequence to ensure that each antenna receives and executes it correctly; and recording the timestamp of the execution completion;

[0126] S63, using a preconfigured performance measurement device, collecting key performance indicators within a predefined time window; calculating statistical characteristics of the key performance indicators; associating the statistical characteristics with the timestamp of the completion of the execution of step S62; and recording environmental parameters, including temperature and humidity;

[0127] S64. Based on statistical characteristics and environmental parameters, calculate performance improvement indicators and generate a performance evaluation report, including the numerical values, trend charts and analysis results of each indicator; store the performance evaluation report as structured data, and extract key parameters as input for the next round of optimization.

[0128] In one embodiment of the present application, the final polarization control parameters are received. According to the hardware specifications, the attenuation code and the phase shift code are converted into specific voltage or current values. A standardized control instruction sequence is generated. The control instruction sequence is packaged and timestamps and verification information are added. A communication link with the polarization simulation hardware is established. The generated control instruction sequence is sent to the hardware according to a predefined protocol. The instruction execution status is monitored to ensure that each antenna correctly receives and executes the control command. The timestamp of the execution completion is recorded. The performance measurement equipment is configured and the measurement parameters are set. Key performance indicators such as SNR, SIR, BER, etc. are collected within a predefined time window. The statistical characteristics of these indicators, such as mean and variance, are calculated. Environmental parameters such as temperature and humidity are recorded. All measurement data are associated with the execution timestamp. Performance data and environmental parameters are integrated. Performance improvement indicators are calculated, such as the percentage of SNR improvement relative to the previous round of optimization. A detailed performance evaluation report is generated, including the numerical values, trend charts and analysis results of each indicator. The report is stored as structured data, and key parameters are extracted as input for the next round of optimization to complete closed-loop control.

[0129] According to another aspect of the present application, it also includes: S7, realizing environmental adaptive polarization simulation. Establishing an electromagnetic environment perception module to collect environmental parameters. Using machine learning technology to build an environmental prediction model. Dynamically adjusting polarization simulation parameters based on the prediction model. Implementing an intelligent frequency selection mechanism to optimize system performance. Specifically:

[0130] Step S71, establish an electromagnetic environment perception module. Configure the broadband receiver, set the frequency sweep range and step. Perform periodic spectrum scanning and record the signal strength of each frequency point. Analyze the scan data to identify the frequency and strength of the main interference source. At the same time, measure the ambient noise floor and time-varying characteristics. Store the collected environmental parameters as time series data, including information such as frequency, signal strength, interference characteristics and noise level.

[0131] Step S72, build an environmental prediction model. Preprocess the time series data collected in step S71, including detrending, standardization and time alignment. Use a long short-term memory network (LSTM) to build a prediction model, with the input being a historical environmental parameter sequence and the output being a prediction of environmental parameters at future moments. Use a sliding window method to train the model to minimize the prediction error. Update the model regularly with newly collected data. Save the trained model parameters for subsequent predictions.

[0132] Step S73, dynamically adjust polarization simulation parameters. Use the prediction model of step S72 to generate environmental parameter predictions for a period of time in the future. Calculate the optimal polarization state based on the predicted environmental parameters. Consider the time-varying characteristics of the environment and design a smooth transition strategy to avoid drastic changes in the polarization state. Generate a series of timestamps and corresponding polarization control instructions, store them as structured data, and allow the polarization simulation device to execute them in sequence.

[0133] Step S74, implement the intelligent frequency selection mechanism. Based on the spectrum scanning data of step S71, identify the available frequency channels. For each available channel, use the prediction model of step S72 to evaluate its quality in the future. Combine the bandwidth requirements and polarization characteristics of the system, and use the multi-attribute decision method to select the optimal operating frequency. Generate frequency switching instructions, and update the polarization simulation parameters accordingly. Combine this information into a control instruction sequence and store it as structured data.

[0134] In one embodiment of the present application, a broadband receiver is configured, and the frequency sweep range [fmin, fmax] and the frequency step Δf are set. A periodic spectrum scan is performed to record the signal power spectrum density P(fi) at each frequency point fi. An energy detection algorithm is used to identify the main signal and the interference source: if P(fi)>γ, it is marked as an active signal, where γ is an adaptive threshold. At the same time, the ambient noise floor N0 and time-varying characteristics are measured, and the short-time Fourier transform (STFT) is used to analyze the time-frequency characteristics of the signal. Key features are extracted, such as the signal-to-noise ratio SNR(fi)=P(fi) / N0, channel occupancy, and signal bandwidth. The collected environmental parameters are stored as time series data {t, f, P(f), SNR(f), ...} for subsequent analysis.

[0135] Preprocess the collected time series data, including detrending, standardization, and time alignment. Use wavelet transform for multi-scale analysis to separate the characteristics of changes at different time scales. Build a deep learning model, such as a long short-term memory network (LSTM). Define the input feature vector X=[f, P(f), SNR(f), ...], and the output is the predicted value Y'=[P'(f), SNR'(f), ...] at the future time t+Δt. Use the sliding window method to generate training samples with a window size of Tw and a prediction step of Δt. Train the model to minimize the prediction error: L =Σ||Y-Y'|| 2 , where Y represents the true value. Use the Adam optimizer to update the model parameters. Regularly update the model with newly collected data, using an online learning strategy. Save the trained model parameters, including the network structure and weights, for subsequent predictions.

[0136] Use the prediction model to generate a prediction sequence of environmental parameters in the future time window [t, t+T]. Based on the predicted environmental parameters, calculate the optimal polarization state sequence {S(t), S(t+Δt), ..., S(t+T)}. Consider the time-varying characteristics of the environment and design a smooth transition strategy. Use the cubic spline interpolation method to construct a continuous conversion curve S(τ), τ∈[t, t+T] between discrete optimal polarization states. At the same time, consider the hardware response characteristics, such as the response time and accuracy limitations of the actuator, and constrain and optimize the conversion curve. Uniformly sample on the time axis to generate a discrete control instruction sequence. Associate the generated polarization control instruction sequence with a timestamp and store it as structured data for the polarization simulation device to execute in time sequence.

[0137] Based on the spectrum scanning data, the available frequency channel set F = {f1, f2, ..., fK} is identified. For each available channel fi, a prediction model is used to evaluate its quality indicators for a period of time [t, t+T] in the future, such as the expected signal-to-noise ratio E[SNR(fi, τ)], τ∈[t, t+T]. Combined with the bandwidth requirements B and polarization characteristics of the system, a multi-attribute decision matrix D is constructed, where Dij represents the score of frequency fi on attribute j (such as signal-to-noise ratio, bandwidth, and polarization matching). The weight vector w of each attribute is determined using the analytic hierarchy process (AHP). The TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method is applied to calculate the proximity Ci of each frequency to the ideal solution. The frequency with the highest proximity is selected as the optimal operating frequency f*. Frequency switching instructions are generated and the polarization simulation parameters are updated accordingly. The frequency selection results and related decision bases are combined into a control instruction sequence and stored as structured data.

[0138] According to another aspect of the present application, step S8 is also included to achieve high-precision phase control. A multi-stage phase-locked loop (PLL) system is constructed to improve phase resolution. An interpolation algorithm is implemented to achieve smooth transition between discrete phase points. An adaptive phase correction mechanism is constructed to compensate for system drift. Multiple phase control technologies are integrated to achieve high-precision and wide-range phase adjustment. Specifically:

[0139] Step S81, construct a multi-stage PLL system. Define the PLL cascade structure, including a reference oscillator, a phase detector, a loop filter, and a voltage-controlled oscillator. Design parameters for each stage of PLL, including the frequency division ratio, loop bandwidth, and phase margin. Use the state space method to establish a mathematical model of the entire system and calculate the system transfer function. Based on the transfer function, analyze the stability and dynamic characteristics of the system. Adjust the PLL parameters through the numerical optimization method to achieve the best phase noise performance and locking time for the system.

[0140] Step S82, implement the phase interpolation algorithm. Based on the PLL system output of step S81, obtain discrete phase points. Use the cubic spline interpolation method to construct a continuous phase curve between adjacent discrete phase points. According to the target phase value, calculate the corresponding control voltage on the interpolation curve. In order to improve the accuracy, use the Newton iteration method to fine-tune the interpolation result. Convert the calculated control voltage into a digital control word and store it as a lookup table for fast query.

[0141] Step S83, construct an adaptive phase correction mechanism. Implement a phase detection module to periodically measure the deviation between the system output phase and the ideal phase. Use the Kalman filter algorithm to estimate the trend of phase deviation and the system drift rate. Based on the estimation results, calculate the required compensation. Design an adaptive control law to dynamically adjust the parameters of the PLL system, such as loop gain and integral time constant. Convert the compensation into control instructions and update the configuration of the PLL system in real time.

[0142] Step S84, integrate multiple phase control technologies. Use the multi-stage PLL system of step S81 as a coarse adjustment mechanism. Use a digitally controlled phase shifter as a fine adjustment mechanism, whose control parameters are generated by the interpolation algorithm of step S82. Implement phase modulation based on direct digital synthesis (DDS) for fast phase adjustment. Design decision logic to select the most suitable control method based on the target phase and accuracy requirements. Integrate the adaptive correction mechanism of step S83 to compensate for the errors of various control methods in real time. Encode the integrated control strategy into a state machine and store it as executable program instructions.

[0143] In one embodiment of the present application, a PLL cascade structure is defined, including N-stage PLLs. Each stage of PLL includes a phase detector, a loop filter, a voltage-controlled oscillator (VCO), and a frequency divider. The parameters of each stage of PLL are designed, including a reference frequency fref i , division ratio Ni, loop bandwidth BW i and phase margin PM i . Use the state space method to establish the mathematical model of the entire system. Define the state vector x = [φ1, φ2, ..., φN, ω1, ω2, ..., ωN] T , where φi and ωi are the phase and frequency errors of the i-th PLL, respectively. Establish the state equation dx / dt =Ax +Bu and the output equation y =Cx +Du, where A, B, C, and D are system matrices. Calculate the system transfer function H(s) = C(sI-A) -1B+D. Analyze the stability and dynamic characteristics of the system, such as lock time and phase noise performance. Use an optimization algorithm (such as particle swarm optimization PSO) to adjust the PLL parameters to minimize the phase noise integral: J=∫L(f)df, where L(f) is the single-sideband phase noise power spectral density. Finally, store the optimized PLL parameters as a configuration file containing all the settings for each PLL stage.

[0144] Based on the PLL system output, a discrete phase point set Φ = {φ1, φ2, ..., φM} is obtained. A continuous phase curve is constructed between adjacent discrete phase points using the cubic Hermite interpolation method. Define the interpolation polynomial: φ(t) = a0 + a1t + a2t 2 +a3t 3 , where the coefficients a0, a1, a2, and a3 are determined by the phase values ​​and derivative values ​​of the adjacent points. For the target phase value φ target , use Newton iteration method on the interpolation curve to solve the corresponding control voltage: V = V0 -f(V0) / f'(V0), where f(V) φ(V) -φ target , f'(V) is its derivative. Iterate until convergence or the maximum number of iterations is reached. To improve accuracy, adaptive step size control is implemented: if the relative error of two consecutive iterations is less than the threshold ε, the step size is reduced. The calculated control voltage is converted into a digital control word, and a mapping table of "target phase-control word" is established and stored as a lookup table data structure for fast query.

[0145] Implement the phase detection module and use an orthogonal phase detector to measure the deviation Δφ(t) between the system output phase and the reference phase. Design a Kalman filter with a state vector of x = [φ, ω, a] T , where φ is the phase deviation, ω is the frequency deviation, and a is the frequency drift rate. Define the state transfer equation x(k+1) =Fx(k)+w(k) and the observation equation z(k)=Hx(k)+v(k), where F and H are the state transfer matrix and the observation matrix, respectively, and w(k) and v(k) are the process noise and observation noise. Perform the prediction and update steps of the Kalman filter to estimate the changing trend of the phase deviation and the system drift rate in real time. At the same time, an adaptive algorithm is used to dynamically adjust the noise covariance matrices Q and R to adapt to different measurement conditions. Specifically, an innovative sequence method is adopted: Q(k)=λQ(k-1)+(1-λ)(K(k)v(k)v(k) T K(k) T), where λ is the forgetting factor, K(k) is the Kalman gain, and v(k) is the innovation sequence. Based on the estimation results, an adaptive control law is designed: ΔV(k) = KpΔφ(k) + Ki∫Δφ(t)dt + Kd*dΔφ(k) / dt, where Kp, Ki, and Kd are PID controller parameters. A fuzzy logic regulator is used to adaptively adjust the PID parameters to cope with different phase deviation conditions. The compensation ΔV(k) is converted into a control command to update the configuration of the PLL system in real time.

[0146] Use a multi-stage PLL system as a coarse adjustment mechanism to cover a wide range of phase adjustments. Use an interpolation algorithm as a fine adjustment mechanism to achieve high-precision adjustment between discrete phase points. Implement phase modulation based on direct digital synthesis (DDS) for fast phase adjustment. The phase accumulator update formula of DDS is: θ(n+1)=(θ(n) +Δθ)mod2π, where Δθ is the phase increment and mod represents the modulo operation. Design decision logic to select the most suitable control method based on the target phase and accuracy requirements. Define decision rules: If |Δφ|>φ threshold_1 , use PLL coarse tuning; if φ threshold_2 <|Δφ|≤φ threshold_1 , use the interpolation algorithm to fine-tune; if |Δφ|≤φ threshold_2 , use DDS to quickly adjust. threshold_1 and φ threshold_2 The control strategy is encoded as a state machine, which defines the state transition conditions and corresponding actions. The state machine logic is converted into executable program instructions and stored as firmware code to run in the embedded system.

[0147] According to another aspect of the present application, it also includes achieving comprehensive performance optimization. Establish a system performance evaluation model, including indicators such as polarization accuracy, adjustment speed and environmental adaptability. Design a multi-objective optimization algorithm to balance various performance indicators. Implement an adaptive parameter adjustment mechanism to dynamically optimize system parameters based on real-time performance feedback. Establish a long-term learning mechanism, accumulate optimization experience, and continuously improve system performance. Specifically:

[0148] Establish a system performance evaluation model. Define a set of performance indicators, including polarization accuracy, adjustment speed, environmental adaptability, power consumption, etc. Design a quantification method for each indicator, such as using the axial ratio error of the polarization ellipse to evaluate polarization accuracy and using the step response time to evaluate adjustment speed. Implement a data acquisition program to collect raw performance data from each subsystem. Use the principal component analysis (PCA) method to reduce the correlation between indicators. Construct a comprehensive performance scoring model to weight the sum of various indicators, and the weight coefficient is determined through expert experience and historical data analysis.

[0149] Construct a multi-objective optimization algorithm. Define multiple objective functions based on the performance evaluation model. Use the multi-objective particle swarm optimization (MOPSO) algorithm to initialize the particle swarm, where each particle represents a set of system parameters. Evaluate the fitness of each particle and update the individual optimal solution and the global optimal solution. Use the non-dominated sorting method to maintain the Pareto optimal solution set. Iterate the optimization process and update the particle speed and position. Select the best parameter combination from the Pareto front as the operating point of the system.

[0150] Implement an adaptive parameter adjustment mechanism. Establish a real-time performance monitoring module based on the performance evaluation model. Design a fuzzy controller with performance indicator deviation as input and parameter adjustment as output. Define a set of fuzzy rules, such as "if polarization accuracy is low and adjustment speed is fast, increase control gain". Implement a fuzzy inference engine to perform fuzzification, rule evaluation, and defuzzification. Perform parameter adjustments periodically and evaluate performance changes after each adjustment. Use reinforcement learning methods to update the weights of fuzzy rules based on the adjustment effect to achieve adaptive optimization of the control strategy.

[0151] Establish a long-term learning mechanism. Design a knowledge base structure to store historical optimization experience, including environmental conditions, system configuration, performance data, and optimization strategies. Implement a case matching algorithm to retrieve similar cases from the knowledge base based on the current environment and performance requirements. Use genetic programming methods to generate new optimization strategies based on the retrieved cases. Perform strategy evaluation and screening to add effective strategies to the knowledge base. Regularly clean up and reorganize the knowledge base to delete outdated or redundant cases. Implement a distributed learning mechanism to allow multiple system instances to share optimization experience and accelerate the learning process.

[0152] In another embodiment of the present application, a method for simulating antenna polarization based on phase and amplitude calibration simulates electromagnetic wave waveforms of different shapes and directions through amplitude and phase components in the horizontal and vertical directions, thereby realizing instantaneous change of polarization form without mechanically adjusting the direction and antenna structure. By calibrating the phase and amplitude in real time, the accuracy of polarization characteristic measurement is improved, and the error caused by amplitude to polarization phase angle is eliminated. Through a calibration method, the errors of amplitude and phase caused by different hardware materials and qualitative changes of hardware over time are eliminated, and the accuracy of planned characteristic measurement is improved. This calibration method can simply improve the granularity of calibration through parameter setting, and through multiple iterations, the measurement accuracy of amplitude and phase will be further improved.

[0153] The vector network analyzer can be used to measure amplitude and phase at the same time, and realize simultaneous calibration. The amplitude calibration is calibrated and controlled separately on the horizontal component and the vertical component, and the phase is calibrated using the difference between the horizontal and vertical components to obtain the initial horizontal and vertical phase difference.

[0154] The amplitude is controlled by the attenuation code, and each attenuation code corresponds to an amplitude value. Therefore, if you want to control a certain amplitude value, you need to get the corresponding attenuation code. Similarly, the phase difference is controlled by the phase shift code, and each phase shift code corresponds to a phase difference value. Therefore, if you want to control a certain phase difference value, you need to get the corresponding phase shift code.

[0155] Different power attenuation code values ​​are sent to the polarization simulation device in a loop on the horizontal and vertical channels respectively, and the frequency value of the internal signal source of the vector network analyzer is circulated to obtain the amplitude value generated by the device under different frequencies and different attenuation codes on the horizontal and vertical channels in turn, and the original data "attenuation code-amplitude value" table is generated. Similarly, phase shift code values ​​are sent to the polarization simulation device in a loop on the horizontal and vertical channels, and the frequency value of the internal signal source of the vector network analyzer is circulated to obtain the phase value generated by the device under different frequencies and different phase shift codes on the horizontal and vertical channels in turn, and the original data "phase shift code-phase value" table is generated. By subtracting the horizontal and vertical, the "phase shift code-phase difference value" table is obtained. A calibration algorithm is designed to realize the reverse calculation of "amplitude value-attenuation code" and "phase difference value-phase shift code".

[0156] At a certain frequency, in the required amplitude value range (including starting amplitude, ending amplitude, and amplitude step), for a certain amplitude value G, in the "attenuation code-power value" table, find the attenuation code closest to the G value at that frequency. Loop in sequence to obtain the attenuation codes corresponding to all power values, and finally obtain the "power value-attenuation code" table. Similarly, loop again to obtain the "power value-attenuation code" table of other frequencies and store them separately. Similarly, at a certain frequency, in a given phase difference value range (including starting phase, ending phase, and phase step), for a certain phase difference value P, in the "phase shift code-phase difference value" table, find the phase shift code closest to the P value at that frequency. Loop in sequence to obtain the phase shift codes corresponding to all phase difference values, and finally obtain the "phase difference value-phase shift code" table. Similarly, loop again to obtain the "phase difference value-phase shift code" table of other frequencies and store them separately.

[0157] In the reverse calculation algorithm, due to the required amplitude or phase difference range of the loop search and the code value of the loop search, when the granularity of the loop is fine, the two layers of loops are time-consuming. This embodiment uses a sorting algorithm combined with a binary search algorithm to reduce the calculation time. That is, all amplitude values ​​and phase difference values ​​are sorted at a certain frequency, and then the closest amplitude and phase are searched by binary search to obtain the corresponding code value, thereby improving the search efficiency.

[0158] By looking up the table obtained from the above calibration, the horizontal component E can be obtained by controlling the code value. x The amplitude E xm , vertical component E y The amplitude Eym , the phase difference is ΔΦ=Φ y -Φ x , thereby simulating linear, circular, elliptical and other polarization simulation waveforms.

[0159] If the horizontal component E x and the vertical component E y The phase difference is the same or 180°, that is, ΔΦ=Φ y -Φ x =0 or π, the composite wave is a directional polarized wave. The electric field strength of the composite wave is E=sqrt(E xm 2 +E ym 2 )cos(wt+Φ x ), where Φ y Represents the vertical component E y The phase, Φ x Represents the horizontal component E x The phase of , w represents the angular frequency, and t represents the time.

[0160] If the horizontal component E x and the vertical component E y The amplitudes are equal, that is, E xm =E ym =E m , but the phase difference is 90°, that is, ΔΦ=Φ y -Φ x =0 or π / 2, the synthesized electric field wave is a circularly polarized wave. Therefore, E x =E m cos(wt+Φ x ), E y =E m cos(wt+Φ x +π / 2) is E=sqrt(E x 2 +E y 2 ) = E m .

[0161] If ΔΦ=Φ y -Φ x When it is not equal to 0 or +π and -π and +π / 2 or -π / 2, or ΔΦ=Φ y -Φ x =π / 2 but E xm ≠E ym When , the synthesized electric field wave is an elliptically polarized wave. The synthesized electromagnetic field strength is E=sqrt(E xm 2 +E ym2 )cos(wt+Φ x +ΔΦ).

[0162] In one embodiment of the present application, the connection of the device is performed. Figure 7 As shown, the polarization simulation device and the vector network analyzer are connected to the calibration and polarization simulation software through a network switch. The output end of the vector network analyzer is connected to the polarization simulation device through an RF line, and the horizontal or vertical output channel of the polarization simulation device is connected to the input end of the vector network analyzer through an RF line. Figure 8 As shown. Open the calibration part of the calibration and polarization simulation software. Cycle frequency, cycle amplitude and phase code value, where the cycle frequency controls the frequency of the internal signal source of the vector network analyzer, and the cycle code value controls the attenuation code and phase shift code of the polarization simulation device, and controls the vector network analyzer to read the corresponding amplitude value and phase value. Until the end of the cycle. Follow the above steps to complete the reading of the amplitude value and phase value in the horizontal and vertical directions in turn. Get the original data of "attenuation code-amplitude value" in the horizontal and vertical directions and the original data of "phase shift code-phase difference value" of the horizontal and vertical phase difference.

[0163] Through the calibration algorithm, the calibration of the original data is completed, and the calibrated data of "amplitude value-attenuation code" (taking 13000MHz frequency as an example) and "phase difference value-phase shift code" are obtained. The polarization simulation function of the calibration and polarization simulation software is turned on. By searching the calibrated data, the amplitude value and phase difference value in the horizontal and vertical directions are controlled by the code value, and the linear polarization simulation, circular polarization simulation, and elliptical polarization simulation experiments are completed through the polarization simulation algorithm.

[0164] Through experiments, the polarization measurement results are as follows: at the frequency of 13000MHz, when the phase difference code is controlled to be 61, the phase difference is 68.1, the corresponding phase value is 67.5, and the phase difference accuracy error is within 1 degree. When the horizontal amplitude code value is controlled to be 121, the horizontal amplitude is -41.9dBm, and the corresponding amplitude value is -42dBm. When the vertical amplitude code value is controlled to be 117, the vertical amplitude is -40.3dBm, and the corresponding amplitude value is -40dBm. The amplitude measurement accuracy error is within 0.5dBm. The use of the present invention to reduce the phase step and increase the number of phase code values, or to reduce the amplitude step and increase the number of amplitude code values, will further improve the accuracy and obtain better accuracy effects.

[0165] The preferred embodiments of the present invention are described in detail above; however, the present invention is not limited to the specific details in the above embodiments. Within the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all belong to the protection scope of the present invention.

Claims

1. An antenna polarization simulation method based on phase and amplitude calibration, characterized in that: The steps include: S1. Collecting raw data using a vector network analyzer, and preprocessing the raw data to obtain preprocessed raw data; wherein the raw data includes amplitude data and phase data; S2. Based on the preprocessed raw data, an initial calibration model is constructed, and the initial calibration model is iteratively optimized to minimize the error between the measured value and the theoretical value to obtain the final calibration parameters; Using the final calibration parameters, generate a post-calibration lookup table, including an "amplitude value-attenuation code" lookup table and a "phase difference value-phase shift code" lookup table; S3, obtaining the target polarization parameters and calculating the corresponding Stokes parameters; Based on the final calibration parameters, the Mueller matrix of the system is constructed; the Stokes parameters are transformed using the Mueller matrix to obtain the output polarization state; According to the output polarization state, the post-calibration lookup table is queried to determine the required attenuation code and the required phase shift code as the ideal polarization state; S4, build a closed-loop control system, monitor the quality of the received signal in real time, and obtain the monitoring results; Based on the monitoring results, a polarization parameter adjustment strategy is constructed using a reinforcement learning algorithm. Based on the polarization parameter adjustment strategy, multi-objective optimization is performed to obtain the optimization results; Based on the optimization results and the ideal polarization state, the optimized polarization parameters are generated; S5. Obtain the location information of each antenna, and build a multi-antenna system model based on the location information of each antenna and the optimized polarization parameters; based on the multi-antenna system model, use a distributed optimization algorithm to coordinate the polarization states of predetermined antennas; based on the coordination result, implement a spatial diversity strategy and generate polarization control parameters that take spatial diversity into consideration; based on the polarization control parameters, use beamforming technology to generate the ultimate control parameters of each antenna; Step S3 is further as follows: S31, obtaining target polarization parameters input by a user, including ellipticity and azimuth, and calculating a normalized Jones vector based on the target polarization parameters; and calculating a Stokes parameter vector based on the Jones vector; Store the Stokes parameter vector as a four-dimensional vector; S32, constructing a frequency-dependent Mueller matrix model based on the final calibration parameters; using the Mueller matrix model, calculating the Mueller matrix elements of each key frequency point; Based on the Mueller matrix elements, a continuous frequency-dependent Mueller matrix function is constructed using an interpolation method; S33, obtaining and based on the current operating frequency, obtaining an applicable Mueller matrix in a continuous frequency-dependent Mueller matrix function; multiplying the Stokes parameter vector by the Mueller matrix to obtain a product result; S34, converting the product result to obtain a converted parameter; Based on the converted parameters, the post-calibration lookup table is queried to determine the attenuation code and phase shift code required for each antenna as the ideal polarization state of each antenna.

2. The antenna polarization simulation method based on phase and amplitude calibration according to claim 1, characterized in that: Step S1 is further as follows: S11. According to a predefined frequency range and step value, the vector network analyzer is configured, including frequency scanning settings, output power, intermediate frequency bandwidth and average factor of the vector network analyzer, to obtain a configured vector network analyzer; and the configured vector network analyzer is calibrated to obtain a calibrated vector network analyzer. S12. At each frequency point, cyclically set the attenuation code and phase shift code of the polarization simulation device; for each attenuation code, use a calibrated vector network analyzer to measure the amplitude data of the horizontal channel and the vertical channel; For each phase-shift code, the phase data of the horizontal channel and the vertical channel are measured using a calibrated vector network analyzer; the amplitude data and the phase data are saved as original data; S13, based on the original data, using the moving window method to calculate the local mean and standard deviation of each data point, and marking the data point that deviates from the local mean by more than a predetermined number of standard deviations as an outlier; Using interpolation to replace outliers, obtaining original data with outliers removed; based on the original data with outliers removed, applying a low-pass filter to remove high-frequency noise, obtaining original data with noise removed; The original data after noise removal is subjected to data standardization processing to obtain preprocessed original data.

3. The antenna polarization simulation method based on phase and amplitude calibration according to claim 2, characterized in that: Step S2 is further as follows: S21, based on the preprocessed raw data, extracting the original amplitude value, phase value, attenuation code and phase shift code; based on the original amplitude value, phase value, attenuation code and phase shift code, constructing an initial calibration model to obtain the calibrated amplitude and phase values; S22, calculating the mean square error between the calibrated amplitude and phase values ​​and the preset theoretical values ​​as the objective function; Based on the objective function, the Levenberg-Marquardt algorithm is used to iteratively optimize the parameters until the preset convergence conditions are reached, and the final optimized parameter values ​​are obtained as the final parameters of the calibration model; S23, based on the final parameters of the calibration model, calibrating the original amplitude value and phase value to obtain the final calibrated amplitude and phase value; S24, based on the final calibrated amplitude, generating a target amplitude value within a predefined amplitude range with a preset fixed step size; For each target amplitude value, the corresponding attenuation code is calculated using the inverse function of the calibration model to generate an "amplitude value-attenuation code" lookup table; S25, based on the final calibrated phase value, generating a target phase difference value with a preset fixed step size within a predefined phase range; For each target phase difference value, the corresponding phase shift code is calculated using the inverse function of the calibration model to generate a "phase difference value-phase shift code" lookup table.

4. The antenna polarization simulation method based on phase and amplitude calibration according to claim 3, characterized in that: Step S4 is further as follows: S41, based on predefined comprehensive performance indicators, using a preconfigured signal quality monitoring module, collecting performance indicator data in real time, including signal-to-noise ratio, signal-to-interference ratio, and bit error rate; S42, constructing a state space based on the performance indicator data and the current polarization parameters; Constructing action space based on preset adjustable attenuation code and phase shift code range; Based on the state space and action space, implement the Q-learning algorithm to obtain the updated Q table; Based on the values ​​in the Q table, the optimal polarization parameter adjustment strategy is selected for each antenna; S43. Based on the performance indicator data, a predetermined objective function is constructed; based on the predetermined objective function and the polarization parameter adjustment strategy, the NSGA-II algorithm is used for optimization, and the Pareto level and the optimized population are iteratively generated; Select the optimal polarization parameters based on Pareto rank and optimized population; S44. Based on the optimal polarization parameter and the ideal polarization state, an adjustment amount is calculated, and the adjustment amount is ensured to be within a preset hardware limit range; based on the adjustment amount, a new attenuation code and a new phase shift code are generated for each antenna as optimized polarization parameters.

5. The antenna polarization simulation method based on phase and amplitude calibration according to claim 4, characterized in that: Step S5 is further as follows: S51, obtaining the position information of each antenna and constructing the geometric structure of the antenna array; obtaining the optimized polarization parameters and initializing them to obtain the initial polarization state of each antenna; Based on the geometric structure of the antenna array and the initial polarization state of each antenna, a coupling matrix between antennas is constructed; Integrate the location information, initial polarization state and coupling matrix of each antenna into structured data to form a multi-antenna system model; S52, dividing the multi-antenna system model into predetermined subsystems, constructing a local objective function for each subsystem; based on the local objective function, implementing a distributed optimization algorithm based on ADMM to obtain an optimized global polarization state; S53, calculating a spatial correlation matrix of each antenna pair based on a multi-antenna system model; Based on the spatial correlation matrix, eigenvalue decomposition is performed to obtain the eigenvector matrix; Selecting orthogonal eigenvectors from the eigenvector matrix as polarization directions of different antennas; combining the orthogonal eigenvectors with the optimized global polarization state to generate polarization control parameters; S54, calculating the beamforming weight of each antenna based on the preset desired beam direction; The beamforming weights are combined with polarization control parameters to generate a composite weight matrix; the composite weight matrix is ​​optimized using the MVDR algorithm to obtain an optimized composite weight matrix; the optimized composite weights are converted into the ultimate control parameters of each antenna.

6. The antenna polarization simulation method based on phase and amplitude calibration according to claim 5, characterized in that: Step S43 is further as follows: S431, constructing a predetermined objective function based on the performance indicator data; generating an initial population using the NSGA-II algorithm based on the predetermined objective function, wherein each individual of the initial population represents a set of polarization parameters; S432, based on the polarization parameter adjustment strategy and the initial population, non-dominated sorting and crowding degree calculation are performed to obtain the Pareto rank and crowding degree of each individual; S433, based on the Pareto level and crowding degree of each individual, selection, crossover and mutation operations are performed, and the operations are repeated until a preset termination condition is reached to obtain an iteratively optimized population; Based on the Pareto rank and the optimized population, the optimal polarization parameters are selected.

7. The antenna polarization simulation method based on phase and amplitude calibration according to claim 5, characterized in that: Step S3 also includes: S3a, obtaining target polarization parameters input by the user, including ellipticity and azimuth, and calculating normalized electric field components based on the target polarization parameters; Based on the electric field components, the Stokes parameter vector is calculated; S3b, obtaining the current operating frequency, and based on the current operating frequency, extracting material characteristic data from pre-stored frequency-related parameters; based on the material characteristic data, calculating the birefringence effect and polarization-related loss of the system; Based on the birefringence effect and polarization-dependent loss of the system, the Mueller matrix is ​​constructed and normalized; the diagonal elements of the Mueller matrix represent attenuation, and the off-diagonal elements represent polarization conversion; S3c, multiply the Stokes parameter vector and the Mueller matrix to obtain a product result; Normalizing the product result to obtain a normalized product result; extracting new ellipsometric ratio and azimuth angle information based on the normalized product result; converting the new ellipsometric ratio and azimuth angle information into the required amplitude ratio and phase difference of the horizontal and vertical components; S3d, based on the required amplitude ratio of the horizontal and vertical components, respectively calculating the target amplitude values ​​required for the horizontal and vertical channels; based on the target amplitude values ​​required for the horizontal and vertical channels, using a binary search algorithm in the "amplitude value-attenuation code" lookup table, finding the attenuation code closest to the target amplitude value as the required attenuation code; Based on the required phase difference between the horizontal and vertical components, a binary search algorithm is used in the "phase difference value-phase shift code" lookup table to find the phase shift code closest to the target phase difference as the required phase shift code.

8. The antenna polarization simulation method based on phase and amplitude calibration according to claim 5, characterized in that: It also includes S6, converting the final polarization control parameters into hardware control commands and generating a performance evaluation report, specifically: S61, obtaining hardware specifications, and based on the hardware specifications, converting the final polarization control parameters into voltage or current values; based on the voltage or current values, generating a standardized control instruction sequence; packaging the control instruction sequence, and adding a timestamp and verification information; S62, establishing a communication link with the polarization simulation hardware, sending the control instruction sequence to the polarization simulation hardware and executing it according to a predefined protocol; monitoring the execution status of the control instruction sequence to ensure that each antenna receives and executes it correctly; Record the timestamp of execution completion; S63, using a preconfigured performance measurement device, collecting key performance indicators within a predefined time window; calculating statistical characteristics of the key performance indicators; associating the statistical characteristics with the timestamp of the completion of the execution of step S62; and recording environmental parameters, including temperature and humidity; S64. Based on statistical characteristics and environmental parameters, calculate performance improvement indicators and generate a performance evaluation report, including the values, trend graphs and analysis results of various indicators; The performance evaluation report is stored as structured data, while key parameters are extracted as input for the next round of optimization.

Images