A narrowband communication device detection apparatus and method

Abstract

The application discloses a kind of narrowband communication equipment detection equipment and detection method. Among them, the method includes: obtaining pre-established wireless channel model data set, the data set contains attenuation value and interference intensity distribution data under different environments;According to the attenuation value and interference intensity distribution data, a multipath propagation simulation model is constructed, a time delay spread range value is obtained by time domain analysis, and an amplitude distortion range value is obtained by frequency domain analysis;Environment noise measured data and the time delay spread range value, amplitude distortion range value are input into adaptive filtering algorithm, and signal quality evaluation data after filtering processing is generated, etc.The method of the present application analyzes and simulates the channel model data under different environments, combines various test methods and algorithms, comprehensively evaluates the performance of the equipment under various conditions, including signal quality, bit error rate, stability and risk, etc., so as to ensure the reliability of the equipment in practical application.

Description

Technical Field

[0001] The present invention relates to the technical field of narrowband communication equipment, and in particular to a narrowband communication equipment detection device and a detection method. Background Art

[0002] Narrowband communication equipment testing is a crucial research area and plays an indispensable role in modern wireless communication technology. In scenarios such as the Internet of Things, industrial control, and intelligent transportation, narrowband communication, with its low power consumption and long-distance transmission characteristics, has become a key technology for ensuring stable data transmission. However, technological progress in this field is directly related to the reliability and application effectiveness of communication systems. Any performance defects can lead to significant losses, making research on its detection and optimization particularly urgent.

[0003] Current testing methods for narrowband communication equipment mostly rely on idealized laboratory testing, which fails to truly reflect the device's performance in complex real-world scenarios. These methods often ignore external environmental interference factors, resulting in significant deviations between test results and actual applications, leading to frequent performance instability in real-world use.

[0004] Against this backdrop, narrowband communication equipment testing faces multiple challenges. The primary issue lies in the complex and volatile wireless transmission environment. Due to the limited signal bandwidth, narrowband communication requires extremely high signal quality. Spectrum congestion and noise interference directly impair signal reception. This degradation in signal quality further exacerbates performance fluctuations in devices exposed to multipath effects, making it difficult for devices to maintain stable communication capabilities in dynamic environments. Furthermore, devices must operate for extended periods in extreme conditions such as high and low temperatures, humidity, and strong vibration. These harsh environments place even higher demands on frequency stability and anti-interference capabilities. Existing testing methods often cannot fully simulate these complex scenarios, making it difficult to detect potential problems in a timely manner.

[0005] Therefore, how to build a comprehensive and effective detection solution under multiple scenarios close to actual applications, taking into account the influence of wireless channel characteristics, environmental noise and extreme working environments, has become a key issue in the performance evaluation and optimization of narrowband communication equipment. Summary of the Invention

[0006] In view of this, an object of the present invention is to provide a narrowband communication device detection device and a detection method, which can solve at least one of the technical problems mentioned in the background technology.

[0007] According to one aspect of the present invention, a method for detecting a narrowband communication device is provided, the method comprising:

[0008] Acquire a pre-established wireless channel model data set, wherein the data set includes attenuation values ​​and interference intensity distribution data under different environments;

[0009] A multipath propagation simulation model is constructed based on the attenuation value and interference intensity distribution data, a time domain analysis is used to obtain a delay spread range value, and a frequency domain analysis is combined to obtain an amplitude distortion range value;

[0010] Inputting the measured environmental noise data, the delay spread range value, and the amplitude distortion range value into an adaptive filtering algorithm to generate signal quality assessment data after filtering;

[0011] If the signal quality evaluation data does not reach the preset signal-to-noise ratio threshold, adjusting the transmit power parameter and the frequency offset parameter to obtain an optimized bit error rate value;

[0012] Constructing an extreme environment test model including temperature, humidity, and vibration parameters, and collecting drift data of the optimized bit error rate value under extreme conditions;

[0013] Performing feature fusion on the drift data and wireless channel characteristic data to generate a multi-scenario comprehensive performance evaluation data set;

[0014] If the fluctuation range value in the performance evaluation data exceeds a preset threshold, the channel model weight parameters are adjusted through an iterative algorithm to generate an updated stability score value;

[0015] Generate a multi-dimensional performance analysis graph based on the stability score value, and extract the bit error rate exceeding standard data and the frequency drift exceeding standard data under a specific environment;

[0016] The exceeding standard data is input into a linear regression model to predict the quantitative risk assessment index of the equipment in a complex environment.

[0017] In the above-mentioned technical solution, the narrowband communication equipment detection method comprehensively evaluates the equipment's performance under various conditions, including signal quality, bit error rate, stability, and risk, by analyzing and simulating channel model data in different environments and combining multiple testing methods and algorithms. This ensures the reliability of the equipment in practical applications. This method covers the entire process from model establishment, simulation analysis, signal processing, environmental testing, and risk prediction.

[0018] Acquisition of wireless channel model data set: Acquisition of a pre-established wireless channel model data set containing attenuation values ​​and interference intensity distribution data under different environments.

[0019] Multipath propagation simulation model construction: Based on the attenuation and interference intensity distribution data, a multipath propagation simulation model was constructed. Time domain analysis was used to determine the delay spread range, and frequency domain analysis was combined to determine the amplitude distortion range. This step simulates the multipath propagation effects in a real-world environment and provides a basis for signal quality assessment.

[0020] Adaptive filtering algorithm application: Measured ambient noise data, along with delay spread and amplitude distortion range values, are fed into the adaptive filtering algorithm to generate filtered signal quality assessment data. The adaptive filtering algorithm effectively reduces noise interference and improves the accuracy of signal quality assessment.

[0021] Transmit power and frequency offset parameter adjustment: If the signal quality assessment data does not meet the preset signal-to-noise ratio threshold, the transmit power and frequency offset parameters are adjusted to obtain an optimized bit error rate value. This step optimizes the device's transmission performance by adjusting key parameters.

[0022] Extreme Environment Test Model Construction: Build an extreme environment test model that includes temperature, humidity, and vibration parameters, and collect data on the drift of the optimized bit error rate under extreme conditions. This step can assess the performance stability of the device in extreme environments.

[0023] Generation of a comprehensive multi-scenario performance evaluation dataset: Drift data is combined with wireless channel characteristic data for feature fusion to generate a comprehensive multi-scenario performance evaluation dataset. This feature fusion considers multiple factors and provides a more comprehensive performance evaluation.

[0024] Iterative Algorithm Application: If the fluctuation range of the performance evaluation data exceeds a preset threshold, the channel model weight parameters are adjusted through an iterative algorithm to generate an updated stability score. The iterative algorithm can gradually optimize the model parameters and improve the accuracy of the stability assessment.

[0025] Multi-dimensional performance analysis chart generation: Based on the stability score, a multi-dimensional performance analysis chart is generated, extracting data on bit error rate and frequency drift exceeding standards under specific conditions. The multi-dimensional performance analysis chart intuitively demonstrates device performance in different environments.

[0026] Predicting Quantitative Risk Assessment Indicators: This step uses historical data analysis to predict potential risks to the equipment.

[0027] This method encompasses a complete process, from model building, simulation analysis, signal processing, to environmental testing and risk prediction, enabling comprehensive performance evaluation of narrowband communication equipment. By utilizing a variety of algorithms and testing methods, such as adaptive filtering, iterative algorithms, and linear regression models, the accuracy of the evaluation results is enhanced. It also considers device performance in diverse environments, including extreme conditions, ensuring device reliability in a wide range of practical application scenarios.

[0028] In some embodiments, a multipath propagation simulation model is constructed based on the attenuation value and the interference intensity distribution data, and a delay spread range value is obtained by using time domain analysis, and an amplitude distortion range value is obtained by combining frequency domain analysis, including:

[0029] By obtaining initial input from attenuation value data and interference intensity value, the data processing layer is used to preprocess the distributed data source to obtain a standardized basic data set;

[0030] Based on the standardized basic data set, a multipath propagation model is constructed, initialized using preset parameters in the simulation model library, and the preliminary structure of the model is determined;

[0031] Based on the multipath propagation model of the preliminary structure, the time domain analysis method is used to calculate the delay spread value and obtain the distribution characteristics of the delay spread range;

[0032] The amplitude distortion value is evaluated by combining the distribution characteristics of the delay spread range with the frequency domain analysis method to obtain the range boundary of the amplitude distortion;

[0033] If the range boundary of the amplitude distortion exceeds the preset threshold, the basic data set is recalibrated through the data processing layer to determine whether the calibrated data meets the constraints of the model construction method;

[0034] Based on the calibrated data, the parameters of the multipath propagation model are readjusted, and the support vector machine algorithm is used to optimize the model to obtain the final simulation results;

[0035] Through the final simulation results, the range outputs of delay spread value and amplitude distortion value are integrated to determine the signal characteristic distribution in the multipath propagation environment.

[0036] In the above technical solution, the above method aims to construct a simulation model that can accurately reflect the multipath propagation characteristics based on the attenuation value and interference intensity distribution data, and determine the range characteristics of delay spread and amplitude distortion through time domain and frequency domain analysis, thereby providing key signal characteristic distribution information for the performance evaluation of narrowband communication equipment.

[0037] Data preprocessing: Initial input is obtained from the attenuation data and interference intensity values. The data processing layer preprocesses the distributed data source to obtain a standardized basic data set. This preprocessing step removes noise and outliers from the data and unifies the data format.

[0038] Multipath propagation model initialization: Based on the standardized basic data set, a multipath propagation model is constructed and initialized using preset parameters from the simulation model library to determine the model's initial structure. The initialization process sets the model's starting state and parameter range.

[0039] Time Domain Analysis and Delay Spread Calculation: Based on the pre-structured multipath propagation model, we use time domain analysis to calculate delay spread and characterize the delay spread distribution. This analysis reveals the signal's spread along the time axis, reflecting the signal delay differences caused by multipath effects.

[0040] Frequency Domain Analysis and Amplitude Distortion Assessment: By analyzing the distribution characteristics of the delay spread range and combining it with frequency domain analysis, we evaluate the amplitude distortion value and determine the amplitude distortion range boundaries. Frequency domain analysis can show the amplitude variations of the signal at different frequencies, reflecting the impact of multipath propagation on the signal amplitude.

[0041] Data calibration and model constraint verification: If the amplitude distortion exceeds a preset threshold, the data processing layer performs a secondary calibration on the underlying dataset to determine if the calibrated data meets the constraints of the model-building method. Data calibration corrects for any deviations in the data, ensuring that the model inputs conform to actual physical properties.

[0042] Model parameter adjustment and optimization: Based on the calibrated data, the multipath propagation model parameters are readjusted and optimized using a support vector machine algorithm to obtain the final simulation results. This optimization step improves the accuracy and reliability of the simulation results by adjusting the model parameters.

[0043] Determining the signal characteristic distribution: The final simulation results are combined with the range outputs of delay spread and amplitude distortion to determine the signal characteristic distribution in a multipath propagation environment. This step summarizes the impact of multipath propagation on the signal and provides key information for subsequent performance evaluation of narrowband communication equipment.

[0044] This method, combining time-domain and frequency-domain analysis, can comprehensively and accurately assess delay spread and amplitude distortion in multipath propagation environments, providing detailed data support for performance evaluation of narrowband communication equipment. Data processing and calibration ensure that the model is based on high-quality data, and the application of support vector machine algorithms further enhances the model's accuracy and reliability. This method can adapt to attenuation values ​​and interference intensity distribution data in diverse environments, demonstrating its broad applicability.

[0045] In some embodiments, the measured environmental noise data, the delay spread range value, and the amplitude distortion range value are input into an adaptive filtering algorithm to generate filtered signal quality assessment data, including:

[0046] The initial input parameter set is formed by combining the collected environmental noise measured data with the delay spread range value and the amplitude distortion range value;

[0047] According to the initial input parameter group, the environmental noise measured data is preliminarily processed using an adaptive filtering algorithm to obtain a filtered first signal data set;

[0048] For the first signal data set, obtaining the delay spread characteristics and amplitude distortion characteristics therein, determining whether the characteristic distribution meets the preset threshold range, and if the characteristic distribution exceeds the threshold range, performing secondary filtering processing on the first signal data set to obtain a second signal data set;

[0049] Extracting signal quality related indicators from the second signal data set, determining whether the signal quality indicators reach a preset standard value, and if not, adjusting parameter configurations of the adaptive filtering algorithm to generate a third signal data set;

[0050] Calculating key parameter values ​​in the signal quality assessment data based on the third signal data set to obtain preliminary structured results of the assessment data;

[0051] Use the structured results, combined with the pre-established signal quality assessment model, to generate the final signal quality assessment data and determine the completeness and accuracy of the assessment data;

[0052] Through the final signal quality assessment data, the distribution characteristics of the outliers are extracted to determine whether the distribution of the outliers affects the overall signal quality. If the impact exceeds the preset range, the outliers are marked and recorded in the assessment data.

[0053] In this technical solution, the method primarily processes measured ambient noise data using an adaptive filtering algorithm, combining it with delay spread and amplitude distortion characteristics to generate data for signal quality assessment. This process aims to improve the accuracy of signal quality assessment and ensure that the performance of narrowband communication equipment in different environments can be accurately evaluated.

[0054] Formation of the initial input parameter group: Collect measured environmental noise data and combine it with the delay spread range value and the amplitude distortion range value to form the initial input parameter group.

[0055] Initial filtering: Based on the initial input parameter set, an adaptive filtering algorithm is used to perform preliminary processing on the measured ambient noise data, generating a filtered first signal data set. Initial filtering can remove some noise interference and initially improve signal quality.

[0056] Feature distribution verification and secondary filtering: For the first signal dataset, the delay spread and amplitude distortion characteristics are obtained to determine whether the feature distribution meets the preset threshold range. If it does, secondary filtering is performed to obtain the second signal dataset. This step further optimizes signal quality through secondary filtering to ensure that the feature distribution meets the requirements.

[0057] Signal Quality Indicator Assessment and Parameter Adjustment: This step extracts signal quality indicators from the second signal dataset and determines whether they meet preset standards. If not, the adaptive filtering algorithm's parameters are adjusted to generate a third signal dataset. This parameter adjustment improves the accuracy of signal quality assessment.

[0058] Calculating key parameter values ​​and obtaining structured results: Based on the third signal dataset, calculate key parameter values ​​in the signal quality assessment data and obtain preliminary structured results of the assessment data. This step converts the signal quality assessment into specific parameter values ​​for subsequent analysis.

[0059] Final signal quality assessment data generation: Combined with the pre-established signal quality assessment model, the preliminary structured results are used to generate the final signal quality assessment data, confirming its completeness and accuracy. This step integrates all information to generate the final assessment results.

[0060] Outlier distribution feature extraction and annotation: This step extracts the distribution features of outliers from the final signal quality assessment data to determine whether they impact overall signal quality. If the impact exceeds a preset range, the outliers are annotated and recorded. This step helps identify and address anomalies in signal quality.

[0061] This method integrates measured ambient noise data with delay spread and amplitude distortion characteristics, comprehensively considering multiple factors affecting signal quality. Through preliminary and secondary filtering, signal quality is gradually improved, ensuring that the characteristic distribution meets preset requirements. Based on the signal quality indicators, filtering algorithm parameters are dynamically adjusted, enhancing the flexibility and adaptability of the assessment.

[0062] In some embodiments, if the signal quality evaluation data does not reach a preset signal-to-noise ratio threshold, adjusting a transmit power parameter and a frequency offset parameter to obtain an optimized bit error rate value includes:

[0063] By processing the initial evaluation data of the signal quality, specific data of the signal-to-noise ratio value is obtained to determine whether it reaches the preset threshold;

[0064] If the signal-to-noise ratio value does not reach the preset threshold, preliminary parameter adjustments are performed on the transmit power and frequency offset to obtain a first set of adjusted parameter data;

[0065] According to the first set of parameter data, a pre-established simulation model is used to simulate signal transmission and obtain simulated signal quality data;

[0066] If the signal quality data after simulation still does not meet the preset threshold, the transmit power and frequency offset are optimized and adjusted twice using the support vector machine algorithm to obtain a second set of parameter data;

[0067] Based on the second set of parameter data, the signal transmission environment is monitored in real time to obtain an optimized bit error rate value;

[0068] By analyzing the bit error rate value, we can determine whether the stability of signal transmission meets the preset standards and obtain the final evaluation results;

[0069] If the final evaluation results show insufficient stability, the second set of parameter data is fine-tuned to obtain a third set of parameter data, and the simulation and monitoring processes are repeated to determine the optimal configuration for signal transmission.

[0070] In this technical solution, the method primarily addresses situations where signal quality assessment data falls below a preset signal-to-noise ratio threshold. By adjusting transmit power and frequency offset parameters and combining simulation with an optimization algorithm, the method gradually achieves an optimized bit error rate (BER) value, ensuring that signal transmission quality meets requirements. This process embodies the refined adjustment and optimization of signal transmission performance, which is crucial for improving the reliability of narrowband communication equipment.

[0071] Signal-to-noise ratio determination: This process processes the initial signal quality assessment data, obtains the specific signal-to-noise ratio value, and determines whether it reaches the preset threshold. This is the trigger condition for the entire adjustment process. Only by accurately determining whether the signal-to-noise ratio value meets the standard can we determine whether subsequent parameter adjustments are necessary.

[0072] Initial parameter adjustment: If the signal-to-noise ratio (SNR) value does not reach the preset threshold, preliminary parameter adjustments are made to the transmit power and frequency offset to obtain the first set of adjusted parameter data. This initial adjustment is based on a preliminary analysis of the current signal quality and provides preliminary optimization directions for subsequent simulations.

[0073] Simulation model simulation: Based on the first set of parameter data, a pre-established simulation model is used to simulate signal transmission and obtain simulated signal quality data. The simulation model can predict the signal transmission effect after parameter adjustment, providing a basis for further parameter optimization.

[0074] Secondary Optimization: If the simulated signal quality data still does not meet the preset threshold, a second optimization adjustment of the transmit power and frequency offset is performed using a support vector machine algorithm to obtain a second set of parameter data. The support vector machine algorithm learns from historical data and simulation results, providing a more accurate optimization solution for parameter adjustment.

[0075] Real-time monitoring and bit error rate acquisition: Based on the second set of parameter data, the signal transmission environment is monitored in real time to obtain the optimized bit error rate value. Real-time monitoring can reflect the effect of parameter adjustments in actual signal transmission. The bit error rate value is an important indicator for evaluating signal transmission quality.

[0076] Stability determination and final evaluation: By analyzing the bit error rate value, we determine whether the stability of signal transmission meets the preset standards and obtain the final evaluation results. Stability determination is a key step in evaluating signal transmission quality. Only stable signal transmission can ensure the reliability of communication.

[0077] Loop Adjustment and Optimization: If the final evaluation results indicate insufficient stability, fine-tune the second set of parameter data, obtain a third set of parameter data, and repeat the simulation and monitoring process to determine the optimal signal transmission configuration. This loop of adjustment and optimization continuously approaches the optimal signal transmission effect, ensuring the thoroughness and effectiveness of parameter adjustments.

[0078] This method, by gradually adjusting transmit power and frequency offset parameters, combined with simulation and real-time monitoring, accurately optimizes signal transmission quality. Using a support vector machine algorithm for secondary optimization, the adjustment process is more scientific and accurate. Bit error rate analysis and stability assessment ensure signal transmission reliability.

[0079] In some embodiments, constructing an extreme environment test model including temperature, humidity, and vibration parameters and collecting drift data of the optimized bit error rate value under extreme conditions includes:

[0080] By building an extreme environment test model, integrating temperature parameters, humidity parameters, and vibration parameters, and simulating test scenarios under various environmental conditions, preliminary environmental parameter combination data is obtained;

[0081] Use acquisition tools to extract key features from preliminary environmental parameter combination data, record the variation range of temperature parameters, humidity parameters, and vibration parameters, and determine the feature data set;

[0082] Based on the feature data set, the support vector machine algorithm is used to model and analyze the correlation between temperature parameters, humidity parameters, vibration parameters and bit error rate values ​​to obtain a bit error rate prediction model;

[0083] If the output value of the bit error rate prediction model exceeds the preset threshold, the collected environmental condition data is recalibrated, and the deviation is corrected in combination with the drift data to determine the bit error rate value after calibration;

[0084] By comparing and analyzing the calibrated bit error rate value with the optimized data, the distribution characteristics of abnormal points in the drift data are extracted to obtain the abnormal drift feature set;

[0085] Based on the abnormal drift feature set, dynamic adjustments are made to the changing trends of environmental conditions, generating corresponding parameter optimization strategies to determine the final bit error rate drift control range;

[0086] Obtain the final bit error rate drift control range, combine the environmental conditions and parameter acquisition data records, iteratively update the test model, and obtain an optimized extreme environment test framework.

[0087] In the above technical solution, the above method aims to comprehensively evaluate the performance stability of narrowband communication equipment under extreme environmental conditions. By constructing a test model containing multiple environmental parameters, simulating real harsh working scenarios, and using advanced data acquisition and analysis technology, it accurately measures and controls the drift of the bit error rate, thereby providing key data support for equipment reliability evaluation and performance optimization.

[0088] Extreme environment test scenario simulation: We build an extreme environment test model, integrating temperature, humidity, and vibration parameters to simulate test scenarios under various environmental conditions and generate preliminary environmental parameter combination data. This step provides diverse environmental conditions for subsequent testing, ensuring that the device's performance under various extreme conditions can be evaluated.

[0089] Key feature data collection: Using collection tools, we extract key features from preliminary environmental parameter data sets. We record the range of temperature, humidity, and vibration parameters, and determine a feature data set. This step, through data collection and feature extraction, provides foundational data for subsequent modeling and analysis.

[0090] Correlation Modeling Analysis: Based on the feature dataset, we use a support vector machine algorithm to model and analyze the correlation between temperature, humidity, and vibration parameters and bit error rate values, thereby obtaining a bit error rate prediction model. This step, by establishing a prediction model, can predict the changing trends of bit error rates under different environmental conditions.

[0091] Data calibration and bias correction: If the output of the prediction model exceeds the preset threshold, the collected environmental condition data is recalibrated and the drift data is combined to perform bias correction. The calibrated bit error rate value is then determined. This step improves the accuracy and reliability of the measurement data through data calibration and bias correction.

[0092] Abnormal drift feature extraction: By comparing and analyzing the calibrated bit error rate values ​​with the optimized data, we extract the distribution characteristics of abnormal points in the drift data and generate an abnormal drift feature set. This step can identify abnormalities in bit error rate drift and provide a basis for further analysis and control.

[0093] Dynamic adjustment and parameter optimization: Based on the abnormal drift signature set, dynamic adjustments are made to the changing trends of environmental conditions, generating corresponding parameter optimization strategies and determining the final BER drift control range. This step, through dynamic adjustment and parameter optimization, effectively controls BER drift and improves stability.

[0094] Iterative test model updates: Obtaining the final BER drift control range, combined with environmental conditions and parameter acquisition data, the test model is iteratively updated to achieve an optimized extreme environment test framework. This step, through iterative model updates, continuously refines the test methodology, improving test accuracy and efficiency.

[0095] This method integrates multiple extreme environmental parameters to comprehensively evaluate device performance under diverse and challenging conditions. Modeling and analysis using support vector machine algorithms improves data processing accuracy and prediction reliability. By extracting abnormal drift signatures and dynamically adjusting parameters, the testing process can be optimized in real time, improving test efficiency.

[0096] In some embodiments, the drift data is feature-fused with wireless channel characteristic data to generate a multi-scenario comprehensive performance evaluation data set, including:

[0097] By acquiring drift data and channel characteristic data from original sources, a pre-processing method is used to clean and unify the formats of the two types of data to obtain a first data set after preliminary processing;

[0098] Based on the first data set, performing feature extraction operations on the drift data and the channel characteristic data respectively to obtain respective key feature values ​​to form a second feature set;

[0099] If the distribution of the eigenvalues ​​in the second feature set meets the preset threshold range, the drift data and the key features of the channel characteristics are integrated by a feature fusion method to determine a third fused feature group;

[0100] If the feature dimensions in the third fused feature group meet the adaptation requirements under multiple scenarios, the fused features are classified using the scenario adaptation rules to obtain a fourth scenario feature set;

[0101] Based on the fourth scenario feature set, a support vector machine algorithm is applied to comprehensively analyze the feature data under multiple scenarios to determine the preliminary results of the performance evaluation and form a fifth evaluation data set;

[0102] Using the fifth evaluation data set and the preset evaluation indicators, the performance in multiple scenarios is quantified to obtain a sixth quantification result set;

[0103] If the indicator values ​​in the sixth quantitative result set meet the preset standards, the data set is finally integrated to determine the complete data results of the multi-scenario comprehensive performance evaluation.

[0104] In the aforementioned technical solution, the method aims to combine drift data obtained from extreme environment testing with wireless channel characteristic data. Through a series of data processing and analysis steps, it generates a comprehensive performance evaluation data set for multiple scenarios. This helps to comprehensively evaluate the performance of narrowband communication equipment in various practical application environments, providing data support for further optimization and reliability improvement of the equipment.

[0105] Data acquisition and preprocessing: Drift data and channel characteristic data are obtained from the original source, and the two types of data are cleaned and formatted using preprocessing methods to obtain the first data set after preliminary processing.

[0106] Feature extraction: Based on the first dataset, feature extraction is performed on the drift data and channel characteristic data, respectively, obtaining their respective key feature values ​​to form a second feature set. Feature extraction simplifies the data, highlights important information, and provides key input for the subsequent fusion process.

[0107] Feature fusion condition determination and execution: If the distribution of feature values ​​in the second feature set meets the preset threshold range, a feature fusion method is used to integrate the key features of the drift data and channel characteristics to determine the third fused feature set. This step ensures that only data that meets the quality requirements is used for fusion, ensuring the reliability of the fusion results.

[0108] Scenario Adaptation: If the feature dimensions in the third fused feature set meet the adaptation requirements for multiple scenarios, the fused features are classified and processed using the scenario adaptation rules to obtain the fourth scenario feature set. Scenario adaptation aligns the fused features with different application scenarios, improving the pertinence and practicality of the evaluation results.

[0109] Comprehensive Analysis and Preliminary Evaluation: Based on the fourth scenario feature set, we applied the support vector machine algorithm to conduct a comprehensive analysis of the feature data from multiple scenarios, determining preliminary performance evaluation results and forming the fifth evaluation dataset. The application of the support vector machine algorithm can effectively classify and evaluate complex data from multiple scenarios, providing preliminary performance evaluation results.

[0110] Performance quantification: Using the fifth evaluation dataset and pre-set evaluation metrics, we quantify performance across multiple scenarios to produce the sixth set of quantitative results. This quantification converts the evaluation results into specific numerical metrics, facilitating comparison and analysis of performance across different scenarios.

[0111] Final Data Integration: If the indicator values ​​in the sixth quantitative result set meet the preset standards, the data set is finally integrated to determine the complete data results of the multi-scenario comprehensive performance evaluation. The final integration step combines all quantitative results that meet the standards into a complete data set, providing a comprehensive summary of the device performance evaluation.

[0112] This method effectively integrates drift data and channel characteristic data from various sources to form a unified evaluation dataset. Through feature extraction and fusion, it highlights key information, reduces data redundancy, and improves evaluation efficiency. It also considers the adaptation requirements of multiple scenarios, making the evaluation results more targeted and practical.

[0113] In some embodiments, if the fluctuation range value in the performance evaluation data exceeds a preset threshold, the channel model weight parameters are adjusted through an iterative algorithm to generate an updated stability score value, including:

[0114] By detecting the fluctuation range of the performance evaluation data, it is determined whether it exceeds the preset threshold and a preliminary range detection result is obtained;

[0115] According to the range detection result, if the fluctuation range exceeds the preset threshold, the current weight parameters of the channel model are obtained to determine the parameter set that needs to be adjusted;

[0116] Adjust the weight parameters in the parameter set one by one through an iterative method to generate an adjusted channel model parameter group;

[0117] Using the adjusted channel model parameter group, recalculate the stability score of the performance evaluation data to obtain an updated score value;

[0118] For the updated score value, if the score value still does not meet the preset standard, the performance evaluation data is further cleaned through the data processing module to obtain an optimized data set;

[0119] Based on the optimized data set, the channel model is re-entered to adjust parameters and update scores to determine whether stability requirements are met and generate the final model optimization results;

[0120] Based on the final model optimization results, the operating configuration of the channel model is updated and the stability score benchmark of the system in subsequent operations is determined.

[0121] In the above technical solution, the method primarily addresses situations where fluctuations in performance evaluation data exceed a preset threshold. An iterative algorithm adjusts the weight parameters of the channel model to generate an updated stability score. This process dynamically optimizes the channel model, improving its stability and ensuring reliable operation of narrowband communication equipment in various scenarios.

[0122] Fluctuation range detection: This process examines the fluctuation range of performance evaluation data to determine whether it exceeds a preset threshold, generating preliminary range detection results. This is the starting point of the entire adjustment process. Only by accurately determining whether the fluctuation range exceeds the standard can we determine whether subsequent parameter adjustments are necessary.

[0123] Parameter Set Determination: Based on the range detection results, if the fluctuation range exceeds the preset threshold, the current weight parameters of the channel model are obtained and the parameter set that needs to be adjusted is determined. This step clarifies the specific parameters that need to be optimized and provides a clear target for subsequent adjustments.

[0124] Weight parameter adjustment: The weight parameters in the parameter set are adjusted through an iterative method to generate an adjusted parameter set. The iterative method can gradually approach the optimal parameter configuration and improve the adjustment accuracy.

[0125] Stability score recalculation: Using the adjusted channel model parameter set, recalculate the stability score of the performance evaluation data to obtain an updated score. This step verifies the effectiveness of the parameter adjustment by recalculating the score.

[0126] Further data cleaning: If the updated score still does not meet the preset standard, the performance evaluation data is further cleaned through the data processing module to obtain an optimized data set. Further data cleaning can remove possible outliers or noise, improving data quality.

[0127] Model re-input and optimization: Based on the optimized data set, the channel model is re-input to adjust parameters and update scores to determine whether stability requirements are met, and generate the final model optimization results. This step continuously improves the stability of the model through iterative optimization.

[0128] Runtime Configuration Update: Based on the final model optimization results, the channel model's operational configuration is updated to determine the stability score benchmark for the system during subsequent operations. This step applies the optimization results to actual system operations to ensure device stability and reliability.

[0129] This method dynamically adjusts the channel model's weight parameters through an iterative algorithm, enabling timely response to performance fluctuations and improving the model's adaptability and stability. Adjusting weight parameters individually and recalculating the stability score ensures the accuracy and effectiveness of these adjustments. Further cleaning of the performance evaluation data improves its accuracy and reliability.

[0130] In some embodiments, a multi-dimensional performance analysis graph is generated based on the stability score value, and bit error rate exceeding standard data and frequency drift exceeding standard data under a specific environment are extracted, including:

[0131] By collecting stability score-related data and using a preset threshold to preliminarily screen the score data, a preliminarily screened score data set is obtained;

[0132] Based on the initially screened scoring dataset, and in response to the requirements for building a multidimensional performance graph, we apply performance analysis methods to generate corresponding multidimensional performance graph data and identify the key performance dimensions in the graph.

[0133] Based on multi-dimensional performance graph data and the influence of specific environmental values, we analyze the effect of environmental impact on performance dimensions and obtain performance fluctuation data under environmental influences.

[0134] For performance fluctuation data, if the fluctuation data exceeds the preset threshold range, the bit error rate data and frequency drift value are separated through data extraction method to obtain the separated abnormal data group;

[0135] Based on the separated abnormal data group, the support vector machine algorithm is used to classify the data points with bit error rate data and frequency drift values ​​exceeding the standard, and the distribution characteristics of the data points exceeding the standard are determined;

[0136] By further processing the distribution characteristics of the data points that exceed the standard, combining the score correlation and drift detection methods, the correlation pattern between the frequency drift value and the stability score is analyzed to determine the final abnormal correlation result;

[0137] According to the final anomaly correlation result, corresponding multi-dimensional performance map update data is generated for the key points in the anomaly correlation result to obtain the updated performance analysis map.

[0138] In the above technical solution, the above method is mainly based on the stability score value. Through the form of a multi-dimensional performance analysis graph, the bit error rate exceeding standard data and the frequency drift exceeding standard data under specific environments are extracted, so as to more intuitively display the performance of narrowband communication equipment in different environments, and provide a visual basis for performance optimization and fault diagnosis.

[0139] Data collection and preliminary screening: Data related to stability scoring is collected and initially screened using pre-set thresholds to obtain a pre-screened scoring dataset. This step ensures the relevance and usability of the data for subsequent analysis and removes data that clearly does not meet the requirements.

[0140] Multidimensional Performance Map Data Generation: Based on the initially screened scoring dataset, we apply performance analysis methods to generate the corresponding multidimensional performance map data and identify the key performance dimensions within the map. This step extracts key performance dimensions through performance analysis, providing the data foundation for building the multidimensional performance map.

[0141] Environmental Impact Analysis: Based on multi-dimensional performance graph data and the impact of specific environmental values, we analyze the impact of environmental influences on performance dimensions and obtain performance fluctuation data under environmental influences. This step reveals the specific impact of environmental factors on device performance and provides guidance for subsequent abnormal data extraction.

[0142] Abnormal data separation: For performance fluctuation data, if the fluctuation data exceeds the preset threshold range, the bit error rate data and frequency drift value are separated through data extraction method to obtain the separated abnormal data group.

[0143] Classification of data points exceeding the standard: Based on the separated abnormal data group, the support vector machine algorithm is used to classify the data points exceeding the standard for bit error rate data and frequency drift values, and to determine the distribution characteristics of the data points exceeding the standard.

[0144] Correlation Pattern Analysis: We further process the distribution characteristics of the out-of-specification data points and, combining score correlation with drift detection methods, analyze the correlation patterns between frequency drift values ​​and stability scores to determine the final abnormal correlation results. This step, through correlation analysis, reveals the inherent connection between frequency drift and stability scores, providing key information for performance optimization.

[0145] Performance Graph Update: Based on the final anomaly correlation results, corresponding multidimensional performance graph update data is generated for key points in the anomaly correlation results, obtaining an updated performance analysis graph. This step feeds the analysis results back into the multidimensional performance graph, enabling dynamic updating and optimization of the graph.

[0146] This method uses multidimensional performance analysis graphs to intuitively display device performance in different environments, facilitating rapid problem identification. Data extraction and classification algorithms accurately isolate and analyze abnormal data, improving the accuracy of anomaly detection. Dynamically updating the performance graph based on analysis results promptly reflects changes in device performance, supporting continuous optimization.

[0147] In some embodiments, inputting the exceeding-standard data into a linear regression model to predict a quantitative risk assessment index of the device in a complex environment includes:

[0148] By acquiring the exceeding standard data from the monitoring system, the data is cleaned and formatted by using pre-processing means to obtain a preliminarily processed data set;

[0149] Based on the data set after preliminary processing, a linear regression model is used for training, and parameters are adjusted according to the data characteristics to determine the model's predictive ability;

[0150] If the deviation between the predicted value of the trained model and the actual risk value exceeds a preset threshold, feature selection is performed on the data set, and the input features are readjusted to obtain an optimized data subset;

[0151] Using the optimized data subset, the linear regression model is re-run to simulate variables in a complex environment and obtain preliminary quantitative indicators of equipment risk.

[0152] Based on the preliminary quantitative indicators, a weighted calculation is performed in combination with environmental impact factors. If the weighted indicators exceed the safety range, the indicators are recalibrated to determine the final risk assessment value;

[0153] After obtaining the final risk assessment value, compare and analyze the value with historical data to determine the risk change trend of the equipment in a complex environment;

[0154] By matching the risk change trend with the preset early warning rules, if the matching result shows that the risk is increasing, a corresponding early warning signal is generated to obtain the dynamic monitoring results of the equipment operation status.

[0155] In the aforementioned technical solution, the method primarily utilizes processed out-of-standard data and a linear regression model to predict the quantitative risk indicators of narrowband communication equipment in complex environments. This process involves not only data preprocessing and model training, but also feature selection, risk quantification, and dynamic monitoring, aiming to provide a scientific, quantitative basis for equipment risk management.

[0156] Data acquisition and preprocessing: Obtain excess data from the monitoring system, and use preprocessing methods to clean and format it to obtain a preliminary processed data set.

[0157] Model training and parameter adjustment: Based on the initially processed data set, a linear regression model is trained. Parameters are adjusted based on the data characteristics to determine the model's predictive capabilities. This step improves prediction accuracy through model training and parameter optimization.

[0158] Model prediction bias check: If the deviation between the trained model's prediction and the actual risk value exceeds a preset threshold, feature selection is performed on the data set, and the input features are readjusted to obtain an optimized data subset. This step improves the model's predictive performance through feature selection and data optimization.

[0159] Risk Quantification Simulation: Using the optimized data subset, we re-run the linear regression model to simulate the variables in a complex environment and obtain preliminary quantitative indicators of device risk. This step provides preliminary quantitative results of device risk by simulating complex environmental variables.

[0160] Risk Assessment Calibration: A weighted calculation is performed based on preliminary quantitative indicators, combined with environmental impact factors. If the weighted indicators exceed the safety range, a secondary calibration is performed to determine the final risk assessment value. This step, through weighted calculation and calibration, ensures the rationality of the risk assessment value.

[0161] Risk Trend Analysis: After obtaining the final risk assessment value, we compare and analyze it with historical data to determine the risk trend of the device in complex environments. This step provides dynamic information on device risk through trend analysis.

[0162] Early warning signal generation and dynamic monitoring: Risk trends are matched against early warning rules. If the matching results indicate an increase in risk, a corresponding early warning signal is generated, providing dynamic monitoring of the equipment's operating status. This step enables real-time monitoring of the equipment's operating status through the generation of early warning signals.

[0163] This method, using a linear regression model, quantifies equipment risks into specific indicators, providing a scientific basis for risk management and decision-making. Combined with risk trend analysis and early warning rules, it enables real-time monitoring of equipment operating status, improving the timeliness and effectiveness of equipment management. Feature selection and data optimization ensure the accuracy and reliability of model predictions.

[0164] According to another aspect of the present invention, a narrowband communication device detection device is provided, based on the above method, and includes:

[0165] An acquisition module is used to acquire a pre-established wireless channel model data set, wherein the data set includes attenuation values ​​and interference intensity distribution data under different environments;

[0166] A simulation module, configured to construct a multipath propagation simulation model based on the attenuation value and the interference intensity distribution data, obtain a delay spread range value by using time domain analysis, and obtain an amplitude distortion range value by combining frequency domain analysis;

[0167] An evaluation module, configured to input the measured environmental noise data, the delay spread range value, and the amplitude distortion range value into an adaptive filtering algorithm to generate signal quality evaluation data after filtering;

[0168] A first determination module is configured to adjust a transmit power parameter and a frequency offset parameter to obtain an optimized bit error rate value if the signal quality evaluation data does not reach a preset signal-to-noise ratio threshold;

[0169] A construction module is used to construct an extreme environment test model including temperature, humidity and vibration parameters, and collect drift data of the optimized bit error rate value under extreme conditions;

[0170] A fusion module is used to perform feature fusion on the drift data and the wireless channel characteristic data to generate a multi-scenario comprehensive performance evaluation data set;

[0171] A second determination module is configured to adjust the channel model weight parameters through an iterative algorithm to generate an updated stability score value if the fluctuation range value in the performance evaluation data exceeds a preset threshold;

[0172] An extraction module, configured to generate a multi-dimensional performance analysis graph based on the stability score value, and extract bit error rate exceeding standard data and frequency drift exceeding standard data under a specific environment;

[0173] The prediction module is used to input the exceeding-standard data into a linear regression model to predict the risk assessment quantitative index of the equipment in a complex environment.

[0174] In the above technical solution, in order to better utilize the above method, the present application proposes a narrowband communication device detection device, wherein each module corresponds to each step of the above method, and its specific principles have been described above and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0175] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0176] Figure 1 This is a flow chart of an embodiment of a narrowband communication device detection method of the present invention;

[0177] Figure 2 It is a structural diagram of an embodiment of a narrowband communication device detection device of the present invention. DETAILED DESCRIPTION

[0178] The present invention will be described in further detail below with reference to the accompanying drawings and examples. It is particularly noted that the following examples are intended only to illustrate the present invention and are not intended to limit the scope of the present invention. Similarly, the following examples are only some embodiments of the present invention and are not intended to be exhaustive. All other embodiments obtained by those of ordinary skill in the art without creative effort are intended to fall within the scope of protection of the present invention.

[0179] Example 1

[0180] See also Figure 1 , a narrowband communication device detection method, the method comprising:

[0181] S1. Acquire a pre-established wireless channel model data set, wherein the data set includes attenuation values ​​and interference intensity distribution data under different environments;

[0182] In this embodiment, S1, a pre-established wireless channel model data set is obtained, where the data set includes attenuation values ​​and interference intensity distribution data under different environments, including:

[0183] S11. Acquire a pre-established wireless channel model data set, where the data set includes attenuation values ​​and interference intensity distribution data under different environments, including:

[0184] S12. Obtain a pre-built wireless channel model data set, classify and organize data under various environments, group attenuation values ​​and interference intensity information in the data set according to environmental differences, and obtain a classified environmental data set;

[0185] S13, using the classified environmental data set, extracting features based on the attenuation values ​​and attenuation features in each set of environmental data, normalizing the attenuation features using signal processing technology, and determining a standardized attenuation feature set;

[0186] S14. Based on the standardized attenuation feature set, combined with the interference intensity and intensity distribution information, if the intensity distribution in a certain environment group exceeds a preset threshold, filtering the interference intensity data of the group to obtain a filtered interference data set;

[0187] S15. Analyze the wireless channel characteristics in each set of environmental data by combining the filtered interference data set with the channel characteristic data, use a support vector machine algorithm to perform correlation modeling on the channel characteristics and the interference data, and determine a correlation strength distribution model.

[0188] S16. Evaluate the data coverage of each set of environmental data based on the correlation strength distribution model. If the data coverage is lower than a preset standard, perform additional sampling on the set of data to obtain an expanded data coverage set.

[0189] S17, using the expanded data coverage set and combining distribution information under various environments, performing a secondary calibration on the distribution characteristics of each set of environmental data to determine a calibrated distribution characteristic set;

[0190] S18. Based on the calibrated distribution feature set, a comprehensive analysis is performed on the attenuation characteristics and intensity distribution of the wireless channel in various environments. A clustering algorithm is used to group and integrate the feature set to obtain the final channel characteristic distribution result.

[0191] For example, to obtain a pre-established wireless channel model data set and process data containing attenuation values ​​and interference intensity distributions in different environments, the following implementation can be used. First, assume a database system stores wireless channel data for various environments, such as urban, suburban, and indoor environments. The data includes attenuation values ​​(in decibels, dB) and interference intensity (in milliwatts, mW). For an urban environment, the attenuation value data range is set to -50 to -120 dB, and the interference intensity distribution range is 0.1 to 5 mW. The system automatically extracts this data by calling a database interface and sorts it by timestamp to ensure timely data acquisition. Next, the raw attenuation data is smoothed using a mean filtering algorithm. The algorithm formula is: Average attenuation value = Σ(attenuation value i) / N, where N is the number of sampling points. Assuming N = 100, the calculated average attenuation value for the urban environment is -85.5 dB. Subsequently, the interference intensity distribution data was statistically analyzed using the normal distribution fitting method to calculate the mean and variance, assuming the mean was 2.3 mW and the variance was 0.8 mW. 2 The fitting results predict peak areas of interference intensity. The analysis shows that urban environments, due to high building density and multiple user access, experience higher attenuation and significant interference intensity fluctuations. Compared to suburban environments (average attenuation of -70 dB, average interference of 1.5 mW), channel quality is poorer. Finally, the processed data is stored in a structured format (such as JSON) for easy access. Urban environment data can be tagged as {"environment":"urban", "average attenuation":-85.5, "average interference":2.3}. The system automatically updates the data every 24 hours to ensure model timeliness.

[0192] S2. Construct a multipath propagation simulation model based on the attenuation value and interference intensity distribution data, obtain a delay spread range value by using time domain analysis, and obtain an amplitude distortion range value by combining frequency domain analysis;

[0193] In this embodiment, S2, constructing a multipath propagation simulation model based on the attenuation value and interference intensity distribution data, using time domain analysis to obtain a delay spread range value, and combining frequency domain analysis to obtain an amplitude distortion range value, includes:

[0194] S21, obtaining initial input from attenuation value data and interference intensity value, and preprocessing the distributed data source using a data processing layer to obtain a standardized basic data set;

[0195] S22. Construct a multipath propagation model based on the standardized basic data set, initialize it using preset parameters in the simulation model library, and determine the preliminary structure of the model;

[0196] S23. Calculate the delay spread value using a time domain analysis method for the multipath propagation model of the preliminary structure to obtain the distribution characteristics of the delay spread range;

[0197] S24. Evaluate the amplitude distortion value by using the distribution characteristics of the delay spread range in combination with a frequency domain analysis method to obtain the range boundary of the amplitude distortion;

[0198] S25. If the range boundary of the amplitude distortion exceeds the preset threshold, the basic data set is recalibrated through the data processing layer to determine whether the calibrated data meets the constraints of the model building method;

[0199] S26. Readjust the parameters of the multipath propagation model based on the calibrated data, optimize the model using a support vector machine algorithm, and obtain the final simulation results.

[0200] S27. Through the final simulation results, the range outputs of the delay spread value and the amplitude distortion value are integrated to determine the signal characteristic distribution in the multipath propagation environment.

[0201] For example, in the process of constructing a multipath propagation simulation model and performing time domain and frequency domain analysis, based on the attenuation value and interference intensity distribution data, assuming that the attenuation value range is -10 dB to -30 dB and the interference intensity distribution conforms to a Gaussian distribution with a mean of -20 dB and a standard deviation of 5 dB, the Monte Carlo method is used to generate 10,000 random sample points to construct a multipath channel model. The attenuation value and interference intensity of each path are determined by random sampling, and the number of paths is set to 5, corresponding to different delay points. Subsequently, in the time domain analysis, the delay spread range is calculated, assuming that the delays of each path are 0 ns, 50 ns, 100 ns, 150 ns, and 200 ns, respectively. The delay spread is calculated using the time domain impulse response, i.e., the difference between the maximum delay and the minimum delay is 200 ns - 0 ns = 200 ns. Combined with the power-weighted average delay formula, the average delay is approximately 90 ns, and the delay spread range is determined to be 0 ns to 200 ns. Next, in the frequency domain analysis, the time domain signal is converted to a frequency domain signal using a Fourier transform. Assuming a signal bandwidth of 10 MHz, the amplitude distortion of each frequency component is calculated, resulting in an amplitude distortion range of 0.5 to 1.5. The distortion value is derived by comparing the amplitude ratio of the ideal signal to the actual signal. For example, at 5 MHz, the amplitude ratio is 0.8. The distortion range is determined by integrating data from all frequency points. To establish a strict logical relationship, the delay spread range and amplitude distortion range are input into the subsequent channel equalization algorithm design. For example, equalizer parameters are adjusted based on the minimum mean square error criterion to ensure signal recovery accuracy exceeds 95%. This process forms a complete technical chain, from data construction to analysis to application, ensuring the accuracy and practicality of the simulation model.

[0202] S3, inputting the measured environmental noise data, the delay spread range value, and the amplitude distortion range value into an adaptive filtering algorithm to generate signal quality assessment data after filtering;

[0203] In this embodiment, S3, inputting the measured environmental noise data, the delay spread range value, and the amplitude distortion range value into an adaptive filtering algorithm to generate filtered signal quality assessment data, includes:

[0204] S31. Forming an initial input parameter group by combining the collected measured environmental noise data with the delay spread range value and the amplitude distortion range value;

[0205] S32. Preliminarily process the measured environmental noise data using an adaptive filtering algorithm based on the initial input parameter group to obtain a filtered first signal data set;

[0206] S33. Obtain delay spread characteristics and amplitude distortion characteristics of the first signal data set, and determine whether the characteristic distribution meets a preset threshold range. If the characteristic distribution exceeds the threshold range, perform secondary filtering on the first signal data set to obtain a second signal data set.

[0207] S34. Extracting signal quality-related indicators from the second signal data set, determining whether the signal quality indicators reach a preset standard value, and if not, adjusting parameter configuration of the adaptive filtering algorithm to generate a third signal data set;

[0208] S35. Calculate key parameter values ​​in the signal quality assessment data based on the third signal data set to obtain preliminary structured results of the assessment data;

[0209] S36. Using the structured results and combining them with a pre-established signal quality assessment model, generate final signal quality assessment data and determine the integrity and accuracy of the assessment data;

[0210] S37. Extract the distribution characteristics of the outliers from the final signal quality evaluation data to determine whether the outlier distribution affects the overall signal quality. If the impact exceeds a preset range, mark the outliers and record them in the evaluation data.

[0211] For example, in processing measured ambient noise data and the delay spread and amplitude distortion range values, the data acquisition system first acquires the measured ambient noise data. Assuming a sampling frequency of 44.1kHz and a collection duration of 10 seconds, data from 441,000 sampling points is obtained, with an average noise signal power of 0.5W. Next, the delay spread range is set to 0.1ms to 5ms, and the amplitude distortion range is set to -3dB to +3dB. These parameters are derived from an analysis of typical urban ambient noise characteristics. This data is then input into an adaptive filtering algorithm using the minimum mean square error (LMS) algorithm, with a step size of 0.01 and a filter order of 32. Through iterative calculations, the filter coefficients gradually converge, and the mean square value of the error signal decreases from an initial 0.25 to 0.05, demonstrating a significant filtering effect. During the filtering process, the algorithm compensates for the delay and adjusts the amplitude of the noise signal. For example, it performs reverse delay processing on a signal with a 2ms delay and performs gain correction on a signal with a +2dB amplitude deviation, ultimately generating the filtered signal. Finally, quality assessment data is calculated based on the filtered signal, using the signal-to-noise ratio (SNR) as the evaluation metric. The calculated SNR for the filtered signal is 15.2dB, a significant improvement from the 8.7dB before filtering. Spectral analysis also reveals a reduction of approximately 40% in high-frequency noise components, demonstrating that filtering effectively suppresses environmental interference.

[0212] S4. If the signal quality evaluation data does not reach a preset signal-to-noise ratio threshold, adjust the transmit power parameter and the frequency offset parameter to obtain an optimized bit error rate value;

[0213] In this embodiment, S4, if the signal quality evaluation data does not reach a preset signal-to-noise ratio threshold, adjusting the transmit power parameter and the frequency offset parameter to obtain an optimized bit error rate value, includes:

[0214] S41, by processing the initial evaluation data of the signal quality, obtaining specific data of the signal-to-noise ratio value, and determining whether it reaches a preset threshold;

[0215] S42. If the signal-to-noise ratio value does not reach the preset threshold, perform preliminary parameter adjustments on the transmit power and frequency offset to obtain a first set of adjusted parameter data;

[0216] S43. Simulate signal transmission using a pre-established simulation model based on the first set of parameter data to obtain simulated signal quality data;

[0217] S44. If the simulated signal quality data still does not meet the preset threshold, the transmit power and frequency offset are optimized and adjusted twice using a support vector machine algorithm to obtain a second set of parameter data.

[0218] S45. Based on the second set of parameter data, the signal transmission environment is monitored in real time to obtain an optimized bit error rate value;

[0219] S46. By analyzing the bit error rate value, determine whether the stability of signal transmission meets the preset standard and obtain a final evaluation result;

[0220] S47. If the final evaluation result shows that the stability is insufficient, fine-tune the second set of parameter data, obtain the third set of parameter data, and loop through the simulation and monitoring process to determine the optimal configuration of signal transmission.

[0221] For example, during the signal quality assessment and optimization process, suppose the initial signal quality assessment data, measured through signal-to-noise ratio (SNR), is 15.2 dB. However, the preset SNR threshold is 20 dB, clearly not meeting the requirement. The system automatically triggers the parameter adjustment mechanism. First, by analyzing historical data and the current ambient noise level, the system calculates that the transmit power needs to be increased by 1.5 times, from 10 mW to 15 mW. Simultaneously, the spectrum analysis algorithm detects that the current frequency offset is 0.3 kHz, exceeding the allowable range by 0.1 kHz and requiring adjustment to the target value of 0.05 kHz. The adjustment process uses the least squares method to fit the frequency offset curve to ensure a smooth transition. The calculation formula is Δf = (f_current - f_target) * 0.8, where f_current is the current offset value and f_target is the target value. After three iterations, the frequency offset stabilizes at 0.06 kHz, close to the ideal value. The system then re-collected signal data and evaluated the optimization results based on the bit error rate (BER). Using Monte Carlo simulation, 10,000 data transmissions were simulated, resulting in a BER of 0.0012 after optimization, a decrease of approximately 78% compared to the pre-optimization BER of 0.0056, demonstrating the effectiveness of the adjustment. Further analysis showed that the signal-to-noise ratio had increased to 19.8 dB, which, while close to the threshold, still fell short of the target. The system automatically recorded the optimized parameters and linked them to the environmental noise database, forming a self-learning model. This model provides a reference for similar scenarios and ensures continuous optimization of the parameter adjustment logic. By integrating with business scenarios, such as in high-noise industrial environments, the system can prioritize historically optimized parameters, reducing calculation time and improving response efficiency.

[0222] S5. Construct an extreme environment test model including temperature, humidity, and vibration parameters, and collect drift data of the optimized bit error rate value under extreme conditions;

[0223] In this embodiment, S5, constructing an extreme environment test model including temperature, humidity, and vibration parameters, and collecting drift data of the optimized bit error rate value under extreme conditions, includes:

[0224] S51. By building an extreme environment test model, integrating temperature parameters, humidity parameters, and vibration parameters, and simulating test scenarios under various environmental conditions, preliminary environmental parameter combination data is obtained;

[0225] S52. Extract key features from the preliminary environmental parameter combination data using an acquisition tool, record the variation ranges of the temperature parameter, humidity parameter, and vibration parameter, and determine a feature data set;

[0226] S53. Based on the feature data set, a support vector machine algorithm is used to model and analyze the correlation between the temperature parameter, the humidity parameter, and the vibration parameter and the bit error rate value to obtain a bit error rate prediction model;

[0227] S54. If the output value of the bit error rate prediction model exceeds a preset threshold, the collected environmental condition data is recalibrated, and the drift data is combined to perform deviation correction, and the calibrated bit error rate value is determined;

[0228] S55. By comparing and analyzing the calibrated bit error rate value with the optimized data, the distribution characteristics of abnormal points in the drift data are extracted to obtain an abnormal drift feature set;

[0229] S56. Based on the abnormal drift feature set, dynamically adjust the changing trend of environmental conditions, generate a corresponding parameter optimization strategy, and determine the final bit error rate drift control range;

[0230] S57. Obtain the final bit error rate drift control range, combine the environmental conditions and parameter acquisition data records, iteratively update the test model, and obtain an optimized extreme environment test framework.

[0231] For example, when building an extreme environment test model that includes temperature, humidity, and vibration parameters, environmental data is first collected through a sensor network. Assuming the temperature range is set to -40°C to 85°C, the humidity range is 10% to 95%, the vibration frequency range is 5Hz to 200Hz, and the vibration acceleration is 1g to 10g. Data is collected at a frequency of 10 times per second for 24 hours, resulting in a dataset containing 864,000 data points. This data is then used to build a test model using a multivariate regression analysis algorithm, with temperature, humidity, and vibration as independent variables and bit error rate (BER) as the dependent variable. The relationship model is established using the formula BER = a*temperature + b*humidity + c*vibration + d, where a, b, c, and d are regression coefficients. Least squares fitting yields the hypothetical values ​​of a = 0.002, b = 0.0015, c = 0.003, and d = 0.01. After the model was built, data on the drift of the optimized bit error rate (BER) under extreme conditions was collected. The test conditions were set to 85°C, 95% humidity, and 10g vibration acceleration. Over a 48-hour period, the BER was observed to have drifted from an initial 0.05% to 0.08%, a drift of 0.03%. Subsequently, the drift data was analyzed using time series analysis to calculate a drift rate of approximately 0.0006% per hour. The influence of vibration frequency on the BER was analyzed using a Fourier transform, revealing that a vibration frequency of 100Hz had the greatest impact, contributing 35%. Finally, the analysis results were fed back into the model optimization process. By adjusting the regression coefficient (for example, increasing c from 0.003 to 0.004) to accommodate the drift characteristics under extreme conditions, the model prediction error was kept within 5%.

[0232] S6. Fusing the drift data with the wireless channel characteristic data to generate a multi-scenario comprehensive performance evaluation data set;

[0233] In this embodiment, S6, performing feature fusion on the drift data and the wireless channel characteristic data to generate a multi-scenario comprehensive performance evaluation data set, includes:

[0234] S61, obtaining drift data and channel characteristic data from original sources, and using a preprocessing method to clean and unify the formats of the two types of data to obtain a first data set after preliminary processing;

[0235] S62: Based on the first data set, perform feature extraction operations on the drift data and the channel characteristic data respectively to obtain their respective key feature values ​​to form a second feature set;

[0236] S63: If the distribution of the feature values ​​in the second feature set meets the preset threshold range, integrating the key features of the drift data and the channel characteristics through a feature fusion method to determine a third fused feature group;

[0237] S64: If the feature dimensions in the third fused feature group meet the adaptation requirements in multiple scenarios, classify the fused features using the scenario adaptation rule to obtain a fourth scenario feature set;

[0238] S65. Based on the fourth scenario feature set, apply a support vector machine algorithm to perform a comprehensive analysis on the feature data under multiple scenarios, determine preliminary results of the performance evaluation, and form a fifth evaluation data set;

[0239] S66. Quantify the performance in multiple scenarios using the fifth evaluation data set and preset evaluation indicators to obtain a sixth quantification result set.

[0240] S67. If the indicator values ​​in the sixth quantitative result set meet the preset standards, the data set is finally integrated to determine the complete data results of the multi-scenario comprehensive performance evaluation.

[0241] Exemplarily, drift data and channel characteristic data are obtained from the operation recording system and wireless channel monitoring system of narrowband communication equipment. Drift data includes metrics such as frequency drift and time drift, while channel characteristic data covers channel gain, noise power, and interference intensity. This data may contain duplicate records, erroneous data, and inconsistent formats. Data cleaning techniques are used to remove duplicate and obviously erroneous data points. Data in different formats are also converted to a standard format, such as standardizing frequency drift data to units of hertz (Hz), time drift data to units of seconds (s), and channel gain data to units of decibel milliwatts (dBm). This yields a preliminarily processed first dataset. Feature extraction operations are then performed on the preliminarily processed drift data and channel characteristic data. For drift data, key feature values ​​are extracted, such as the frequency drift rate (frequency change per unit time) and time drift (time deviation within a specific time interval). For channel characteristic data, key feature values ​​are extracted, such as the average and fluctuation range of channel gain, the mean and peak values ​​of noise power, and the frequency and intensity level of interference intensity. A second feature set containing these feature values ​​is then formed. Preset threshold ranges for eigenvalue distributions, such as ±10 Hz / s for frequency drift rate and -100 dBm to 0 dBm for channel gain. If the eigenvalue distributions in the second feature set meet these preset threshold ranges, feature fusion methods, such as principal component analysis (PCA) or linear combination, are used to integrate the drift data and key channel characteristics to determine a third fused feature set containing these fused features. These fused features can more comprehensively reflect the device's overall performance in a specific scenario. Consider the adaptation requirements for multiple scenarios. For example, narrowband communication equipment may be used in different scenarios, such as indoor, outdoor, and mobile, each with different requirements for frequency stability and channel adaptability. Preset adaptation criteria for feature dimensions are also established. For example, in indoor scenarios, sensitivity to time drift is low, but adaptability to channel gain variations is high. If the feature dimensions in the third fused feature set meet these adaptation requirements for multiple scenarios, scenario adaptation rules are used to classify the fused features based on the scenario type in which the device is located, resulting in a fourth scenario feature set tailored to each scenario. Based on the fourth scenario feature set, a support vector machine algorithm is used to build a classification model. The fused feature data is then input to perform a comprehensive analysis of the feature data across multiple scenarios. The trained model determines preliminary performance evaluation results for the device in different scenarios, categorizing performance as excellent, good, fair, or poor, for example. This generates a fifth evaluation dataset containing these preliminary evaluation results. Pre-set evaluation metrics include communication stability indicators (such as packet loss rate and bit error rate) and transmission efficiency indicators (such as data transmission rate and latency) for the device in different scenarios.Using the fifth evaluation data set and combining these preset evaluation indicators, the performance in multiple scenarios is quantified. For example, the communication stability indicator is quantified as a specific packet loss rate percentage and bit error rate value, and the transmission efficiency indicator is quantified as the amount of data transmitted per second and the average latency value, thereby obtaining a sixth quantified result set. Preset performance evaluation qualification criteria are set, such as a packet loss rate of less than 5% and a bit error rate of less than 10^-6 in the communication stability indicator, and a data transmission rate of greater than 10kbps and a latency of less than 500ms in the transmission efficiency indicator. If the indicator values ​​in the sixth quantified result set meet these preset criteria, the data set is finally integrated, and the quantified results for different scenarios are organized according to a specific format and logical relationship to determine the complete data results of the comprehensive multi-scenario performance evaluation, including detailed performance indicators, evaluation levels, and comprehensive performance scores for the device in each scenario.

[0242] S7. If the fluctuation range value in the performance evaluation data exceeds a preset threshold, adjusting the channel model weight parameter through an iterative algorithm to generate an updated stability score value;

[0243] In this embodiment, S7, if the fluctuation range value in the performance evaluation data exceeds a preset threshold, adjusting the channel model weight parameter through an iterative algorithm to generate an updated stability score value, including:

[0244] S71. Detect the fluctuation range of the performance evaluation data to determine whether it exceeds a preset threshold, thereby obtaining a preliminary range detection result.

[0245] S72. If the fluctuation range exceeds a preset threshold according to the range detection result, obtain the current weight parameters of the channel model and determine the parameter set that needs to be adjusted;

[0246] S73, adjusting the weight parameters in the parameter set one by one through an iterative method to generate an adjusted channel model parameter group;

[0247] S74. Recalculate the stability score of the performance evaluation data using the adjusted channel model parameter group to obtain an updated score value;

[0248] S75. If the updated score value still does not meet the preset standard, the performance evaluation data is further cleaned by the data processing module to obtain an optimized data set;

[0249] S76. Re-enter the channel model based on the optimized data set to adjust parameters and update the score, determine whether the stability requirements are met, and generate the final model optimization result;

[0250] S77. Based on the final model optimization results, update the operating configuration of the channel model and determine the stability score benchmark for the system in subsequent operations.

[0251] For example, in the performance evaluation data processing, suppose the channel performance fluctuation range value we collected is 15.2, and the preset threshold is set to 10.0. Through automatic comparison by the system, it is found that the fluctuation range value exceeds the threshold, triggering the subsequent adjustment mechanism. For the adjustment of the channel model weight parameters, the system adopts a gradient descent iterative algorithm. The initial weight parameter is set to 0.5, the learning rate is 0.01, and the goal is to minimize the loss function of the fluctuation range and stability, defined as L=|(fluctuation range-target value)|, and the target value is set to 8.0. The loss value calculated in the first iteration is L=|(15.2-8.0)|=7.2, and the gradient descent updates the weight to 0.5-0.01*7.2=0.428. The system records the updated weight and recalculates the channel model output to obtain a new fluctuation range value of 12.5. The second iteration calculates the loss value to be L = |(12.5-8.0)| = 4.5, and the updated weight is 0.428-0.01*4.5 = 0.383. After the fifth iteration, the fluctuation range value drops to 9.1, and the loss value is less than 1.0, meeting the convergence condition, and the system stops iterating. Finally, the updated stability score is generated using the formula S = 100-10*(fluctuation range - target value). Substituting the data into S = 100-10*(9.1-8.0) = 89.0. The system stores this score in the database and compares it with historical scores to analyze stability trends. If the score falls below 90.0 for three consecutive times, an alarm mechanism is triggered, notifying the relevant service modules to optimize channel configuration parameters to ensure communication quality.

[0252] S8. Generate a multi-dimensional performance analysis graph based on the stability score value, and extract the bit error rate exceeding standard data and the frequency drift exceeding standard data under a specific environment;

[0253] In this embodiment, S8, generating a multi-dimensional performance analysis graph based on the stability score value, extracting bit error rate exceeding standard data and frequency drift exceeding standard data under a specific environment, including:

[0254] S81. Collecting stability score-related data and performing preliminary screening on the score data using a preset threshold to obtain a preliminary screened score data set;

[0255] S82. Based on the initially screened scoring data set and in accordance with the requirements for constructing a multidimensional performance graph, apply a performance analysis method to generate corresponding multidimensional performance graph data and determine the key performance dimensions in the graph;

[0256] S83. Analyze the effect of environmental impact on performance dimensions based on the multi-dimensional performance map data and the influence of specific environmental values, and obtain performance fluctuation data under environmental influences.

[0257] S84. For the performance fluctuation data, if the fluctuation data exceeds a preset threshold range, the bit error rate data and the frequency drift value are separated by a data extraction method to obtain a separated abnormal data group;

[0258] S85. Based on the separated abnormal data group, a support vector machine algorithm is used to classify the data points of the bit error rate data and the frequency drift value that exceed the standard, and the distribution characteristics of the data points that exceed the standard are determined;

[0259] S86. By further processing the distribution characteristics of the data points exceeding the standard, combining the score correlation and drift detection method, analyzing the correlation pattern between the frequency drift value and the stability score, and determining the final abnormal correlation result;

[0260] S87. Based on the final abnormality correlation result, corresponding multi-dimensional performance map update data is generated for key points in the abnormality correlation result, and an updated performance analysis map is obtained.

[0261] For example, the technical implementation of generating a multidimensional performance analysis graph based on stability scores and extracting data on bit error rate and frequency drift exceeding standards in specific environments can be implemented through the following fusion method. First, assume that we collect stability scores for devices in an environment with a temperature range of -20°C to 60°C and a humidity range of 30% to 90%. The scores range from 0 to 100, with scores below 60 considered unstable. Using the Python NumPy library, we matrix-process the scores to generate a three-dimensional array containing temperature, humidity, and score values. For example, the score matrix might be [[25, 50, 75], [30, 55, 62], [35, 60, 58]]. Using Matplotlib, we plot a three-dimensional scatter plot to visually display the changing trends of the scores as they change with environmental variables. Analysis reveals that areas with scores below 60 are concentrated in high-temperature and high-humidity environments. For example, a score of 58 is obtained at a temperature of 50°C and a humidity of 85%. Next, we extracted data with bit error rates exceeding the specified limit. We set a bit error rate threshold of 0.001. The collected data showed a bit error rate of 0.0015 at a temperature of 55°C and a humidity of 88%, exceeding the threshold. We used the K-means clustering algorithm to classify the bit error rate data and calculated that the exceeding data were primarily distributed in the high-temperature and high-humidity range, with the cluster center at (temperature 52°C, humidity 87%). We also recorded five exceeding-the-standard data, accounting for 20% of the total data. We also extracted data with frequency drift exceeding the specified limit, setting a threshold of ±10Hz. The measured data showed a frequency drift of 12Hz at a temperature of 58°C, exceeding the threshold. A linear regression algorithm was used to analyze the correlation between frequency drift and temperature, yielding a regression coefficient of 0.8, indicating that frequency drift increases by 0.8Hz for every 1°C increase in temperature. Furthermore, we predicted that the drift could reach 13.6Hz at 60°C. Finally, the bit error rate and frequency drift exceeding standard data were associated with the stability score value to construct a comprehensive performance database. The decision tree algorithm was used for analysis and it was found that when the score value was lower than 60, the probability of bit error rate exceeding standard was 75%, and the probability of frequency drift exceeding standard was 60%. Thus, a logical mapping relationship was formed between environmental variables, score values ​​and performance indicators, providing data support for subsequent optimization.

[0262] S9. Input the exceeding-standard data into a linear regression model to predict the quantitative risk assessment index of the equipment in a complex environment.

[0263] In this embodiment, S9, inputting the exceeding-standard data into a linear regression model to predict the risk assessment quantitative index of the equipment in a complex environment, includes:

[0264] S91. Acquire excess data from the monitoring system and clean and format the data using a pre-processing method to obtain a preliminarily processed data set.

[0265] S92. Based on the preliminarily processed data set, a linear regression model is used for training, parameters are adjusted according to data characteristics, and the predictive ability of the model is determined;

[0266] S93. If the deviation between the predicted value of the trained model and the actual risk value exceeds a preset threshold, feature selection is performed on the data set, and input features are readjusted to obtain an optimized data subset;

[0267] S94. Re-run the linear regression model using the optimized data subset to simulate the variables in the complex environment and obtain preliminary quantitative indicators of equipment risk;

[0268] S95. Perform weighted calculation based on the preliminary quantitative indicators in combination with environmental impact factors. If the weighted indicators exceed the safety range, perform a secondary calibration on the indicators to determine the final risk assessment value.

[0269] S96. After obtaining the final risk assessment value, compare and analyze the value with historical data to determine the risk change trend of the device in a complex environment;

[0270] S97. Match the risk change trend with the preset warning rules. If the matching result shows that the risk is increasing, a corresponding warning signal is generated to obtain a dynamic monitoring result of the equipment operation status.

[0271] For example, the specific implementation method for inputting out-of-standard data into a linear regression model to predict quantitative risk assessment indicators for equipment in complex environments can be integrated into a complete logical process. First, assume we have collected a set of equipment operating data, including parameters such as temperature, humidity, and vibration frequency. For example, the temperature exceeded the standard by 45.2°C (the upper limit is 40°C), the humidity by 85.3% (the upper limit is 80%), and the vibration frequency by 12.7Hz (the upper limit is 10Hz). This out-of-standard data is organized into an input feature matrix X, where each row represents a sample and each column represents a parameter value. A target variable Y is constructed, representing the historical risk assessment value. For example, the risk value of a particular record is 0.75 (out of a maximum score of 1.0). Next, a linear regression model is trained, and the regression coefficient is calculated using the least squares method, using the formula β = (X^TX)^(-1) X^TY. Matrix operations are used to determine the weights of each parameter, such as 0.35 for temperature, 0.25 for humidity, and 0.40 for vibration frequency. The analysis process shows that vibration frequency has the greatest impact on risk and requires priority attention. Subsequently, the newly collected data exceeding the standard is input into the model. For example, the new data is a temperature of 46.1°C, a humidity of 86.2%, and a vibration frequency of 13.2Hz. The risk value is predicted by the model and calculated to be 0.82, indicating that the equipment is at high risk in the current complex environment. Further combined with business associations, the prediction results are compared with the equipment maintenance cycle database. If the risk value exceeds 0.8, a maintenance reminder signal is automatically triggered and pushed to the equipment management system to ensure timely intervention. Finally, through error analysis of the prediction results, the mean square error (MSE) is calculated to be 0.05, verifying the accuracy of the model. If the error exceeds the threshold of 0.1, the model retraining process is automatically triggered to update the data set to improve the prediction accuracy.

[0272] Example 2

[0273] See also Figure 2 A narrowband communication device detection device, based on the method described in one of the embodiments, comprising:

[0274] An acquisition module is used to acquire a pre-established wireless channel model data set, wherein the data set includes attenuation values ​​and interference intensity distribution data under different environments;

[0275] A simulation module, configured to construct a multipath propagation simulation model based on the attenuation value and the interference intensity distribution data, obtain a delay spread range value by using time domain analysis, and obtain an amplitude distortion range value by combining frequency domain analysis;

[0276] An evaluation module, configured to input the measured environmental noise data, the delay spread range value, and the amplitude distortion range value into an adaptive filtering algorithm to generate signal quality evaluation data after filtering;

[0277] A first determination module is configured to adjust a transmit power parameter and a frequency offset parameter to obtain an optimized bit error rate value if the signal quality evaluation data does not reach a preset signal-to-noise ratio threshold;

[0278] A construction module is used to construct an extreme environment test model including temperature, humidity and vibration parameters, and collect drift data of the optimized bit error rate value under extreme conditions;

[0279] A fusion module is used to perform feature fusion on the drift data and the wireless channel characteristic data to generate a multi-scenario comprehensive performance evaluation data set;

[0280] A second determination module is configured to adjust the channel model weight parameters through an iterative algorithm to generate an updated stability score value if the fluctuation range value in the performance evaluation data exceeds a preset threshold;

[0281] An extraction module, configured to generate a multi-dimensional performance analysis graph based on the stability score value, and extract bit error rate exceeding standard data and frequency drift exceeding standard data under a specific environment;

[0282] The prediction module is used to input the exceeding-standard data into a linear regression model to predict the risk assessment quantitative index of the equipment in a complex environment.

[0283] In the above technical solution, in order to better utilize the method described in one of the embodiments, the present application proposes a narrowband communication device detection device, each module of which corresponds to each step of the above method. The specific principles have been described above and will not be repeated here.

[0284] The above descriptions are only some embodiments of the present invention and do not limit the scope of protection of the present invention. Any equivalent device or equivalent process transformation made by using the contents of the description and drawings of the present invention, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A narrowband communication device detection method, characterized in that: The method comprises: Acquire a pre-established wireless channel model data set, wherein the data set includes attenuation values ​​and interference intensity distribution data under different environments; A multipath propagation simulation model is constructed based on the attenuation value and interference intensity distribution data, a time domain analysis is used to obtain a delay spread range value, and a frequency domain analysis is combined to obtain an amplitude distortion range value; Inputting the measured environmental noise data, the delay spread range value, and the amplitude distortion range value into an adaptive filtering algorithm to generate signal quality assessment data after filtering; If the signal quality evaluation data does not reach the preset signal-to-noise ratio threshold, adjusting the transmit power parameter and the frequency offset parameter to obtain an optimized bit error rate value; Constructing an extreme environment test model including temperature, humidity, and vibration parameters, and collecting drift data of the optimized bit error rate value under extreme conditions; Performing feature fusion on the drift data and wireless channel characteristic data to generate a multi-scenario comprehensive performance evaluation data set; If the fluctuation range value in the performance evaluation data exceeds a preset threshold, the channel model weight parameters are adjusted through an iterative algorithm to generate an updated stability score value; Generate a multi-dimensional performance analysis graph based on the stability score value, and extract the bit error rate exceeding standard data and the frequency drift exceeding standard data under the preset environment; Inputting the exceeding standard data into a linear regression model to predict the quantitative risk assessment index of the equipment in a complex environment; A multipath propagation simulation model is constructed based on the attenuation value and interference intensity distribution data, and a delay spread range value is obtained by time domain analysis, and an amplitude distortion range value is obtained by frequency domain analysis, including: By obtaining initial input from attenuation value data and interference intensity value, the data processing layer is used to preprocess the distributed data source to obtain a standardized basic data set; Based on the standardized basic data set, a multipath propagation model is constructed, initialized using preset parameters in the simulation model library, and the preliminary structure of the model is determined; Based on the multipath propagation model of the preliminary structure, the time domain analysis method is used to calculate the delay spread value and obtain the distribution characteristics of the delay spread range; The amplitude distortion value is evaluated by combining the distribution characteristics of the delay spread range with the frequency domain analysis method to obtain the range boundary of the amplitude distortion; If the range boundary of the amplitude distortion exceeds the preset threshold, the basic data set is recalibrated through the data processing layer to determine whether the calibrated data meets the constraints of the model construction method; Based on the calibrated data, the parameters of the multipath propagation model are readjusted, and the support vector machine algorithm is used to optimize the model to obtain the final simulation results; Through the final simulation results, the range outputs of delay spread value and amplitude distortion value are integrated to determine the signal characteristic distribution in the multipath propagation environment.

2. A narrowband communication device detection method according to claim 1, characterized in that: Inputting the measured environmental noise data, the delay spread range value, and the amplitude distortion range value into an adaptive filtering algorithm to generate filtered signal quality assessment data, including: The initial input parameter set is formed by combining the collected environmental noise measured data with the delay spread range value and the amplitude distortion range value; According to the initial input parameter group, the environmental noise measured data is preliminarily processed using an adaptive filtering algorithm to obtain a filtered first signal data set; For the first signal data set, obtaining the delay spread characteristics and amplitude distortion characteristics therein, determining whether the characteristic distribution meets the preset threshold range, and if the characteristic distribution exceeds the threshold range, performing secondary filtering processing on the first signal data set to obtain a second signal data set; Extracting signal quality related indicators from the second signal data set, determining whether the signal quality indicators reach a preset standard value, and if not, adjusting parameter configurations of the adaptive filtering algorithm to generate a third signal data set; Calculating key parameter values ​​in the signal quality assessment data based on the third signal data set to obtain preliminary structured results of the assessment data; Use the structured results, combined with the pre-established signal quality assessment model, to generate the final signal quality assessment data and determine the completeness and accuracy of the assessment data; Through the final signal quality assessment data, the distribution characteristics of the outliers are extracted to determine whether the distribution of the outliers affects the overall signal quality. If the impact exceeds the preset range, the outliers are marked and recorded in the assessment data.

3. The narrowband communication device detection method according to claim 1, wherein: If the signal quality evaluation data does not reach a preset signal-to-noise ratio threshold, adjusting a transmit power parameter and a frequency offset parameter to obtain an optimized bit error rate value, including: By processing the initial evaluation data of the signal quality, specific data of the signal-to-noise ratio value is obtained to determine whether it reaches the preset threshold; If the signal-to-noise ratio value does not reach the preset threshold, preliminary parameter adjustments are performed on the transmit power and frequency offset to obtain a first set of adjusted parameter data; According to the first set of parameter data, a pre-established simulation model is used to simulate signal transmission and obtain simulated signal quality data; If the signal quality data after simulation still does not meet the preset threshold, the transmit power and frequency offset are optimized and adjusted twice using the support vector machine algorithm to obtain a second set of parameter data; Based on the second set of parameter data, the signal transmission environment is monitored in real time to obtain an optimized bit error rate value; By analyzing the bit error rate value, we can determine whether the stability of signal transmission meets the preset standards and obtain the final evaluation results; If the final evaluation results show insufficient stability, the second set of parameter data is fine-tuned to obtain a third set of parameter data, and the simulation and monitoring processes are repeated to determine the optimal configuration for signal transmission.

4. The narrowband communication device detection method according to claim 1, wherein: Construct an extreme environment test model including temperature, humidity, and vibration parameters, and collect drift data of the optimized bit error rate value under extreme conditions, including: By building an extreme environment test model, integrating temperature parameters, humidity parameters, and vibration parameters, and simulating test scenarios under various environmental conditions, preliminary environmental parameter combination data is obtained; Use acquisition tools to extract key features from preliminary environmental parameter combination data, record the variation range of temperature parameters, humidity parameters, and vibration parameters, and determine the feature data set; Based on the feature data set, the support vector machine algorithm is used to model and analyze the correlation between temperature parameters, humidity parameters, vibration parameters and bit error rate values ​​to obtain a bit error rate prediction model; If the output value of the bit error rate prediction model exceeds the preset threshold, the collected environmental condition data is recalibrated, and the deviation is corrected in combination with the drift data to determine the bit error rate value after calibration; By comparing and analyzing the calibrated bit error rate value with the optimized data, the distribution characteristics of abnormal points in the drift data are extracted to obtain the abnormal drift feature set; Based on the abnormal drift feature set, dynamic adjustments are made to the changing trends of environmental conditions, generating corresponding parameter optimization strategies to determine the final bit error rate drift control range; Obtain the final bit error rate drift control range, combine the environmental conditions and parameter acquisition data records, iteratively update the test model, and obtain an optimized extreme environment test framework.

5. The narrowband communication device detection method according to claim 1, wherein: The drift data is subjected to feature fusion with the wireless channel characteristic data to generate a multi-scenario comprehensive performance evaluation data set, including: By acquiring drift data and channel characteristic data from original sources, a pre-processing method is used to clean and unify the formats of the two types of data to obtain a first data set after preliminary processing; Based on the first data set, performing feature extraction operations on the drift data and the channel characteristic data respectively to obtain respective key feature values ​​to form a second feature set; If the distribution of the eigenvalues ​​in the second feature set meets the preset threshold range, the drift data and the key features of the channel characteristics are integrated by a feature fusion method to determine a third fused feature group; If the feature dimensions in the third fused feature group meet the adaptation requirements under multiple scenarios, the fused features are classified using the scenario adaptation rules to obtain a fourth scenario feature set; Based on the fourth scenario feature set, a support vector machine algorithm is applied to comprehensively analyze the feature data under multiple scenarios to determine the preliminary results of the performance evaluation and form a fifth evaluation data set; Using the fifth evaluation data set and the preset evaluation indicators, the performance in multiple scenarios is quantified to obtain a sixth quantification result set; If the indicator values ​​in the sixth quantitative result set meet the preset standards, the data set is finally integrated to determine the complete data results of the multi-scenario comprehensive performance evaluation.

6. The narrowband communication device detection method according to claim 1, characterized in that: If the fluctuation range value in the performance evaluation data exceeds a preset threshold, the channel model weight parameters are adjusted through an iterative algorithm to generate an updated stability score value, including: By detecting the fluctuation range of the performance evaluation data, it is determined whether it exceeds the preset threshold and a preliminary range detection result is obtained; According to the range detection result, if the fluctuation range exceeds the preset threshold, the current weight parameters of the channel model are obtained to determine the parameter set that needs to be adjusted; Adjust the weight parameters in the parameter set one by one through an iterative method to generate an adjusted channel model parameter group; Using the adjusted channel model parameter group, recalculate the stability score of the performance evaluation data to obtain an updated score value; For the updated score value, if the score value still does not meet the preset standard, the performance evaluation data is further cleaned through the data processing module to obtain an optimized data set; Based on the optimized data set, the channel model is re-entered to adjust parameters and update scores to determine whether stability requirements are met and generate the final model optimization results; Based on the final model optimization results, the operating configuration of the channel model is updated and the stability score benchmark of the system in subsequent operations is determined.

7. A narrowband communication device detection method according to claim 1, characterized in that: A multi-dimensional performance analysis graph is generated based on the stability score value, and bit error rate exceeding standard data and frequency drift exceeding standard data under a preset environment are extracted, including: By collecting stability score-related data and using a preset threshold to preliminarily screen the score data, a preliminarily screened score data set is obtained; Based on the initially screened scoring dataset, and in response to the requirements for building a multidimensional performance graph, we apply performance analysis methods to generate corresponding multidimensional performance graph data and identify the key performance dimensions in the graph. Based on the multi-dimensional performance map data and the influence of preset environmental values, we analyze the effect of environmental impact on performance dimensions and obtain performance fluctuation data under environmental influences. For performance fluctuation data, if the fluctuation data exceeds the preset threshold range, the bit error rate data and frequency drift value are separated through data extraction method to obtain the separated abnormal data group; Based on the separated abnormal data group, the support vector machine algorithm is used to classify the data points with bit error rate data and frequency drift values ​​exceeding the standard, and the distribution characteristics of the data points exceeding the standard are determined; By further processing the distribution characteristics of the data points that exceed the standard, combining the score correlation and drift detection methods, the correlation pattern between the frequency drift value and the stability score is analyzed to determine the final abnormal correlation result; According to the final anomaly correlation result, corresponding multi-dimensional performance map update data is generated for the key points in the anomaly correlation result to obtain the updated performance analysis map.

8. The narrowband communication device detection method according to claim 1, wherein: Inputting the exceeding data into a linear regression model to predict the quantitative risk assessment indicators of the equipment in a complex environment includes: By acquiring the exceeding standard data from the monitoring system, the data is cleaned and formatted by using pre-processing means to obtain a preliminarily processed data set; Based on the data set after preliminary processing, a linear regression model is used for training, and parameters are adjusted according to the data characteristics to determine the model's predictive ability; If the deviation between the predicted value of the trained model and the actual risk value exceeds a preset threshold, feature selection is performed on the data set, and the input features are readjusted to obtain an optimized data subset; Using the optimized data subset, the linear regression model is re-run to simulate variables in a complex environment and obtain preliminary quantitative indicators of equipment risk. Based on the preliminary quantitative indicators, a weighted calculation is performed in combination with environmental impact factors. If the weighted indicators exceed the safety range, the indicators are recalibrated to determine the final risk assessment value; After obtaining the final risk assessment value, compare and analyze the value with historical data to determine the risk change trend of the equipment in a complex environment; By matching the risk change trend with the preset early warning rules, if the matching result shows that the risk is increasing, a corresponding early warning signal is generated to obtain the dynamic monitoring results of the equipment operation status.

9. A narrowband communication device detection device, characterized in that: Based on the method according to any one of claims 1 to 8, the device comprises: An acquisition module is used to acquire a pre-established wireless channel model data set, wherein the data set includes attenuation values ​​and interference intensity distribution data under different environments; A simulation module is used to construct a multipath propagation simulation model based on the attenuation value and interference intensity distribution data, obtain a delay spread range value by using time domain analysis, and obtain an amplitude distortion range value by combining frequency domain analysis, including: By obtaining initial input from attenuation value data and interference intensity value, the data processing layer is used to preprocess the distributed data source to obtain a standardized basic data set; Based on the standardized basic data set, a multipath propagation model is constructed, initialized using preset parameters in the simulation model library, and the preliminary structure of the model is determined; Based on the multipath propagation model of the preliminary structure, the time domain analysis method is used to calculate the delay spread value and obtain the distribution characteristics of the delay spread range; The amplitude distortion value is evaluated by combining the distribution characteristics of the delay spread range with the frequency domain analysis method to obtain the range boundary of the amplitude distortion; If the range boundary of the amplitude distortion exceeds the preset threshold, the basic data set is recalibrated through the data processing layer to determine whether the calibrated data meets the constraints of the model construction method; Based on the calibrated data, the parameters of the multipath propagation model are readjusted, and the support vector machine algorithm is used to optimize the model to obtain the final simulation results; Through the final simulation results, the range output of delay spread value and amplitude distortion value is integrated to determine the signal characteristic distribution in the multipath propagation environment; An evaluation module, configured to input the measured environmental noise data, the delay spread range value, and the amplitude distortion range value into an adaptive filtering algorithm to generate signal quality evaluation data after filtering; A first determination module is configured to adjust a transmit power parameter and a frequency offset parameter to obtain an optimized bit error rate value if the signal quality evaluation data does not reach a preset signal-to-noise ratio threshold; A construction module is used to construct an extreme environment test model including temperature, humidity and vibration parameters, and collect drift data of the optimized bit error rate value under extreme conditions; A fusion module is used to perform feature fusion on the drift data and the wireless channel characteristic data to generate a multi-scenario comprehensive performance evaluation data set; A second determination module is configured to adjust the channel model weight parameters through an iterative algorithm to generate an updated stability score value if the fluctuation range value in the performance evaluation data exceeds a preset threshold; An extraction module is used to generate a multi-dimensional performance analysis map based on the stability score value, and extract the bit error rate exceeding standard data and the frequency drift exceeding standard data under a preset environment; The prediction module is used to input the exceeding-standard data into a linear regression model to predict the risk assessment quantitative index of the equipment in a complex environment.

Images