Radio equipment error calibration method under complex magnetic field interference

Abstract

The invention relates to the technical field of radio equipment error calibration, in particular to a radio equipment error calibration method under complex magnetic field interference, which comprises the following steps: S1, magnetic field interference detection: detecting the magnetic field intensity and direction around radio equipment in real time, and recording magnetic field data; s2, interference source positioning: positioning the position and intensity of a magnetic field interference source; s3, interference model construction: constructing a magnetic field interference model; s4, error analysis and calibration: dynamically correcting errors caused by complex magnetic field interference; s5, error calibration: correcting errors caused by complex magnetic field interference in real time; s6, evaluating a calibration effect: verifying the calibration effect; and S7, adaptive calibration: adaptively adjusting and calibrating parameters of the nonlinear calibration model. According to the invention, accurate and reliable basic data is provided for error calibration, and the performance and effectiveness of the whole calibration system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of radio equipment error calibration, and in particular to a radio equipment error calibration method under complex magnetic field interference. Background Art

[0002] With the widespread application of radio technology, radio equipment plays a vital role in various fields such as communication, navigation, and positioning. However, the complex magnetic field environment has a significant impact on the performance of radio equipment. Magnetic field interference can cause the frequency, phase, amplitude and other performance parameters of radio equipment to shift, thereby affecting the normal operation of the equipment. Therefore, how to perform high-precision error calibration of radio equipment under complex magnetic field interference has become an important research topic.

[0003] At present, research on error calibration of radio equipment mainly focuses on simple linear models and traditional algorithms. These methods often do not work well when dealing with nonlinear interference in complex magnetic field environments. The existing technology lacks effective distinction and targeted calibration strategies for different types of magnetic field interference (static and dynamic interference), resulting in insufficient calibration accuracy and stability. In addition, the existing technology also has major limitations in real-time and adaptability, and cannot meet the high requirements of calibration accuracy and response speed in dynamically changing complex magnetic field environments.

[0004] The present invention aims to provide a method for calibrating radio equipment errors under complex magnetic field interference, which not only improves the accuracy and stability of calibration, but also effectively copes with different types of magnetic field interference, thereby ensuring the performance and reliability of radio equipment in complex magnetic field environments. Summary of the invention

[0005] Based on the above purpose, the present invention provides a method for calibrating radio equipment errors under complex magnetic field interference.

[0006] A method for calibrating radio equipment errors under complex magnetic field interference comprises the following steps: S1, magnetic field interference detection: use magnetic field sensors to detect the magnetic field strength and direction around the radio equipment in real time and record the magnetic field data; S2, interference source positioning: Based on the magnetic field data, the position and intensity of the magnetic field interference source are located through a multi-point data fusion algorithm; S3, interference model construction: based on the detected magnetic field data and the location and strength of the magnetic field interference source, a magnetic field interference model is constructed, including two types of static interference and dynamic interference; S4, Error Analysis and Calibration: Analyze the errors caused by magnetic field interference on the performance parameters of radio equipment, generate nonlinear calibration models, and dynamically correct the errors caused by complex magnetic field interference, including: S41, Error characteristic analysis: Analyze the errors caused by magnetic field interference on the performance parameters of radio equipment using the constructed magnetic field interference model. The performance parameters include frequency, phase and amplitude. The error distribution and characteristics under different interference conditions are obtained through simulation and experimental data. S42, data preprocessing: collect historical magnetic field data and performance parameters for preprocessing, including data cleaning and normalization; S43, model generation: using the preprocessed historical magnetic field data and performance parameters to generate a nonlinear calibration model; S44, calibration strategy design: design corresponding calibration strategies according to different types of magnetic field interference, and optimize the nonlinear calibration model in combination with the real-time detected magnetic field data; S5, Error Calibration: Apply nonlinear calibration models to radio equipment to correct errors caused by complex magnetic field interference in real time; S6, Calibration effect evaluation: Evaluate the performance of the calibrated radio equipment through test signals and field measurements to verify the calibration effect; S7, adaptive calibration: adaptively adjust the parameters of the nonlinear calibration model according to the real-time detected magnetic field changes.

[0007] Optionally, the magnetic field interference detection in S1 includes: S11, sensor arrangement: arranging a plurality of three-axis magnetic field sensors around the radio device to form a sensor array; S12, data collection: real-time collection of magnetic field intensity and direction data at different locations through a sensor array; S13, data aggregation and recording: the magnetic field data collected by the sensor are aggregated and recorded.

[0008] Optionally, the interference source positioning in S2 includes: S21, multi-point data fusion: The magnetic field data of each sensor is fused using a weighted average algorithm, expressed as: ; in, is the fused magnetic field data, For the The magnetic field data of each sensor, is the weight corresponding to the sensor data, is the number of sensors; S22, interference source location calculation: using the inversion algorithm, the location and intensity of the magnetic field interference source are calculated through the fused magnetic field data, expressed as: ; in, is the position vector of the interference source, For the The position vector of each sensor, For the The magnetic field strength detected by each sensor.

[0009] Optionally, the interference model building in S3 includes: S31, static interference model construction: Assuming the interference source position is fixed and using the least squares method to fit and optimize, a model simulating the static magnetic field interference distribution is constructed, specifically including: S311, fixed position assumption: Assume that the interference source is fixed, and use the magnetic field intensity attenuation algorithm to calculate the theoretical magnetic field intensity of each detection point, expressed as: ; in, The distance from the interference source is The magnetic field strength at is the magnetic field strength at the interference source; S312, fitting and optimization: Based on the multi-point magnetic field data, the least squares method is used to fit the interference source position and intensity, and the static interference model is optimized, which is expressed as: ; in, For the The magnetic field strength detected by each sensor, is the magnetic field strength at the interference source, For the The distance from each sensor to the interference source; S32, dynamic interference model construction: Assuming that the interference source position changes with time and using the Kalman filter algorithm, a model simulating the dynamic magnetic field interference change is constructed, which specifically includes: S321, time variation hypothesis: Assume that the interference source position changes with time, record the magnetic field data and interference source position at different times, and use the dynamic variation model to calculate the theoretical magnetic field strength of each detection point, expressed as: ; in, For time and the distance from the interference source is The magnetic field strength at For time The magnetic field strength at the interference source is For time The location of the interference source; S322, dynamic data processing: Use the Kalman filter algorithm to process dynamic data and predict changes in the location and intensity of interference sources.

[0010] Optionally, the error feature analysis in S41 includes: S411, analyzing errors using magnetic field interference models: using the constructed static and dynamic magnetic field interference models, analyzing the errors caused by magnetic field interference on the performance parameters of the radio equipment, specifically including: Frequency error analysis: The frequency offset caused by interference is calculated using the magnetic field interference model, expressed as: ; in, is the frequency offset, is the sensitivity coefficient of frequency to magnetic field changes, is the interference magnetic field strength; Phase error analysis: The phase shift caused by interference is calculated using the magnetic field interference model and is expressed as: ; in, is the phase shift, is the sensitivity coefficient of phase to magnetic field changes, is the interference magnetic field strength, For time; Amplitude error analysis: The amplitude change caused by interference is calculated using the magnetic field interference model, expressed as: ; in, is the amplitude change, is the sensitivity coefficient of amplitude to magnetic field changes, is the interference magnetic field strength; S412, simulation and experimental data acquisition: Through simulation and experiments, record the performance parameter changes under different interference conditions, and obtain error distribution and characteristics, including: Simulation data: Use simulation tools to simulate the performance of radio equipment under different magnetic field interference conditions and generate simulation data; Experimental data: In actual environments, the performance of radio equipment under different magnetic field interference conditions is measured experimentally to obtain experimental data; Data analysis: Based on the simulation and experimental data, the error distribution and characteristics under different interference conditions are analyzed using the K-means clustering algorithm to form an error feature library.

[0011] Optionally, the nonlinear calibration model in S43 adopts a recursive neural network (RNN) model, and the recursive neural network (RNN) model includes: S431, data preprocessing: normalize the raw magnetic field data and performance parameters to eliminate scale differences between different features; S432, feature fusion: combine the normalized magnetic field data and performance parameters to form a multi-feature input vector, expressed as: ; in, is the magnetic field data, is the performance parameter, is the merged normalized input feature vector; S433, Multi-layer recursive neural network calculation: Increase the depth of the recursive neural network layer, expressed as: First layer: ; Second layer: ; in, and The time The hidden states of the first and second layers of the recurrent neural network, and The time The hidden states of the first and second layers of the recurrent neural network, is the weight matrix, is the bias vector; S434, Attention Mechanism: Introduce the attention mechanism, expressed as: Attention weight calculation: , is the time step Ratings; Attention Score: , It is the row vector in the attention mechanism used to perform a linear mapping on the output of the tanh function to generate a score; Weighted hidden states: ; in, The weighted hidden state, is the attention weight, Score attention. and is the weight matrix, is the bias vector; S435, residual connection: Using residual connection, the original hidden state is added to the hidden state weighted by the attention mechanism, expressed as: ; in, is the hidden state after residual connection; S436, error calibration output: The adjusted hidden state is mapped to the calibrated error value through the output layer to achieve the calibration of the radio equipment error under complex magnetic field interference, which is expressed as: ; in, For time The calibration output, is the output weight matrix, is the output bias vector.

[0012] Optionally, the calibration strategy design in S44 includes: S441, identifying interference type: distinguishing static interference from dynamic interference according to the magnetic field data detected in real time; S442, static interference calibration strategy design: For static interference, set the interference intensity and location Fixed,calibrated using the constructed static interference model; S443, Dynamic interference calibration strategy design: For dynamic interference, set the interference intensity and location As time changes, the dynamic interference model is updated in real time using the particle filter algorithm and calibrated; S444, calibration strategy optimization: Combined with the real-time detected magnetic field data, dynamically adjust the calibration parameters of static and dynamic interference, and optimize the nonlinear calibration model, expressed as: ; in, The optimized time The calibration parameters, For time The calibration parameters, is the learning rate, is the gradient of the calibration parameter, is the loss function (such as the sum of squared errors).

[0013] Optionally, the error calibration in S5 includes: S51, model application: deploy the optimized nonlinear calibration model to the control platform of the radio equipment, and perform error calibration by inputting magnetic field data and equipment performance parameters in real time; S52, real-time data input: real-time collection of magnetic field data and equipment performance parameters , and input into the nonlinear calibration model; S53, error correction calculation: using the nonlinear calibration model to calculate the correction value in real time, expressed as: ; ; in, For time The input feature vector (including normalized magnetic field data and equipment performance parameters), For time The hidden state of For time The hidden state of is the weight matrix input to the hidden layer, is the weight matrix from hidden layer to hidden layer, is the bias vector of the hidden layer, is the output weight matrix, is the output bias vector, For time The calibration output (corrected error value); S54, calibration feedback: The calculated calibration output Feedback is sent to the control platform of the radio equipment to adjust the radio equipment parameters in real time.

[0014] Optionally, the calibration effect evaluation in S6 includes: S61, test signal generation: generating a plurality of test signals with known characteristics, covering various magnetic field interference conditions; S62, Calibration Equipment Test: Input the test signal into the calibrated radio equipment and record the equipment's response data under different interference conditions; S63, Field Measurement: Conduct field measurements on calibrated radio equipment in actual use environments to collect performance data of the equipment in real magnetic field environments; S64, performance parameter comparison: compare the test signal and the field measured data with the performance parameters under normal conditions, calculate the error, and express it as: ; in, is the number of test samples, is the performance parameter value under normal conditions, is the actual measured value of the radio equipment after calibration; S65, Statistical analysis: Evaluate the calibration effect by statistically analyzing the error distribution and mean square error (MSE).

[0015] Optionally, the adaptive calibration in S7 includes: S71, real-time magnetic field detection: use the magnetic field sensor to detect the changes in magnetic field strength and direction in real time and record real-time magnetic field data ; S72, error calculation: Calculate the error of the calibrated device, expressed as: ; in, For time The error, is the output value of the device after calibration, is the performance parameter value under normal conditions; S73, gradient calculation: Calculate the gradient of the nonlinear calibration model parameters according to the error, expressed as: ; in, are the parameters of the nonlinear calibration model, is the loss function (such as the sum of squared errors); S74, adaptive learning rate update: adaptively adjust the learning rate according to real-time error and gradient , expressed as: ; in, is the initial learning rate, For the The gradient of the step; S75, parameter update: Update the model parameters using adaptive learning rate, expressed as: .

[0016] Beneficial effects of the present invention: The present invention detects and collects magnetic field strength and direction data in real time, and uses a multi-point data fusion algorithm to accurately locate the position and strength of the magnetic field interference source, constructs a magnetic field model including static interference and dynamic interference, ensures a comprehensive understanding of the complex magnetic field environment, provides accurate and reliable basic data for error calibration, and helps to improve the performance and effectiveness of the overall calibration system.

[0017] The present invention realizes efficient processing of complex time series data and dynamic calibration strategy design through error feature analysis, data preprocessing and recursive neural network model generation, combined with attention mechanism and residual connection, and designs and optimizes different types of magnetic field interference calibration strategies, ensuring high-precision error calibration under static and dynamic interference, which not only improves the accuracy and stability of calibration, but also effectively responds to calibration requirements under different interference conditions, and significantly improves the performance and reliability of radio equipment in complex magnetic field environments.

[0018] The present invention uses calibration effect evaluation and adaptive calibration mechanism, utilizes test signals and field measurement data to comprehensively verify the calibration effect, and dynamically adjusts calibration model parameters in combination with real-time detected magnetic field changes to ensure the real-time and accuracy of the calibration. The adaptive learning rate gradient descent algorithm optimizes the calibration parameters and improves the efficient calibration capability of the device in a complex magnetic field environment, thereby significantly improving the performance stability and reliability of the device and meeting the demand for high-precision calibration in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 A schematic diagram of a calibration method flow chart of an embodiment of the present invention; Figure 2 Schematic diagram of error analysis and calibration according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0022] It should be noted that the references to "one embodiment", "embodiment", "exemplary embodiments", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0023] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0024] like Figure 1-Figure 2 As shown, a method for calibrating radio equipment errors under complex magnetic field interference includes the following steps: S1, magnetic field interference detection: use magnetic field sensors to detect the magnetic field strength and direction around the radio equipment in real time and record the magnetic field data; S2, interference source positioning: Based on the magnetic field data, the position and intensity of the magnetic field interference source are located through a multi-point data fusion algorithm; S3, interference model construction: based on the detected magnetic field data and the location and strength of the magnetic field interference source, a magnetic field interference model is constructed, including two types of static interference and dynamic interference; S4, Error Analysis and Calibration: Analyze the errors caused by magnetic field interference on the performance parameters of radio equipment, generate nonlinear calibration models, and dynamically correct the errors caused by complex magnetic field interference, including: S41, Error characteristic analysis: Analyze the errors caused by magnetic field interference on the performance parameters of radio equipment using the constructed magnetic field interference model. The performance parameters include frequency, phase and amplitude. The error distribution and characteristics under different interference conditions are obtained through simulation and experimental data. S42, data preprocessing: collect historical magnetic field data and performance parameters for preprocessing, including data cleaning and normalization; S43, model generation: using the preprocessed historical magnetic field data and performance parameters to generate a nonlinear calibration model; S44, calibration strategy design: according to different types of magnetic field interference, design corresponding calibration strategies to ensure effective calibration under both static and dynamic interference, and optimize the nonlinear calibration model in combination with real-time detected magnetic field data to ensure real-time correction of errors caused by complex magnetic field interference; S5, Error Calibration: Apply nonlinear calibration models to radio equipment to correct errors caused by complex magnetic field interference in real time; S6, Calibration effect evaluation: Evaluate the performance of the calibrated radio equipment through test signals and field measurements to verify the calibration effect; S7, Adaptive calibration: Adaptively adjust the parameters of the nonlinear calibration model according to the real-time detected magnetic field changes to ensure accurate calibration of the radio equipment in different magnetic field environments; Through the above steps, high-precision, dynamic and real-time error calibration is ensured in a complex magnetic field environment, which significantly improves the performance and reliability of the radio equipment.

[0025] Magnetic field interference detection in S1 includes: S11, sensor arrangement: arranging a plurality of three-axis magnetic field sensors around the radio device to form a sensor array; S12, data collection: real-time collection of magnetic field intensity and direction data at different locations through a sensor array; S13, data aggregation and recording: aggregation and recording of magnetic field data collected by the sensor; Through the above steps, the magnetic field strength and direction data are collected and summarized in real time, ensuring the comprehensiveness and real-time nature of magnetic field detection, providing accurate and reliable basic data for subsequent interference source positioning and model building, and helping to improve the overall calibration effect.

[0026] Interference source positioning in S2 includes: S21, multi-point data fusion: The magnetic field data of each sensor is fused using a weighted average algorithm, expressed as: ; in, is the fused magnetic field data, For the The magnetic field data of each sensor, is the weight corresponding to the sensor data, is the number of sensors, weight Distribution based on sensor accuracy and location; S22, interference source location calculation: using the inversion algorithm, the location and intensity of the magnetic field interference source are calculated through the fused magnetic field data, expressed as: ; in, is the position vector of the interference source, For the The position vector of each sensor, For the The magnetic field strength detected by each sensor; Through the above steps, the position and strength of the magnetic field interference source can be accurately located, ensuring the accuracy and reliability of the positioning process, providing accurate basic data for subsequent error calibration, thereby improving the performance and effectiveness of the entire calibration system.

[0027] The interference model construction in S3 includes: S31, static interference model construction: Assuming the interference source position is fixed and using the least squares method to fit and optimize, a model simulating the static magnetic field interference distribution is constructed, specifically including: S311, fixed position assumption: Assume that the interference source is fixed, and use the magnetic field intensity attenuation algorithm to calculate the theoretical magnetic field intensity of each detection point, expressed as: ; in, The distance from the interference source is The magnetic field strength at is the magnetic field strength at the interference source; S312, fitting and optimization: Based on the multi-point magnetic field data, the least squares method is used to fit the interference source position and intensity, and the static interference model is optimized, which is expressed as: ; in, For the The magnetic field strength detected by each sensor, is the magnetic field strength at the interference source, For the The distance from each sensor to the interference source; S32, dynamic interference model construction: Assuming that the interference source position changes with time and using the Kalman filter algorithm, a model simulating the dynamic magnetic field interference change is constructed, which specifically includes: S321, time variation hypothesis: Assume that the interference source position changes with time, record the magnetic field data and interference source position at different times, and use the dynamic variation model to calculate the theoretical magnetic field strength of each detection point, expressed as: ; in, For time and the distance from the interference source is The magnetic field strength at For time The magnetic field strength at the interference source is For time The location of the interference source; S322, dynamic data processing: Use the Kalman filter algorithm to process dynamic data and predict the changes in the location and intensity of the interference source, expressed as: predict: ; ; renew: ; ; ; in, and are the predicted and updated state estimates, respectively, and is the corresponding covariance matrix, is the Kalman gain, is the observed value, is the state transfer matrix, is the control matrix, is the control input, is the process noise covariance, is the observation matrix, is the observation noise covariance, is the identity matrix, is the state estimate, which means Moment based on The state estimation is performed based on the information at each moment. is the state covariance matrix, expressed in Moment based on The uncertainty of the state of information at the moment, is the state transfer matrix The transpose of Through the above content, the magnetic field distribution under different interference sources can be fully simulated. This method not only improves the accuracy and reliability of the model, but also provides reliable basic data for error calibration, thereby significantly improving the overall calibration effect and the performance of radio equipment in complex magnetic field environments.

[0028] Error signature analysis in S41 includes: S411, analyzing errors using magnetic field interference models: using the constructed static and dynamic magnetic field interference models, analyzing the errors caused by magnetic field interference on the performance parameters of the radio equipment, specifically including: Frequency error analysis: The frequency offset caused by interference is calculated using the magnetic field interference model, expressed as: ; in, is the frequency offset, is the sensitivity coefficient of frequency to magnetic field changes, is the interference magnetic field strength; Phase error analysis: The phase shift caused by interference is calculated using the magnetic field interference model and is expressed as: ; in, is the phase shift, is the sensitivity coefficient of phase to magnetic field changes, is the interference magnetic field strength, For time; Amplitude error analysis: The amplitude change caused by interference is calculated using the magnetic field interference model, expressed as: ; in, is the amplitude change, is the sensitivity coefficient of amplitude to magnetic field changes, is the interference magnetic field strength; S412, simulation and experimental data acquisition: Through simulation and experiments, record the performance parameter changes under different interference conditions, and obtain error distribution and characteristics, including: Simulation data: Use simulation tools to simulate the performance of radio equipment under different magnetic field interference conditions and generate simulation data; Experimental data: In actual environments, the performance of radio equipment under different magnetic field interference conditions is measured experimentally to obtain experimental data; Data analysis: Combine simulation and experimental data, analyze the error distribution and characteristics under different interference conditions through K-means clustering algorithm, and form an error feature library; The K-means clustering algorithm specifically includes: Initialization: Select K initial cluster centers (i.e. centroids), which are ,in Indicates Cluster centers (centroids); Assign samples: Assign each data point to the cluster center closest to it and update the cluster members, expressed as: ; ; in, For data points To Cluster Centers The distance Indicates the first data points, Represents data points The cluster number to which it belongs; Update cluster centers: Recalculate the position of each cluster center based on the currently assigned cluster members, expressed as: ; in, It is The set of members of a cluster, It is The number of members in a cluster; Iteration: Repeat the sample allocation and cluster center update until the cluster center no longer changes or the change is less than the preset threshold; The threshold is determined through experiments, including: Set the initial threshold: Choose a reasonable small value as the initial threshold, usually between 0.001 and 0.0001; Calculate the change in cluster center: In each iteration, calculate the change in cluster center between the current iteration and the previous iteration. The calculation formula is: ; in, It is The first iteration The location of the cluster center, It is The first iteration The location of the cluster centers; Compare the change to the threshold: If the change If the change is less than the preset threshold, the iteration is stopped and the clustering result is considered to have converged. If it is greater than the preset threshold, the iteration continues; Through the above content, not only the accuracy and comprehensiveness of error analysis are improved, but also a detailed reference basis is provided for subsequent error calibration, thereby significantly improving the calibration effect and performance stability of radio equipment in complex magnetic field environments.

[0029] The nonlinear calibration model in S43 adopts a recurrent neural network (RNN) model, which includes: S431, data preprocessing: normalize the original magnetic field data and performance parameters to eliminate the scale differences between different features and ensure the stability and effectiveness of the input data during the model training process, expressed as: ; in, is the normalized input data, is the original input data, Indicates normalization processing; S432, feature fusion: Combine the normalized magnetic field data and performance parameters to form a multi-feature input vector, so that the model can consider multiple influencing factors at the same time and improve the calibration effect, which is expressed as: ; in, is the magnetic field data, is the performance parameter, is the merged normalized input feature vector; S433, Multi-layer recurrent neural network calculation: Increasing the depth of the recurrent neural network layer enables the model to capture more complex time dependencies and improve the ability to express time series data, thereby more accurately modeling and calibrating errors, expressed as: First layer: ; Second layer: ; in, and The time The hidden states of the first and second layers of the recurrent neural network, and The time The hidden states of the first and second layers of the recurrent neural network, is the weight matrix, is the bias vector; S434, attention mechanism: The attention mechanism is introduced to enable the model to dynamically adjust the degree of attention to the input data at different times, focusing on the time that has a greater impact on error calibration, and improving the calibration accuracy, which is expressed as: Attention weight calculation: , is the time step Ratings; Attention Score: , It is the row vector in the attention mechanism used to perform a linear mapping on the output of the tanh function to generate a score; Weighted hidden states: ; in, The weighted hidden state, is the attention weight, Score attention. and is the weight matrix, is the bias vector; S435, residual connection: Using residual connection, the original hidden state is added to the hidden state weighted by the attention mechanism to alleviate the gradient vanishing problem in deep networks and improve the training effect and stability of the model, expressed as: ; in, is the hidden state after residual connection; S436, error calibration output: The adjusted hidden state is mapped to the calibrated error value through the output layer to calibrate the error of the radio equipment under complex magnetic field interference and improve the performance of the equipment, which is expressed as: ; in, For time The calibration output, is the output weight matrix, is the output bias vector; Through the above content, the model's ability to process complex time series data has been improved, enabling it to more accurately capture and correct the impact of complex magnetic field interference on the performance parameters of radio equipment, dynamically adjust the focus on key time, alleviate the gradient vanishing problem of deep networks, and ensure the stability and high precision of calibration results, thereby significantly improving the calibration effect and performance reliability of radio equipment in complex magnetic field environments.

[0030] The calibration strategy design in S44 includes: S441, identifying interference type: distinguishing static interference from dynamic interference based on real-time detected magnetic field data. Static interference refers to interference whose interference source position and intensity remain unchanged within a certain period of time, while dynamic interference refers to interference whose interference source position and intensity change over time. S442, static interference calibration strategy design: For static interference, set the interference intensity and location Fixed, calibrated using the constructed static interference model, expressed as: ; in, A static interference calibration function adjusts the parameters of the radio equipment according to the static interference model to minimize its error; S443, Dynamic interference calibration strategy design: For dynamic interference, set the interference intensity and location As time changes, the dynamic interference model is updated in real time using the particle filter algorithm and calibrated, which is expressed as: Initialization: Initialization Particles and weight , so that the initial weights satisfy ; Particle state prediction: predict each particle according to the state transition model, expressed as: ; in, For particles In time status, is the state transition probability; Weight update: Update the weight of each particle according to the observation model, expressed as: ; in, For particles In time The weight of is the observation probability; Weight normalization: Normalize the weights, expressed as: ; Resampling: Resampling is performed according to the particle weights to avoid particle degradation. The resampled particles are evenly distributed in high-weight areas. State estimation: The system state is estimated based on the weighted average of the particles, expressed as: ; S444, calibration strategy optimization: Combined with the real-time detected magnetic field data, dynamically adjust the calibration parameters of static and dynamic interference, optimize the nonlinear calibration model, so that the calibration strategy can effectively respond to different interference conditions, expressed as: ; in, The optimized time The calibration parameters, For time The calibration parameters, is the learning rate, is the gradient of the calibration parameter, is the loss function (such as the sum of squared errors); Through the above steps, different interference conditions can be effectively dealt with, the accuracy and stability of calibration are improved, and the reliable performance and high-precision calibration of radio equipment in complex magnetic field environments are ensured.

[0031] Error calibration in S5 includes: S51, model application: deploy the optimized nonlinear calibration model to the control platform of the radio equipment, and perform error calibration by inputting magnetic field data and equipment performance parameters in real time; S52, real-time data input: real-time collection of magnetic field data and equipment performance parameters , and input into the nonlinear calibration model; S53, error correction calculation: using the nonlinear calibration model to calculate the correction value in real time, expressed as: ; ; in, For time The input feature vector (including normalized magnetic field data and equipment performance parameters), For time The hidden state of For time The hidden state of is the weight matrix input to the hidden layer, is the weight matrix from hidden layer to hidden layer, is the bias vector of the hidden layer, is the output weight matrix, is the output bias vector, For time The calibration output (corrected error value); S54, calibration feedback: The calculated calibration output Feedback to the control platform of the radio equipment to adjust the radio equipment parameters in real time to correct errors caused by complex magnetic field interference; Through the above steps, the real-time and accuracy of calibration are improved, ensuring that the radio equipment maintains stable performance and high-precision error calibration in a complex magnetic field environment, significantly improving the reliability and operating efficiency of the equipment.

[0032] The calibration effect evaluation in S6 includes: S61, test signal generation: generating a plurality of test signals with known characteristics, covering various magnetic field interference conditions; S62, Calibration Equipment Test: Input the test signal into the calibrated radio equipment and record the equipment's response data under different interference conditions; S63, Field Measurement: Conduct field measurements on calibrated radio equipment in actual use environments to collect performance data of the equipment in real magnetic field environments; S64, performance parameter comparison: compare the test signal and the field measured data with the performance parameters under normal conditions, calculate the error, and express it as: ; in, is the number of test samples, is the performance parameter value under normal conditions, is the actual measured value of the radio equipment after calibration; S65, Statistical analysis: The calibration effect is evaluated by statistically analyzing the error distribution and mean square error (MSE), expressed as: ; Through the above content, the calibration effect can be fully verified to ensure the high accuracy and stability of the equipment in complex magnetic field environments, providing a strong guarantee for the reliable operation of the equipment.

[0033] Adaptive calibration in S7 includes: S71, real-time magnetic field detection: use the magnetic field sensor to detect the changes in magnetic field strength and direction in real time and record real-time magnetic field data ; S72, error calculation: Calculate the error of the calibrated device, expressed as: ; in, For time The error, is the output value of the device after calibration, is the performance parameter value under normal conditions; S73, gradient calculation: Calculate the gradient of the nonlinear calibration model parameters according to the error, expressed as: ; in, are the parameters of the nonlinear calibration model, is the loss function (such as the sum of squared errors); S74, adaptive learning rate update: adaptively adjust the learning rate according to real-time error and gradient , expressed as: ; in, is the initial learning rate, For the The gradient of the step; S75, parameter update: Update the model parameters using adaptive learning rate, expressed as: ; Through the above contents, the real-time and accuracy of calibration are improved, ensuring the efficient calibration of radio equipment in complex magnetic field environments. By continuously optimizing the calibration parameters, the performance stability and reliability of the equipment are significantly improved, meeting the needs of high-precision calibration in practical applications.

[0034] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0035] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for calibrating radio equipment errors under complex magnetic field interference, characterized in that: The following steps are involved: S1, magnetic field interference detection: use magnetic field sensors to detect the magnetic field strength and direction around the radio equipment in real time and record the magnetic field data; S2, interference source positioning: Based on the magnetic field data, the position and intensity of the magnetic field interference source are located through a multi-point data fusion algorithm; S3, interference model construction: based on the detected magnetic field data and the location and strength of the magnetic field interference source, a magnetic field interference model is constructed, including two types of static interference and dynamic interference; S4, Error Analysis and Calibration: Analyze the errors caused by magnetic field interference on the performance parameters of radio equipment, generate nonlinear calibration models, and dynamically correct the errors caused by complex magnetic field interference, including: S41, Error characteristic analysis: Analyze the errors caused by magnetic field interference on the performance parameters of radio equipment using the constructed magnetic field interference model. The performance parameters include frequency, phase and amplitude. The error distribution and characteristics under different interference conditions are obtained through simulation and experimental data. S42, data preprocessing: collect historical magnetic field data and performance parameters for preprocessing, including data cleaning and normalization; S43, model generation: using the preprocessed historical magnetic field data and performance parameters to generate a nonlinear calibration model; S44, calibration strategy design: design corresponding calibration strategies according to different types of magnetic field interference, and optimize the nonlinear calibration model in combination with the real-time detected magnetic field data; S5, Error Calibration: Apply nonlinear calibration models to radio equipment to correct errors caused by complex magnetic field interference in real time; S6, Calibration effect evaluation: Evaluate the performance of the calibrated radio equipment through test signals and field measurements to verify the calibration effect; S7, adaptive calibration: adaptively adjust the parameters of the nonlinear calibration model according to the real-time detected magnetic field changes.

2. The method for calibrating radio equipment errors under complex magnetic field interference according to claim 1, characterized in that: The magnetic field interference detection in S1 includes: S11, sensor arrangement: arranging a plurality of three-axis magnetic field sensors around the radio device to form a sensor array; S12, data collection: real-time collection of magnetic field intensity and direction data at different locations through a sensor array; S13, data aggregation and recording: the magnetic field data collected by the sensor are aggregated and recorded.

3. The method for calibrating radio equipment errors under complex magnetic field interference according to claim 2, characterized in that: The interference source positioning in S2 includes: S21, multi-point data fusion: The magnetic field data of each sensor is fused using a weighted average algorithm, expressed as: ； in, is the fused magnetic field data, For the The magnetic field data of each sensor, is the weight corresponding to the sensor data, is the number of sensors; S22, interference source location calculation: using the inversion algorithm, the location and intensity of the magnetic field interference source are calculated through the fused magnetic field data, expressed as: ； in, is the position vector of the interference source, For the The position vector of each sensor, For the The magnetic field strength detected by each sensor.

4. The method for calibrating radio equipment errors under complex magnetic field interference according to claim 1, characterized in that: The interference model construction in S3 includes: S31, static interference model construction: Assuming the interference source position is fixed and using the least squares method to fit and optimize, a model simulating the static magnetic field interference distribution is constructed, specifically including: S311, fixed position assumption: Assume that the interference source is fixed, and use the magnetic field intensity attenuation algorithm to calculate the theoretical magnetic field intensity of each detection point, expressed as: ； in, The distance from the interference source is The magnetic field strength at is the magnetic field strength at the interference source; S312, fitting and optimization: Based on the multi-point magnetic field data, the least squares method is used to fit the interference source position and intensity, and the static interference model is optimized, which is expressed as: ； in, For the The magnetic field strength detected by each sensor, is the magnetic field strength at the interference source, For the The distance from each sensor to the interference source; S32, dynamic interference model construction: Assuming that the interference source position changes with time and using the Kalman filter algorithm, a model simulating the dynamic magnetic field interference change is constructed, which specifically includes: S321, time variation hypothesis: Assume that the interference source position changes with time, record the magnetic field data and interference source position at different times, and use the dynamic variation model to calculate the theoretical magnetic field strength of each detection point, expressed as: ； in, For time and the distance from the interference source is The magnetic field strength at For time The magnetic field strength at the interference source is For time The location of the interference source at the time; S322, dynamic data processing: Use the Kalman filter algorithm to process dynamic data and predict changes in the location and intensity of interference sources.

5. The method for calibrating radio equipment errors under complex magnetic field interference according to claim 1, characterized in that: The error characteristic analysis in S41 includes: S411, analyzing errors using magnetic field interference models: using the constructed static and dynamic magnetic field interference models, analyzing the errors caused by magnetic field interference on the performance parameters of the radio equipment, specifically including: Frequency error analysis: The frequency offset caused by interference is calculated using the magnetic field interference model, expressed as: ； in, is the frequency offset, is the sensitivity coefficient of frequency to magnetic field changes, is the interference magnetic field strength; Phase error analysis: The phase shift caused by interference is calculated using the magnetic field interference model and is expressed as: ； in, is the phase shift, is the sensitivity coefficient of phase to magnetic field changes, is the interference magnetic field strength, For time; Amplitude error analysis: The amplitude change caused by interference is calculated using the magnetic field interference model, expressed as: ； in, is the amplitude change, is the sensitivity coefficient of amplitude to magnetic field changes, is the interference magnetic field strength; S412, simulation and experimental data acquisition: Through simulation and experiments, record the performance parameter changes under different interference conditions, and obtain error distribution and characteristics, including: Simulation data: Use simulation tools to simulate the performance of radio equipment under different magnetic field interference conditions and generate simulation data; Experimental data: In actual environments, the performance of radio equipment under different magnetic field interference conditions is measured experimentally to obtain experimental data; Data analysis: Based on the simulation and experimental data, the error distribution and characteristics under different interference conditions are analyzed using the K-means clustering algorithm to form an error feature library.

6. The method for calibrating radio equipment errors under complex magnetic field interference according to claim 5, characterized in that: The nonlinear calibration model in S43 adopts a recursive neural network model, and the recursive neural network model includes: S431, data preprocessing: normalize the raw magnetic field data and performance parameters to eliminate scale differences between different features; S432, feature fusion: combine the normalized magnetic field data and performance parameters to form a multi-feature input vector, expressed as: ； in, is the magnetic field data, is the performance parameter, is the merged normalized input feature vector; S433, Multi-layer recursive neural network calculation: Increase the depth of the recursive neural network layer, expressed as: First layer: ; Second layer: ; in, and The time The hidden states of the first and second layers of the recurrent neural network, and The time The hidden states of the first and second layers of the recurrent neural network, is the weight matrix, is the bias vector; S434, Attention Mechanism: Introduce the attention mechanism, expressed as: Attention weight calculation: , is the time step Ratings; Attention Score: , It is the row vector in the attention mechanism used to perform a linear mapping on the output of the tanh function to generate a score; Weighted hidden states: ; in, The weighted hidden state, is the attention weight, Score attention. and is the weight matrix, is the bias vector; S435, residual connection: Using residual connection, the original hidden state is added to the hidden state weighted by the attention mechanism, expressed as: ； in, is the hidden state after residual connection; S436, error calibration output: The adjusted hidden state is mapped to the calibrated error value through the output layer to achieve the calibration of the radio equipment error under complex magnetic field interference, which is expressed as: ； in, For time The calibration output, is the output weight matrix, is the output bias vector.

7. The method for calibrating radio equipment errors under complex magnetic field interference according to claim 6, characterized in that: The calibration strategy design in S44 includes: S441, identifying interference type: distinguishing static interference from dynamic interference according to the magnetic field data detected in real time; S442, static interference calibration strategy design: For static interference, set the interference intensity and location Fixed,calibrated using the constructed static interference model; S443, Dynamic interference calibration strategy design: For dynamic interference, set the interference intensity and location As time changes, the dynamic interference model is updated in real time using the particle filter algorithm and calibrated; S444, calibration strategy optimization: Combined with the real-time detected magnetic field data, dynamically adjust the calibration parameters of static and dynamic interference, and optimize the nonlinear calibration model, expressed as: ； in, The optimized time The calibration parameters, For time The calibration parameters, is the learning rate, is the gradient of the calibration parameter, is the loss function.

8. The method for calibrating radio equipment errors under complex magnetic field interference according to claim 7, characterized in that: The error calibration in S5 includes: S51, model application: deploy the optimized nonlinear calibration model to the control platform of the radio equipment, and perform error calibration by inputting magnetic field data and equipment performance parameters in real time; S52, real-time data input: real-time collection of magnetic field data and equipment performance parameters , and input into the nonlinear calibration model; S53, error correction calculation: using the nonlinear calibration model to calculate the correction value in real time, expressed as: ； ； in, For time The input feature vector is For time The hidden state of For time The hidden state of is the weight matrix input to the hidden layer, is the weight matrix from hidden layer to hidden layer, is the bias vector of the hidden layer, is the output weight matrix, is the output bias vector, For time The calibration output of S54, calibration feedback: The calculated calibration output Feedback is sent to the control platform of the radio equipment to adjust the radio equipment parameters in real time.

9. The method for calibrating radio equipment errors under complex magnetic field interference according to claim 1, characterized in that: The calibration effect evaluation in S6 includes: S61, test signal generation: generating a plurality of test signals with known characteristics, covering various magnetic field interference conditions; S62, Calibration Equipment Test: Input the test signal into the calibrated radio equipment and record the equipment's response data under different interference conditions; S63, Field Measurement: Conduct field measurements on calibrated radio equipment in actual use environments to collect performance data of the equipment in real magnetic field environments; S64, performance parameter comparison: compare the test signal and the field measured data with the performance parameters under normal conditions, calculate the error, and express it as: ； in, is the number of test samples, is the performance parameter value under normal conditions, is the actual measured value of the radio equipment after calibration; S65, Statistical analysis: Evaluate the calibration effect by statistically analyzing the error distribution and mean square error.

10. The method for calibrating radio equipment errors under complex magnetic field interference according to claim 9, characterized in that: The adaptive calibration in S7 includes: S71, real-time magnetic field detection: use magnetic field sensors to detect changes in magnetic field strength and direction in real time and record real-time magnetic field data ; S72, error calculation: Calculate the error of the calibrated device, expressed as: ； in, For time The error, is the output value of the device after calibration, is the performance parameter value under normal conditions; S73, gradient calculation: Calculate the gradient of the nonlinear calibration model parameters according to the error, expressed as: ； in, are the parameters of the nonlinear calibration model, is the loss function; S74, adaptive learning rate update: adaptively adjust the learning rate according to real-time error and gradient , expressed as: ； in, is the initial learning rate, For the The gradient of the step; S75, parameter update: Update the model parameters using adaptive learning rate, expressed as: 。

Images