Intelligent monitoring method and system for performance of radio frequency relay

Abstract

This application provides an intelligent monitoring method and system for the performance of radio frequency (RF) relays, applied in the field of RF relay technology. It collects the service signals of RF relays and establishes characteristic fingerprint baselines under different service loads in a healthy state. During normal operation, it extracts characteristic fingerprints in real time and identifies the current service load. By comparing the real-time characteristic fingerprints with the baseline fingerprints, it determines deviations and dynamically evaluates the deviations of adjacent RF relays, establishing a dynamic reference deviation benchmark. Finally, it compares the deviations of the RF relays with the dynamic reference deviation benchmark to quantify the degradation contribution of the RF relays themselves, thereby effectively distinguishing the degradation of the RF relays themselves from the impact of other parts of the system or environmental changes, and issuing a performance degradation warning. Therefore, this application has the beneficial effects of achieving non-stop detection, accurately distinguishing the degradation of RF relays themselves from the influence of external factors, and significantly reducing false alarms and missed alarms.

Description

Technical Field

[0001] This application relates to the field of radio frequency relay technology, and in particular to a method and system for intelligent monitoring of radio frequency relay performance. Background Technology

[0002] Radio frequency (RF) relays are responsible for signal path switching in communication systems, but their performance degrades over time due to wear and tear on mechanical and electrical components. Traditional maintenance requires system shutdown for inspection, which is costly and impacts critical systems. To achieve uninterrupted online monitoring, existing technologies initially attempted to inject weak probe signals into service signals, but in practical applications, these weak probe signals interfere with normal communication. Therefore, existing technologies have shifted to inferring the relay's health status using the characteristics of the service signals themselves.

[0003] However, this technology still faces three major challenges: First, the power, frequency, and modulation method of service signals are constantly changing, and the performance of relay degradation varies greatly under different signals; second, degradation often exhibits nonlinearity, such as intermodulation distortion that may occur at high power, which is difficult to capture by simply monitoring the average value of insertion loss; third, changes in other components in the system link (amplifiers, filters, etc.) and the environment can also affect signal characteristics, making it difficult to accurately attribute subtle changes to specific relays. These problems make online monitoring prone to false alarms or missed alarms, affecting the reliability of predictive maintenance.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] In view of the shortcomings of the prior art, this application provides a method and system for intelligent monitoring of radio frequency relay performance, which aims to solve the technical problems of high cost, business interruption and false alarms or missed alarms in the existing radio frequency relay maintenance methods and online monitoring systems.

[0006] In a first aspect, a method for intelligent monitoring of the performance of a radio frequency relay, the method comprising the following steps:

[0007] S1: Acquire service signals transmitted via radio frequency relays;

[0008] S2: When the radio frequency relay is in a healthy state, establish a characteristic fingerprint baseline of the service signal for different service loads;

[0009] S3: During normal operation of the radio frequency relay, extract the feature fingerprint of the service signal in real time and identify the current service load;

[0010] S4: Compare the real-time extracted feature fingerprint with the baseline fingerprint in the feature fingerprint baseline that corresponds to the current service load, and determine the deviation of the feature fingerprint;

[0011] S5: Based on the deviation of the feature fingerprint, dynamically evaluate the deviation of the feature fingerprint of adjacent RF relays, and based on the evaluation results, establish a dynamic reference deviation benchmark for the RF relay;

[0012] S6: Compare the characteristic fingerprint deviation of the RF relay with the dynamic reference deviation benchmark to obtain the relative deviation amount, and quantify the degradation contribution of the RF relay itself based on the relative deviation amount and the characteristics of the current service load, so as to distinguish the degradation of the RF relay itself from the impact of other parts of the system or environmental changes.

[0013] S7: Based on the differentiation results, issue a warning about the performance degradation of the radio frequency relay.

[0014] Furthermore, step S2 includes:

[0015] S21: When the radio frequency relay is in a healthy state, establish an initial feature fingerprint baseline for the service signal for different service loads;

[0016] S22: During normal operation of the radio frequency relay, continuously monitor the characteristic fingerprint of the service signal and its deviation from the initial characteristic fingerprint baseline;

[0017] S23: Based on the deviation, identify the period during which the radio frequency relay is in a stable and healthy operating state, and there is no performance degradation warning during the period;

[0018] S24: During the stable and healthy operation cycle, if it is determined that the characteristic fingerprint of the service signal continuously and generally deviates from the initial characteristic fingerprint baseline, but the degree of deviation does not reach the performance degradation warning threshold, then it is determined that the characteristic fingerprint baseline has undergone natural drift.

[0019] S25: Based on the judgment result of the natural drift, and based on the service signal feature fingerprint data collected during the stable and healthy operation cycle, the initial feature fingerprint baseline is incrementally adjusted to obtain the feature fingerprint baseline of the service signal.

[0020] Furthermore, in step S3, the real-time extraction of the feature fingerprint of the service signal includes the following steps:

[0021] S31: Adjust the feature fingerprint extraction parameters of the service signal according to the identified current service load characteristics, the parameters including sampling rate, processing bandwidth or calculation accuracy;

[0022] S32: Assign the extraction tasks of different feature fingerprints of the service signal to independent hardware processing units, and perform parallel synchronous calculations on the service signal according to the adjusted feature fingerprint extraction parameters to obtain the feature fingerprint of the service signal;

[0023] S33: When the data processing load of the service signal exceeds a preset threshold, the higher-order modulation error vector amplitude or signal envelope distortion mode in the service signal is extracted first, and the extraction of other features is downgraded to ensure the real-time extraction of the feature fingerprint of the service signal.

[0024] Furthermore, in step S32, the characteristic fingerprint of the service signal includes: harmonic distortion components, intermodulation distortion components, in-band noise power, phase noise spectral density, gain compression characteristics, or amplitude-phase conversion characteristics of the service signal.

[0025] Furthermore, in step S3, identifying the current service load includes the following steps:

[0026] S34: Analyze the modulation type, bandwidth, power spectral density, or data throughput parameters of the service signal to obtain the service signal parameters;

[0027] S35: Match the service signal parameters with a preset service load feature library to obtain a matching result;

[0028] S36: Identify the current service load based on the matching results.

[0029] Furthermore, step S4 includes:

[0030] S41: Obtain the correlation pattern between different feature fingerprints of the service signals contained in the baseline fingerprint corresponding to the current service load in the feature fingerprint baseline and the temporal evolution pattern of a single feature fingerprint;

[0031] S42: Real-time calculation of the real-time correlation patterns between different feature fingerprints of the service signals contained in the real-time extracted feature fingerprints and the real-time temporal evolution patterns of individual feature fingerprints;

[0032] S43: Compare the real-time correlation pattern with the correlation pattern, and the real-time time-series evolution pattern with the time-series evolution pattern to obtain the pattern comparison result;

[0033] S44: Based on the pattern comparison results, determine the pattern deviation of the feature fingerprint and quantify the pattern deviation.

[0034] Furthermore, in step S7, issuing the warning of performance degradation of the radio frequency relay includes the following steps:

[0035] S71: Send the warning information about the performance degradation of the radio frequency relay to the equipment management system or maintenance work order system;

[0036] S72: Based on the warning information, automatically generate maintenance tasks or fault reports.

[0037] Furthermore, the method also includes:

[0038] S8: Predict the time window when the radio frequency relay reaches a critical fault state.

[0039] Furthermore, step S8 includes:

[0040] S81: Continuously monitor the temporal changes of multiple degradation indicators reflected by the degradation of the radio frequency relay itself, and obtain the current values ​​of the multiple degradation indicators.

[0041] S82: For each degradation index, calculate the rate of change and acceleration of change of the degradation index based on the time-series changes of the degradation index;

[0042] S83: Based on the characteristics of the current service load, dynamically set a critical threshold and degradation acceleration coefficient for each degradation index;

[0043] S84: Identify that the current value of any of the degradation indicators, the rate of change, or the acceleration of change reaches a preset accelerated degradation trigger condition;

[0044] S85: When the accelerated degradation triggering condition is identified, the time required for the degradation index to reach the critical threshold is calculated based on the current value of the degradation index, the rate of change, the acceleration of change, the critical threshold, and the degradation acceleration coefficient.

[0045] S86: Compare all the calculated times and select the time value with the smallest value as the time window for the radio frequency relay to reach the critical fault state.

[0046] Secondly, a radio frequency relay performance intelligent monitoring system, the system being used to implement the steps of any of the above methods, the system comprising:

[0047] Acquisition module: Acquires service signals transmitted via radio frequency relays;

[0048] First baseline establishment module: When the radio frequency relay is in a healthy state, establish the characteristic fingerprint baseline of the service signal for different service loads;

[0049] Fingerprint extraction module: During normal operation of the radio frequency relay, extracts the characteristic fingerprint of the service signal in real time and identifies the current service load;

[0050] Comparison module: compares the real-time extracted feature fingerprint with the baseline fingerprint in the feature fingerprint baseline that corresponds to the current service load, and determines the deviation of the feature fingerprint;

[0051] The second baseline establishment module dynamically evaluates the deviation of the characteristic fingerprints of adjacent RF relays based on the deviation of the characteristic fingerprints, and establishes a dynamic reference deviation benchmark for the RF relays based on the evaluation results.

[0052] Differentiation module: compares the deviation of the characteristic fingerprint of the RF relay with the dynamic reference deviation benchmark to obtain the relative deviation amount, and quantifies the degradation contribution of the RF relay itself based on the relative deviation amount and the characteristics of the current service load, so as to distinguish the degradation of the RF relay itself from the impact of other parts of the system or environmental changes.

[0053] Early warning module: Based on the differentiation results, issue an early warning of performance degradation of the radio frequency relay.

[0054] Beneficial Effects: This application proposes an intelligent monitoring method and system for RF relay performance. It collects the service signals of RF relays and establishes characteristic fingerprint baselines under different service loads in a healthy state. During normal operation, it extracts characteristic fingerprints in real time and identifies the current service load. By comparing the real-time characteristic fingerprints with the baseline fingerprints, it determines deviations and dynamically evaluates the deviations of adjacent RF relays, establishing a dynamic reference deviation benchmark. Finally, it compares the deviations of the RF relays with the dynamic reference deviation benchmark to quantify the degradation contribution of the RF relays themselves. This effectively distinguishes between the degradation of the RF relays themselves and the impact of changes in other parts of the system or the environment, and issues a performance degradation warning. Therefore, this application has the beneficial effects of achieving non-stop detection, accurately distinguishing between the degradation of the RF relays themselves and the influence of external factors, and significantly reducing false alarms and missed alarms. Attached Figure Description

[0055] Figure 1 This is a flowchart of a method for intelligent monitoring of the performance of a radio frequency relay proposed in this application.

[0056] Figure 2 This is a structural diagram of an intelligent monitoring system for the performance of a radio frequency relay proposed in this application.

[0057] Figure 3 This is a schematic diagram of an intelligent monitoring system for the performance of a radio frequency relay proposed in this application.

[0058] Labeling Explanation: 201. Data Acquisition Module; 202. First Baseline Establishment Module; 203. Fingerprint Extraction Module; 204. Comparison Module; 205. Second Baseline Establishment Module; 206. Differentiation Module; 207. Early Warning Module. Detailed Implementation

[0059] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and marked in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0060] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0061] Firstly, a method for intelligent monitoring of the performance of a radio frequency relay, the method comprising the following steps:

[0062] S1: Acquire service signals transmitted via radio frequency relays;

[0063] S2: When the RF relay is in a healthy state, establish a characteristic fingerprint baseline for the service signal for different service loads;

[0064] S3: During normal operation of the RF relay, extract the characteristic fingerprint of the service signal in real time and identify the current service load;

[0065] S4: Compare the real-time extracted feature fingerprint with the baseline fingerprint in the feature fingerprint baseline that corresponds to the current business load to determine the deviation of the feature fingerprint;

[0066] S5: Based on the deviation of the feature fingerprint, dynamically evaluate the deviation of the feature fingerprint of adjacent RF relays, and based on the evaluation results, establish a dynamic reference deviation benchmark for the RF relays;

[0067] S6: Compare the characteristic fingerprint deviation of the RF relay with the dynamic reference deviation benchmark to obtain the relative deviation amount. Based on the relative deviation amount and the characteristics of the current service load, quantify the degradation contribution of the RF relay itself to distinguish the degradation of the RF relay itself from the impact of other parts of the system or environmental changes.

[0068] S7: Based on the differentiation results, issue a warning about the performance degradation of the radio frequency relay.

[0069] This application effectively solves the accuracy and reliability problems of RF relay performance monitoring in existing technologies through a series of steps, including collecting service signals, establishing a characteristic fingerprint baseline, extracting and comparing characteristic fingerprints in real time, dynamically evaluating deviations, quantifying degradation contributions, and issuing early warnings. This method can distinguish between the degradation of the RF relay itself and the impact of changes in other parts of the system or the environment, thereby avoiding false alarms and missed alarms and improving the accuracy of predictive maintenance.

[0070] In step S1, the service signal can be acquired by connecting a high-precision oscilloscope, spectrum analyzer, or dedicated RF signal acquisition equipment to the input or output of the RF relay. For example, a high-speed analog-to-digital converter (ADC) can be used to digitize the analog RF signal and store it as a digital signal data stream. Alternatively, existing monitoring ports or test points within the communication system can be used to extract a portion of the service signal via a coupler or directional coupler and transmit it to the signal processing unit for analysis.

[0071] In step S2, the characteristic fingerprint baseline serves as a reference for judging the performance deviation of the RF relay. When establishing the baseline, the RF relay can first be ensured to be in optimal health after being put into use or calibrated. Then, different service loads are simulated or actually run; for example, different signal power levels, modulation types (such as QPSK, 16QAM, 64QAM, etc.), bandwidths, or data rates can be set. Under each service load, the service signal is collected and its characteristic fingerprint is extracted. For example, harmonic distortion components, intermodulation distortion components, in-band noise power, phase noise spectral density, gain compression characteristics, or amplitude-phase transition characteristics can be extracted. These characteristic fingerprints and their corresponding service load information are stored to form a characteristic fingerprint baseline database.

[0072] Specifically, step S2 includes:

[0073] S21: When the RF relay is in a healthy state, establish an initial characteristic fingerprint baseline for the service signal for different service loads;

[0074] S22: During normal operation of the radio frequency relay, continuously monitor the characteristic fingerprint of the service signal and its deviation from the initial characteristic fingerprint baseline;

[0075] S23: Based on the deviation, identify the period during which the RF relay is in a stable and healthy operating state, with no performance degradation warning within the period;

[0076] S24: During a stable and healthy operating cycle, if the characteristic fingerprint of the business signal is continuously and generally deviates from the initial characteristic fingerprint baseline, but the degree of deviation does not reach the performance degradation warning threshold, then the characteristic fingerprint baseline is judged to have undergone natural drift.

[0077] S25: Based on the judgment result of natural drift, the initial feature fingerprint baseline is incrementally adjusted based on the service signal feature fingerprint data collected during the stable and healthy operation cycle, so as to obtain the feature fingerprint baseline of the service signal.

[0078] Specifically, in step S21, the initial feature fingerprint baseline refers to the reference set of service signal feature fingerprints collected and established for different service load conditions (such as different modulation methods, bandwidths, power levels, etc.) when the RF relay is first put into use or after a comprehensive overhaul and confirmed to be in optimal health. This initial baseline serves as the starting point for subsequent monitoring.

[0079] In step S22, the deviation can be quantified as numerical differences, pattern differences, or statistical distances in the characteristic fingerprints.

[0080] In step S23, the period of stable and healthy operation refers to a time period in which the RF relay does not issue any performance degradation warnings, and the deviation of its service signal characteristic fingerprint remains within the normal fluctuation range. Identifying this period helps to eliminate anomalies caused by transient interference or occasional events, ensuring the accuracy of subsequent baseline adjustments.

[0081] In step S24, the determination of natural drift is crucial to this solution. When, during a stable and healthy operating cycle, the characteristic fingerprint of the service signal exhibits a continuous and widespread deviation, but the degree of this deviation has not yet reached the preset performance degradation warning threshold, it can be determined that a natural drift has occurred in the characteristic fingerprint baseline. This drift is usually slow and cumulative, rather than a sudden failure.

[0082] In step S25, incremental adjustment refers to making small, gradual updates to the initial feature fingerprint baseline based on the identified natural drift and using the latest service signal feature fingerprint data collected during a stable and healthy operating cycle. This adjustment can be a weighted average, exponential smoothing, or other adaptive algorithm to ensure that the baseline better reflects the true state of the RF relay under long-term healthy operation.

[0083] In some preferred embodiments, it is assumed that an RF relay establishes an initial characteristic fingerprint baseline for its service signals during its initial deployment. Over the following months, the monitoring system continues to operate without issuing any performance degradation warnings. However, the system observes that the in-band noise power of the RF relay under specific service loads exhibits a slow and continuous upward trend, but its value remains below the preset performance degradation warning threshold. At this point, the system identifies this as a natural drift phenomenon within a stable and healthy operating cycle. Based on the service signal data collected during this cycle, the system incrementally adjusts the corresponding in-band noise power component in the initial characteristic fingerprint baseline to reflect the actual noise level of the RF relay under healthy conditions. Thus, the new characteristic fingerprint baseline will better reflect the long-term operating characteristics of the RF relay, avoiding potential misjudgments due to baseline rigidity.

[0084] In step S3, real-time extraction of feature fingerprints can be achieved through dedicated signal processing hardware or software. For example, a digital signal processor (DSP) or field-programmable gate array (FPGA) can be used to perform spectral analysis such as Fast Fourier Transform (FFT) on the real-time acquired service signal to extract distortion components such as harmonics and intermodulation. Simultaneously, to accurately compare with a baseline, the current service load needs to be identified. This can be achieved by analyzing parameters such as the modulation type, bandwidth, power spectral density, or data throughput of the service signal.

[0085] Specifically, in step S3, the real-time extraction of the feature fingerprint of the service signal includes the following steps:

[0086] S31: Adjust the feature fingerprint extraction parameters of the service signal according to the identified current service load characteristics. The parameters include sampling rate, processing bandwidth or calculation accuracy.

[0087] S32: The task of extracting different feature fingerprints of the service signal is assigned to an independent hardware processing unit, and the service signal is calculated in parallel and synchronously according to the adjusted feature fingerprint extraction parameters to obtain the feature fingerprint of the service signal.

[0088] S33: When the data processing load of the service signal exceeds the preset threshold, the higher-order modulation error vector amplitude or signal envelope distortion mode in the service signal is extracted first, and the extraction of other features is downgraded to ensure the real-time extraction of the feature fingerprint of the service signal.

[0089] In step S31, system parameters for feature extraction can be dynamically configured based on the specific characteristics of the current service signal, such as its modulation scheme, bandwidth, and data rate. For example, when the service load is a high-bandwidth, high-order modulation signal, the sampling rate and calculation accuracy can be increased to capture more refined signal features; conversely, when the service load is a low-bandwidth, low-order modulation signal, the sampling rate and calculation accuracy can be appropriately reduced to save computational resources. The purpose of this is to enable the feature extraction process to adaptively match different service scenarios, thereby improving extraction efficiency and accuracy.

[0090] In step S32, dedicated hardware resources such as multi-core processors, digital signal processors (DSPs), or field-programmable gate arrays (FPGAs) can be used to simultaneously process different types of feature fingerprints, including harmonic distortion components, intermodulation distortion components, in-band noise power, phase noise spectral density, gain compression characteristics, and amplitude-phase conversion characteristics of the service signal. This parallel processing method can significantly shorten the total time for feature extraction, ensuring that feature fingerprints can be calculated in real time and efficiently even with continuous input of service signals.

[0091] In step S33, when the system's processing capacity faces a bottleneck, the system intelligently prioritizes processing the key features most sensitive to or indicative of RF relay performance degradation. Higher-order modulation error vector amplitude (EVM) and signal envelope distortion mode typically directly reflect nonlinear distortion in the RF link and power amplifier performance degradation, serving as important indicators for assessing the health of the RF relay. Degrading the extraction of other secondary features, such as reducing their sampling rate or calculation frequency, aims to ensure real-time acquisition of core degradation indicators even with limited resources, avoiding overall delays caused by comprehensive processing, thereby ensuring timely performance degradation warnings.

[0092] Furthermore, in step S32, the characteristic fingerprint of the service signal includes: harmonic distortion components, intermodulation distortion components, in-band noise power, phase noise spectral density, gain compression characteristics, or amplitude phase conversion characteristics of the service signal.

[0093] Among them, harmonic distortion components refer to the signal components with frequencies that are integer multiples of the fundamental frequency generated when the service signal passes through the radio frequency relay due to nonlinear effects, which reflect the linearity of the radio frequency relay.

[0094] Intermodulation distortion (IMD) refers to the signal components of non-integer multiple frequencies generated due to nonlinear effects when multiple frequency service signals pass through an RF relay simultaneously. It is also an important indicator for measuring the linear performance of an RF relay.

[0095] In-band noise power refers to the random noise power present within the effective bandwidth of the service signal, which directly affects the signal-to-noise ratio of the signal.

[0096] Phase noise spectral density describes the randomness of signal phase variation with frequency and has a significant impact on the demodulation performance of high-order modulated signals.

[0097] Gain compression characteristic refers to the phenomenon that the output power gain of an RF relay decreases as the input power increases, reflecting its maximum linear output capability.

[0098] Amplitude-phase conversion refers to the phenomenon that changes in the amplitude of the input signal lead to changes in the phase of the output signal. For some modulation methods, this can introduce additional signal distortion.

[0099] Through the above technical solutions, this application can obtain a more refined and comprehensive performance characterization of RF relays. Compared with methods that only monitor a single or a few macroscopic indicators, this application significantly improves the ability to identify potential degradation modes of RF relays by extracting multiple key features such as harmonic distortion components, intermodulation distortion components, in-band noise power, phase noise spectral density, gain compression characteristics, and amplitude-phase conversion characteristics.

[0100] Furthermore, in step S3, identifying the current service load includes the following steps:

[0101] S34: Analyze the modulation type, bandwidth, power spectral density, or data throughput parameters of the service signal to obtain the service signal parameters;

[0102] S35: Match the service signal parameters with the preset service load feature library to obtain the matching result;

[0103] S36: Identify the current service load based on the matching results.

[0104] Specifically, in step S34, the analysis of the service signal aims to obtain its core characteristics, which directly reflect the current operating status of the communication network and data transmission requirements. Among these, modulation type can refer to different digital modulation methods such as QPSK, 16QAM, and 64QAM; its analysis helps to understand the signal's coding efficiency and anti-interference capability. Bandwidth refers to the frequency range occupied by the signal, reflecting the data transmission capacity. Power spectral density describes the distribution of signal power across frequencies and can be used to assess the signal's energy concentration and potential interference. The data throughput parameter directly quantifies the amount of data transmitted per unit time. Through comprehensive analysis of these parameters, service signal parameters can be fully obtained, providing basic data for subsequent service load identification.

[0105] In step S35, the service load feature library pre-stores combination patterns of typical modulation types, bandwidths, power spectral density, and data throughput parameters of service signals under different service scenarios (e.g., voice calls, video streaming, data downloads, etc.). The matching process can employ various algorithms, such as pattern recognition, to determine which preset service load pattern in the feature library most closely matches the current service signal parameters, thereby obtaining the matching result.

[0106] In one specific embodiment, the matching process can be performed by a data processing unit. This unit receives current service signal parameters and accesses a storage unit containing a feature library. The feature library contains multiple preset service load patterns. When using a pattern recognition algorithm, the data processing unit can perform feature extraction to extract specific feature vectors from the current service signal parameters. Subsequently, the data processing unit compares this feature vector with feature vectors of each preset service load pattern in the feature library, for example, by calculating cosine similarity to quantify their proximity. The service load pattern corresponding to the feature vector with the highest cosine similarity is selected as the matching result.

[0107] Therefore, in step S36, the current service load can be identified based on the matching result obtained in step S35. For example, if the matching result shows that the current service signal parameters highly match the characteristic mode of high-definition video streaming, then the current service load can be identified as high-definition video streaming. This identification mechanism ensures accurate perception of the actual operating environment of the RF relay.

[0108] The above technical solution enables accurate identification of the current service load of RF relays. This accurate identification is crucial for intelligent monitoring of RF relay performance because it ensures that subsequent baseline comparisons of feature fingerprints are performed under the correct service load context, avoiding misjudgments caused by changes in service load. This improves the accuracy and reliability of RF relay performance degradation early warning, allowing maintenance personnel to intervene more promptly and accurately, thereby effectively extending equipment lifespan and ensuring the stable operation of the communication network.

[0109] In step S4, after identifying the current service load, the baseline fingerprint corresponding to the service load is retrieved from the feature fingerprint baseline database. Then, the real-time extracted feature fingerprint is compared with the baseline fingerprint one by one. The comparison method may include calculating the absolute difference, relative percentage deviation, or statistical distance (such as Euclidean distance) of each feature parameter. For example, if the real-time extracted harmonic distortion component is higher than the baseline value by a certain threshold, it can be determined that there is a deviation.

[0110] Specifically, step S4 includes:

[0111] S41: Obtain the correlation patterns between different feature fingerprints of the service signals contained in the baseline fingerprint corresponding to the current service load in the feature fingerprint baseline and the temporal evolution patterns of individual feature fingerprints;

[0112] S42: Real-time calculation of the real-time correlation patterns between different feature fingerprints of the business signals contained in the real-time extracted feature fingerprints and the real-time temporal evolution patterns of individual feature fingerprints;

[0113] S43: Compare real-time correlation patterns with each other, and real-time time series evolution patterns with each other, to obtain the pattern comparison results;

[0114] S44: Based on the pattern comparison results, determine the pattern deviation of the feature fingerprint and quantify the pattern deviation.

[0115] In step S41, the correlation pattern can be understood as the statistical dependence or mathematical relationship between different feature fingerprints (e.g., harmonic distortion components, intermodulation distortion components, in-band noise power, phase noise spectral density, gain compression characteristics, or amplitude phase transition characteristics) in the service signal when the RF relay is in a healthy state. For example, in some specific embodiments, a correlation coefficient calculation unit can be set up, which receives all extracted feature fingerprint data. The correlation coefficient calculation unit performs Pearson correlation coefficient calculation on each pair of feature fingerprints. For example, it calculates the correlation coefficient between harmonic distortion components and intermodulation distortion components, the correlation coefficient between in-band noise power and phase noise spectral density, and the correlation coefficient between gain compression characteristics and amplitude phase transition characteristics. The calculation results form a correlation coefficient matrix, which characterizes the statistical dependence between each feature fingerprint in the healthy state of the RF relay. This matrix serves as the correlation pattern benchmark for the healthy state and is stored in the data storage unit.

[0116] Temporal evolution patterns refer to the patterns of change of a single feature fingerprint over time in a healthy state, such as its mean, variance, trend, periodicity, or autoregressive characteristics. These patterns can be modeled using time series analysis methods (such as ARIMA models). These patterns are established when the RF relay is in a healthy state and stored as baseline patterns. In some specific embodiments, the process of establishing and storing the temporal evolution patterns of the RF relay in a healthy state is as follows:

[0117] First, a spectrum analyzer is connected to the input and output terminals of the RF relay to acquire its RF signal characteristics. Simultaneously, current and voltage sensor modules, such as Hall effect sensors and voltage divider circuits, are connected to the RF relay's power supply circuit to acquire its operating current and voltage characteristics. These sensors continuously acquire data at a fixed sampling frequency (e.g., 100Hz) to form a time series.

[0118] Then, a digital signal processor (DSP) is used to receive raw data from the data acquisition unit. This module filters and reduces noise from the raw data. Subsequently, mature feature extraction algorithms in existing technologies, such as Fast Fourier Transform (FFT), are used to extract characteristic fingerprints representing the state of the RF relay, such as the power spectral density, harmonic distortion, and insertion loss of the RF signal, as well as the root mean square value and peak factor of the operating current. These characteristic fingerprints form their respective time series.

[0119] Then, an embedded industrial computer was used to run time series analysis software. This software analyzed the time series data for each feature fingerprint and built an ARIMA model. For example, for the time series data of insertion loss characteristics under healthy conditions of an RF relay, the computational unit determined the order (p, d, q) of the ARIMA model using the autocorrelation function (ACF) and partial autocorrelation function (PACF). Then, the ARIMA model was trained using methods such as maximum likelihood estimation to obtain the model's parameters. These parameters collectively describe the mean, variance, trend, and autoregressive characteristics of the insertion loss under healthy conditions.

[0120] Finally, the computation unit stores the trained ARIMA model parameters, along with descriptive information about other time-series evolution patterns (such as mean, standard deviation, and Fourier coefficients of periodic components), into a non-volatile memory, such as a solid-state drive (SSD) or flash memory. This stored information constitutes the baseline pattern of the RF relay's health status. Each feature fingerprint corresponds to one or a set of baseline patterns.

[0121] In step S43, for comparing correlation patterns, the distance (e.g., Frobenius norm) between the real-time correlation coefficient matrix and the baseline correlation coefficient matrix can be calculated. For comparing temporal evolution patterns, the differences between the parameters of the real-time time series model and the parameters of the baseline model can be compared, or the residuals between the real-time feature fingerprint sequence and the predicted values ​​of the baseline time series model can be evaluated. Thus, one or more pattern comparison results can be obtained, which quantify the degree of deviation of the real-time pattern from the baseline pattern.

[0122] In step S44, one or more thresholds can be set. When the pattern comparison result exceeds these thresholds, it is considered that there is a pattern deviation. Quantifying the pattern deviation means converting the degree of pattern deviation into a numerical indicator, such as a deviation score, distance value, or probability value, for further analysis and decision-making.

[0123] Through the above technical solutions, this application can significantly improve the accuracy and sensitivity of RF relay performance monitoring. By analyzing the correlation patterns and time-series evolution patterns of characteristic fingerprints, it can effectively identify performance degradation caused by complex coupling effects or gradual degradation that is difficult to detect using traditional methods. In addition, by quantifying pattern deviations, it can provide more accurate input for subsequent degradation contribution assessment, helping to more accurately distinguish the impact of RF relay degradation itself on other parts of the system or environmental changes, reducing false alarms and missed alarms, and improving maintenance efficiency and equipment reliability.

[0124] In step S5, to more accurately determine the degradation of the RF relay itself, it is necessary to consider changes in the overall system environment. For example, in a typical RF relay array, multiple relays are usually closely arranged and share part of the upstream signal link. The monitoring device simultaneously evaluates the deviation of the characteristic fingerprint of the target relay and several physically adjacent relays. If a characteristic fingerprint, such as in-band noise power, is found to deviate in a similar magnitude and with a synchronized trend from that of the target relay and all its adjacent relays within the same time period, then this generalized change is very likely caused by a common upstream device, such as gain drift of a preamplifier low-noise amplifier, or an increase in the ambient temperature of the entire chassis. In this case, this generalized deviation is defined as a dynamic reference deviation benchmark. Conversely, if only the characteristic fingerprint of the target relay deviates significantly, while its adjacent relays remain normal, then this deviation is very likely caused by a localized problem within the target relay itself.

[0125] The establishment of the dynamic reference deviation benchmark is based on the principle of spatial correlation and is achieved through collaborative analysis of adjacent RF relay groups on the same physical link or the same backplane. In a specific embodiment, adjacent RF relays refer to relays that are physically adjacent to the target relay within the same chassis or rack and share the same power supply bus or upstream signal distribution network. Typically, two adjacent relays to the left and two to the right of the target relay, for a total of four, are selected as the reference group.

[0126] The dynamic evaluation of the characteristic fingerprint deviation of adjacent radio frequency relays includes: within the same time window (e.g., the most recent 10 minutes), calculating the characteristic fingerprint deviation of the target relay and each adjacent relay respectively, wherein the deviation can be calculated using Euclidean distance or Mahalanobis distance.

[0127] The process of establishing a dynamic reference deviation benchmark includes: performing median filtering on the deviations of adjacent relays, removing outliers outside the range of ±30% of the median deviation, and then calculating the weighted average of the remaining deviations. The weights can be determined based on the physical distance or signal coupling strength between the adjacent relays and the target relay. For example, the closer the distance, the higher the weight, thus obtaining the dynamic reference deviation benchmark.

[0128] In step S6, a relative deviation is obtained by comparing the overall characteristic fingerprint deviation of the target relay with a dynamic reference deviation benchmark, for example, by performing a subtraction operation. This relative deviation effectively isolates the combined effects of systemic and environmental factors, thus reflecting the degradation contribution of the RF relay itself more purely. Furthermore, this relative deviation is analyzed in conjunction with the characteristics of the current service load. For example, if the relative deviation is found to increase significantly only when processing signals with a peak-to-average power ratio, this strongly points to a nonlinear response problem of the relay contacts under high instantaneous power surges, such as micro-discharge or unstable contact resistance. In this way, not only can the problem be confirmed to be with the relay itself, but a preliminary inference of the physical mechanism of the fault can also be made.

[0129] Furthermore, in step S7, issuing a warning about the performance degradation of the radio frequency relay includes the following steps:

[0130] S71: Send early warning information about the performance degradation of the RF relay to the equipment management system or maintenance work order system;

[0131] S72: Automatically generate maintenance tasks or fault reports based on early warning information.

[0132] Specifically, sending RF relay performance degradation warnings to the equipment management system or maintenance work order system means automatically pushing detailed information about the monitored RF relay performance degradation, including but not limited to the degree of degradation, degradation indicators, possible causes, and occurrence time, to the software platform responsible for equipment operation management or maintenance processes through a preset communication interface and protocol. The equipment management system can be understood as a platform for centralized management and monitoring of various devices in the network, aiming to provide visualization of equipment status, configuration management, and performance data analysis. The maintenance work order system can be understood as a system for automating the processing and tracking of maintenance requests, aiming to transform warnings into specific maintenance tasks and manage their lifecycle.

[0133] Furthermore, automatically generating maintenance tasks or fault reports based on early warning information means that once early warning information is received by the equipment management system or maintenance work order system, the system will automatically create corresponding maintenance work orders or fault reports according to preset rules and strategies. For example, if the early warning level is high, the system may automatically generate an emergency maintenance work order and assign it to the appropriate maintenance personnel; if the early warning level is low, it may generate a regular fault report for subsequent analysis. The purpose is to transform performance degradation early warnings into actionable actions, reduce manual intervention, and improve the efficiency and accuracy of fault response.

[0134] In some embodiments described above, this application proposes a scheme to monitor the performance of radio frequency (RF) relays and issue performance degradation warnings. However, in its implementation, simply issuing performance degradation warnings may not be sufficient to support efficient preventative maintenance and resource scheduling. Traditional warning mechanisms typically trigger when performance indicators reach preset thresholds, but fail to provide predictive information about when the RF relay might reach a critical fault state. If these problems are not addressed, maintenance personnel may find it difficult to plan repairs or replacements in advance, potentially leading to sudden failures and affecting the stability and reliability of system operation. Therefore, this application further proposes an intelligent monitoring method for RF relay performance.

[0135] S8: Predicts the time window for the RF relay to reach a critical fault state.

[0136] Specifically, the predicted time window for the RF relay to reach a critical failure state refers to estimating, based on the analysis of the RF relay's performance degradation trend, the probability that the RF relay will reach a critical state of functional failure or require emergency intervention within a certain period of the future. This critical failure state can be understood as the RF relay failing to meet its design performance specifications, or its performance degradation severely impacting the normal operation of the system. Its purpose is to provide maintenance personnel with forward-looking information to plan maintenance activities, spare parts procurement, or system switchover in advance, thereby avoiding unplanned downtime and reducing operating costs.

[0137] Specifically, step S8 includes:

[0138] S81: Continuously monitor the time-series changes of multiple degradation indicators reflected by the degradation of the RF relay itself, and obtain the current values ​​of multiple degradation indicators.

[0139] S82: For each degradation index, calculate the rate of change and acceleration of change of the degradation index based on the time-series changes of the degradation index.

[0140] S83: Based on the characteristics of the current business load, dynamically set the critical threshold and degradation acceleration coefficient for each degradation indicator;

[0141] S84: Identify that the current value, rate of change, or acceleration of change of any of the degradation indicators reaches the preset accelerated degradation trigger condition.

[0142] S85: When an accelerated degradation trigger condition is identified, the time required for the degradation index to reach the critical threshold is calculated based on the current value, rate of change, acceleration of change, critical threshold, and degradation acceleration coefficient of the degradation index.

[0143] S86: Compare all the calculated times and select the time with the smallest value as the time window for the RF relay to reach the critical fault state.

[0144] Specifically, in step S81, multiple degradation indicators may include, but are not limited to, a decrease in the gain of the RF relay, an increase in the VSWR, a decrease in isolation, an increase in intermodulation distortion, or a deterioration in the noise figure. These indicators directly reflect the performance degradation of the RF relay during long-term operation. Continuous monitoring refers to periodically or in real-time collecting data on these indicators and recording their trends over time to construct complete time-series change data.

[0145] In step S82, the rate of change refers to the amount of change in the degradation index per unit time, for example, a decrease in gain of 0.01 dB per hour. The acceleration of change refers to the trend of the rate of change, i.e., whether the degradation rate is accelerating. These parameters can be obtained by differential calculation of historical monitoring data, thereby revealing the dynamic characteristics of the degradation process.

[0146] In step S83, the critical threshold refers to the value at which the RF relay is considered to be in a critical fault state and requires maintenance or replacement. The degradation acceleration factor quantifies the degree of acceleration of the degradation process relative to a standard load under a specific service load. For example, under high-power, high-bandwidth, or high-data-throughput service loads, the degradation rate of the RF relay may be significantly accelerated, requiring a higher degradation acceleration factor to more accurately predict its remaining lifespan. This dynamic setting can be adjusted based on historical data analysis and expert experience.

[0147] In step S84, the accelerated degradation triggering condition can be the current value of one or more degradation indicators, the rate of change, or the acceleration of change exceeding a preset specific threshold. For example, when the rate of change of a certain degradation indicator suddenly increases significantly, or its acceleration reaches a certain critical value, it can be determined that the RF relay has entered the accelerated degradation stage, which usually indicates that a fault is about to occur.

[0148] In step S85, the time required for the degradation index to reach the critical threshold can be calculated using a statistical regression model. This calculation comprehensively considers the current degradation state, degradation trend, and the impact of workload on the degradation process, thereby obtaining a more accurate estimate of remaining lifetime.

[0149] In step S86, all calculated times are compared, and the time value with the smallest value is selected as the time window for the radio frequency relay to reach the critical fault state. This is to ensure that, among all possible fault modes, the indicator that reaches the critical state first is used as the standard, thereby providing the most conservative and safest prediction and avoiding unexpected failures caused by the rapid deterioration of a single indicator.

[0150] Please refer to Figure 2 , Figure 3A radio frequency relay performance intelligent monitoring system, the system being used to implement the steps of any of the above methods, the system comprising:

[0151] Acquisition module 201: Acquires service signals transmitted via radio frequency relays;

[0152] First baseline establishment module 202: When the RF relay is in a healthy state, establish characteristic fingerprint baselines of service signals for different service loads;

[0153] Fingerprint extraction module 203: During normal operation of the radio frequency relay, it extracts the characteristic fingerprint of the service signal in real time and identifies the current service load;

[0154] Comparison module 204: Compares the real-time extracted feature fingerprint with the baseline fingerprint in the feature fingerprint baseline that corresponds to the current service load, and determines the deviation of the feature fingerprint;

[0155] Second baseline establishment module 205: Based on the deviation of the feature fingerprint, dynamically evaluate the deviation of the feature fingerprint of adjacent RF relays, and based on the evaluation results, establish a dynamic reference deviation benchmark for the RF relays;

[0156] Differentiation Module 206: Compares the deviation of the characteristic fingerprint of the RF relay with the dynamic reference deviation benchmark to obtain the relative deviation amount, and quantifies the degradation contribution of the RF relay itself based on the relative deviation amount and the characteristics of the current service load, so as to distinguish the degradation of the RF relay itself from the impact of other parts of the system or environmental changes.

[0157] Early warning module 207: Based on the differentiation results, issue an early warning of performance degradation of the radio frequency relay.

[0158] The intelligent monitoring system for RF relay performance disclosed in this application achieves comprehensive, real-time, and intelligent monitoring of RF relay performance through a modular design. Specifically, the acquisition module 201 continuously acquires service signals, providing a data source for the entire monitoring process. The first baseline establishment module 202 constructs characteristic fingerprint baselines under multiple load conditions while the relay is in a healthy state, providing a reference standard for subsequent performance comparisons. During normal relay operation, the fingerprint extraction module 203 extracts characteristic fingerprints of service signals in real time and identifies service loads, ensuring the timeliness and relevance of monitoring. Subsequently, the comparison module 204 compares the real-time fingerprints with the baseline fingerprints under the corresponding loads to preliminarily determine performance deviations. To improve the accuracy and robustness of the judgment, the second baseline establishment module 205 introduces performance data from adjacent relays and dynamically adjusts the reference deviation benchmark, thereby effectively avoiding interference from common environmental or system changes on the performance evaluation of a single relay. Finally, the differentiation module 206 compares the relay's deviation with the dynamic reference deviation benchmark and, combined with service load characteristics, accurately quantifies the relay's own degradation contribution, thereby accurately distinguishing the relay's own degradation from the impact of other parts of the system or environmental changes. Therefore, the early warning module 207 can issue timely warnings of performance degradation based on accurate differentiation results, avoiding false alarms and missed alarms, and providing a reliable basis for predictive maintenance of radio frequency relays.

[0159] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for intelligent monitoring of the performance of a radio frequency relay, characterized in that, The method includes the following steps: S1: Acquire service signals transmitted via radio frequency relays; S2: When the radio frequency relay is in a healthy state, establish a characteristic fingerprint baseline of the service signal for different service loads; Step S2 includes: S21: When the radio frequency relay is in a healthy state, establish an initial feature fingerprint baseline for the service signal for different service loads; S22: During normal operation of the radio frequency relay, continuously monitor the characteristic fingerprint of the service signal and its deviation from the initial characteristic fingerprint baseline; S23: Based on the deviation, identify the period during which the radio frequency relay is in a stable and healthy operating state, and there is no performance degradation warning during the period; S24: During the stable and healthy operation cycle, if it is determined that the characteristic fingerprint of the service signal continuously and generally deviates from the initial characteristic fingerprint baseline, but the degree of deviation does not reach the performance degradation warning threshold, then it is determined that the characteristic fingerprint baseline has undergone natural drift. S25: Based on the judgment result of the natural drift, and based on the service signal feature fingerprint data collected during the stable and healthy operation cycle, the initial feature fingerprint baseline is incrementally adjusted to obtain the feature fingerprint baseline of the service signal. S3: During normal operation of the radio frequency relay, extract the feature fingerprint of the service signal in real time and identify the current service load; S4: Compare the real-time extracted feature fingerprint with the baseline fingerprint in the feature fingerprint baseline that corresponds to the current service load, and determine the deviation of the feature fingerprint; S5: Based on the deviation of the feature fingerprint, dynamically evaluate the deviation of the feature fingerprint of adjacent RF relays, and based on the evaluation results, establish a dynamic reference deviation benchmark for the RF relay; S6: Compare the characteristic fingerprint deviation of the RF relay with the dynamic reference deviation benchmark to obtain the relative deviation amount, and quantify the degradation contribution of the RF relay itself based on the relative deviation amount and the characteristics of the current service load, so as to distinguish the degradation of the RF relay itself from the impact of other parts of the system or environmental changes. S7: Based on the differentiation results, issue a warning about the performance degradation of the radio frequency relay.

2. The method for intelligent monitoring of radio frequency relay performance according to claim 1, characterized in that, In step S3, the real-time extraction of the feature fingerprint of the service signal includes the following steps: S31: Adjust the feature fingerprint extraction parameters of the service signal according to the identified current service load characteristics, the parameters including sampling rate, processing bandwidth or calculation accuracy; S32: Assign the extraction tasks of different feature fingerprints of the service signal to independent hardware processing units, and perform parallel synchronous calculations on the service signal according to the adjusted feature fingerprint extraction parameters to obtain the feature fingerprint of the service signal; S33: When the data processing load of the service signal exceeds a preset threshold, the higher-order modulation error vector amplitude or signal envelope distortion mode in the service signal is extracted first, and the extraction of other features is downgraded to ensure the real-time extraction of the feature fingerprint of the service signal.

3. The method for intelligent monitoring of radio frequency relay performance according to claim 2, characterized in that, In step S32, the characteristic fingerprint of the service signal includes: harmonic distortion components, intermodulation distortion components, in-band noise power, phase noise spectral density, gain compression characteristics, or amplitude phase conversion characteristics of the service signal.

4. The method for intelligent monitoring of radio frequency relay performance according to claim 1, characterized in that, In step S3, identifying the current service load includes the following steps: S34: Analyze the modulation type, bandwidth, power spectral density, or data throughput parameters of the service signal to obtain the service signal parameters; S35: Match the service signal parameters with a preset service load feature library to obtain a matching result; S36: Identify the current service load based on the matching results.

5. The method for intelligent monitoring of radio frequency relay performance according to claim 1, characterized in that, Step S4 includes: S41: Obtain the correlation pattern between different feature fingerprints of the service signals contained in the baseline fingerprint corresponding to the current service load in the feature fingerprint baseline and the temporal evolution pattern of a single feature fingerprint; S42: Real-time calculation of the real-time correlation patterns between different feature fingerprints of the service signals contained in the real-time extracted feature fingerprints and the real-time temporal evolution patterns of individual feature fingerprints; S43: Compare the real-time correlation pattern with the correlation pattern, and the real-time time-series evolution pattern with the time-series evolution pattern to obtain the pattern comparison result; S44: Based on the pattern comparison results, determine the pattern deviation of the feature fingerprint and quantify the pattern deviation.

6. The method for intelligent monitoring of radio frequency relay performance according to claim 1, characterized in that, In step S7, issuing the warning of performance degradation of the radio frequency relay includes the following steps: S71: Send the warning information about the performance degradation of the radio frequency relay to the equipment management system or maintenance work order system; S72: Based on the warning information, automatically generate maintenance tasks or fault reports.

7. The method for intelligent monitoring of radio frequency relay performance according to claim 1, characterized in that, The method further includes: S8: Predict the time window when the radio frequency relay reaches a critical fault state.

8. The method for intelligent monitoring of radio frequency relay performance according to claim 7, characterized in that, Step S8 includes: S81: Continuously monitor the temporal changes of multiple degradation indicators reflected by the degradation of the radio frequency relay itself, and obtain the current values ​​of the multiple degradation indicators. S82: For each degradation index, calculate the rate of change and acceleration of change of the degradation index based on the time-series changes of the degradation index; S83: Based on the characteristics of the current service load, dynamically set a critical threshold and degradation acceleration coefficient for each degradation index; S84: Identify that the current value of any of the degradation indicators, the rate of change, or the acceleration of change reaches a preset accelerated degradation trigger condition; S85: When the accelerated degradation triggering condition is identified, the time required for the degradation index to reach the critical threshold is calculated based on the current value of the degradation index, the rate of change, the acceleration of change, the critical threshold, and the degradation acceleration coefficient. S86: Compare all the calculated times and select the time value with the smallest value as the time window for the radio frequency relay to reach the critical fault state.

9. A radio frequency relay performance intelligent monitoring system, characterized in that, The system is used to implement the steps of the method of any one of claims 1-8, the system comprising: Acquisition module: Acquires service signals transmitted via radio frequency relays; First baseline establishment module: When the radio frequency relay is in a healthy state, establish the characteristic fingerprint baseline of the service signal for different service loads; Fingerprint extraction module: During normal operation of the radio frequency relay, extracts the characteristic fingerprint of the service signal in real time and identifies the current service load; The fingerprint extraction module is also used to establish an initial feature fingerprint baseline of the service signal for different service loads when the radio frequency relay is in a healthy state. During normal operation of the radio frequency relay, the characteristic fingerprint of the service signal and its deviation from the initial characteristic fingerprint baseline are continuously monitored. Based on the deviation, identify the period during which the radio frequency relay is in a stable and healthy operating state, and there is no performance degradation warning during the period; During the stable and healthy operation cycle, if it is determined that the characteristic fingerprint of the service signal continuously and generally deviates from the initial characteristic fingerprint baseline, but the degree of deviation does not reach the performance degradation warning threshold, then it is determined that the characteristic fingerprint baseline has undergone natural drift. Based on the natural drift judgment result, and based on the service signal feature fingerprint data collected during the stable and healthy operation cycle, the initial feature fingerprint baseline is incrementally adjusted to obtain the feature fingerprint baseline of the service signal. Comparison module: compares the real-time extracted feature fingerprint with the baseline fingerprint in the feature fingerprint baseline that corresponds to the current service load, and determines the deviation of the feature fingerprint; The second baseline establishment module dynamically evaluates the deviation of the characteristic fingerprints of adjacent RF relays based on the deviation of the characteristic fingerprints, and establishes a dynamic reference deviation benchmark for the RF relays based on the evaluation results. Differentiation module: compares the deviation of the characteristic fingerprint of the RF relay with the dynamic reference deviation benchmark to obtain the relative deviation amount, and quantifies the degradation contribution of the RF relay itself based on the relative deviation amount and the characteristics of the current service load, so as to distinguish the degradation of the RF relay itself from the impact of other parts of the system or environmental changes. Early warning module: Based on the differentiation results, issue an early warning of performance degradation of the radio frequency relay.

Images