A method for link state monitoring of a radio frequency combiner
By analyzing downlink transmit power, uplink utilization, and total received power, a target load component and idle thermal noise baseline are constructed. The baseline value is dynamically corrected, excess noise component is extracted, and RF link status is determined. This solves the problem of the impact of service noise and temperature rise interference, and improves the accuracy and robustness of link status monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CIXI ZHIDE COMM TECH CO LTD
- Filing Date
- 2026-05-09
- Publication Date
- 2026-07-21
AI Technical Summary
During peak business hours, existing technologies suffer from noise that masks weak PIM fault signals, and thermal noise drift caused by equipment temperature rise is misjudged as a fault. Static thresholding methods cannot effectively separate fault types, reducing the accuracy of link fault identification.
By acquiring downlink transmit power, uplink utilization, and total received power at different times, the expected load component is constructed. The baseline value is dynamically corrected using the idle thermal noise benchmark and residual power, the excess noise component is extracted, and the RF link status is determined by combining regression feature values, thus separating the influence of service noise and temperature rise interference.
It significantly improves the accuracy and robustness of RF combiner link status monitoring, accurately identifies power-dependent faults such as PIM and external independent interference, and overcomes the misjudgment and omission problems of the traditional static threshold method.
Smart Images
Figure CN122160821B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of wireless communication detection technology, and more specifically to a method for monitoring the link status of an RF combiner. Background Technology
[0002] In mobile communication base station systems, radio frequency combiners are responsible for combining and transmitting multi-band signals and receiving weak uplink signals. With the increase in network frequency bands, the power density carried by radio frequency links significantly increases. Under the thermal stress of long-term high-power signals, connectors and joints are prone to microscopic aging, leading to passive intermodulation (PIM) effects. If the frequency of interference signals generated by PIM falls into the uplink receiving frequency band, it will cause an increase in the received noise floor, affecting communication quality. Existing technologies generally identify fault types using static thresholds. However, during peak traffic periods, a large amount of traffic noise can mask weak PIM fault signals, and thermal noise drift caused by equipment temperature rise can easily be misjudged as a fault. Therefore, existing static threshold methods cannot effectively separate fault types, reducing the accuracy of link fault identification. Summary of the Invention
[0003] To address the issue that during peak business hours, heavy traffic noise can mask weak PIM fault signals, and thermal noise drift caused by equipment temperature rise can easily be misjudged as a fault, existing static threshold methods cannot effectively separate fault types, thus reducing the accuracy of link fault identification. The present invention aims to provide a link status monitoring method for radio frequency combiners, and the specific technical solution adopted is as follows: This invention proposes a link status monitoring method for an RF combiner, the method comprising: In radio frequency link monitoring, downlink transmit power, uplink utilization, and total received power are obtained at different times. The expected load component is obtained by using the linear relationship between uplink utilization and downlink transmit power in the triggered state; the residual power is obtained based on the difference between the total received power and the expected load component in the linear power domain. Based on the linear power domain distribution of the total received power at different times, an idle-time thermal noise reference is obtained; based on the residual power distribution state of the buffer region at different times, and combined with the idle-time thermal noise reference and residual power, a correction basis value is obtained; based on the difference between the residual power and the correction basis value, the excess noise component is obtained. Based on the distribution of the excess noise component, obtain the feature sample set; obtain the regression feature values through the numerical distribution of the elements in the feature sample set. The status of the radio frequency link is determined based on the numerical distribution of the regression feature values.
[0004] Furthermore, the method for obtaining the expected load component includes: Data pairs are filtered based on the difference between a preset non-triggered threshold and downlink transmit power. The data pairs include uplink utilization and total received power. A linear function is obtained based on the linear relationship of all data pairs. The uplink utilization at the current moment is substituted into the linear function to obtain the expected load component at the current moment.
[0005] Furthermore, the method for obtaining the residual power includes: Calculate the linear power values of the total received power and the expected load component respectively; obtain the residual power at different times based on the difference between the linear power values of the total received power and the expected load component.
[0006] Furthermore, the method for obtaining the idle-time thermal noise reference includes: The total received power is filtered by a preset idle time period; the linear power of the filtered total received power is calculated and the mean is obtained to obtain the idle thermal noise benchmark for the idle time period.
[0007] Furthermore, the method for obtaining the correction basis value includes: Based on the distribution of residual power in the buffer, the temperature rise baseline value is obtained; the temperature rise baseline value is then limited by the temperature rise residual threshold to obtain the correction baseline value.
[0008] Furthermore, the method for obtaining the temperature rise residual threshold includes: The maximum temperature rise tolerance is obtained by measuring the fluctuation of the downlink transmit power; a first feature value is obtained by performing a linear power domain transformation based on the feature fusion value of the standard state linear power and the maximum temperature rise tolerance; and a temperature rise residual threshold is obtained based on the first feature value and the linear power value of the expected load component.
[0009] Furthermore, the method for obtaining the excess noise component includes: The excess noise component is obtained by truncating the difference between the residual power and the correction basis value and converting it to the logarithmic domain; the power unit in the logarithmic domain is dBm.
[0010] Furthermore, the method for obtaining the feature sample set includes: Excess noise components are filtered by setting a preset effective detection threshold. The filtering method is as follows: when the excess noise component is greater than the preset effective detection threshold, the excess noise component and its corresponding downlink transmit power are used to form a sample data pair; a feature sample set is formed based on all sample data pairs.
[0011] Furthermore, the method for obtaining the regression feature values includes: By fitting a linear function to the elements in the feature sample set, regression feature values are obtained. The fitting coefficient is denoted as the regression sensitivity coefficient, and the mean of the fitted dependent variable is denoted as the average excess strength of the fit. The regression eigenvalues include the regression sensitivity coefficient and the average excess strength of the fit.
[0012] Furthermore, the method for identifying the radio frequency link includes: The fault level is determined based on preset fault alarm thresholds and slope thresholds. When the average fitted excess strength is less than the fault alarm threshold, the RF link is considered to be in a normal state; when the average fitted excess strength is greater than or equal to the fault alarm threshold, the RF link is considered to be in a fault state. Furthermore, in the fault state, when the regression sensitivity coefficient is greater than or equal to the slope threshold, the fault cause is determined to be linear interference dependent on transmit power; when the regression sensitivity coefficient is less than the slope threshold, the fault cause is determined to be external interference or independent noise.
[0013] The present invention has the following beneficial effects: This invention leverages the superposition characteristics of service load, thermal noise, and potential fault interference in the power domain within an RF link. By analyzing downlink transmit power, uplink utilization, and total received power at different times, it constructs an expected load component characterizing normal service load and decouples the residual power accordingly. Using data from low-load off-peak periods, an off-peak thermal noise benchmark unaffected by service fluctuations is established. Combined with the distribution of residual power in the buffer, the baseline value is dynamically corrected, effectively separating thermal noise drift caused by equipment temperature rise. Based on this, the excess noise component reflecting abnormal interference or faults is accurately extracted, and by analyzing its correlation with downlink transmit power, regression characteristic values such as regression sensitivity coefficient and average fitted excess strength are obtained. This method enables the link status discrimination process to adaptively isolate the influence of service noise and temperature rise interference, accurately identifying and distinguishing power-dependent faults such as passive intermodulation (PIM) from external independent interference. The final discrimination result effectively overcomes the misjudgment and omission problems of traditional static threshold methods in scenarios with high service load and equipment temperature rise during busy periods, significantly improving the accuracy and robustness of RF combiner link status monitoring. Attached Figure Description
[0014] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0015] Figure 1 This is a flowchart of a link status monitoring method for an RF combiner provided in one embodiment of the present invention. Detailed Implementation
[0016] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a link status monitoring method for an RF combiner proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.
[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0018] The specific scheme of the link status monitoring method for an RF combiner provided by the present invention will be described in detail below with reference to the accompanying drawings.
[0019] Please see Figure 1 The diagram illustrates a flowchart of a link status monitoring method for an RF combiner provided in one embodiment of the present invention.
[0020] Step S101: In RF link monitoring, obtain downlink transmit power, uplink utilization, and total receive power at different times.
[0021] The downlink transmit power, uplink utilization, and total receive power of the radio frequency combiner are collected in real time by the base station at a frequency of 15 minutes per acquisition. Downlink transmit power represents the downlink carrier transmit power, and the original unit is usually decibel-milliwatt (dBm); uplink utilization represents the uplink PRB (Physical Resource Block) utilization rate, and the unit is percentage (%), which characterizes the service load; total receive power represents the total receive channel power (RTWP), and the original unit is usually decibel-milliwatt (dBm).
[0022] Step S102: Obtain the expected load component by the linear relationship between uplink utilization and downlink transmit power in the triggered state; obtain the residual power based on the difference between the total received power and the expected load component in the linear power domain.
[0023] Because the uplink data load and RF transmit power are coupled during each communication process, a significant linear deviation between uplink utilization and downlink transmit power in the triggered state indicates an abnormal superposition of passive intermodulation products in the current link, with power-dependent fault components mixed in the signal noise floor. Therefore, a linear function is obtained by examining the linear relationship between uplink utilization and downlink transmit power in the untriggered state. By substituting the current uplink utilization into the linear function, the expected load component is obtained. The larger this value, the higher the current load level of the RF combiner.
[0024] Since the total power of the receiving channel is an incoherent superposition of the service signal, thermal noise, and potential interference in the energy domain (mW), directly subtracting it in the logarithmic domain (dBm) violates physical laws. Therefore, the residual power is obtained by the difference between the total received power and the expected load component in the linear power domain. A positive value indicates that the actual noise floor at the current moment is higher than the historical statistical expectation (possibly due to faults or temperature rise); a negative value indicates that the actual noise floor at the current moment is lower than the historical statistical expectation (possibly due to lower temperature or random fluctuations). The unit of power in the energy domain is mW, and the unit of power in the logarithmic domain is dBm.
[0025] Step S103: Obtain the idle-time thermal noise reference based on the linear power domain distribution of the total received power at different times; obtain the correction basis value based on the residual power distribution state of the buffer region at different times, combined with the idle-time thermal noise reference and the residual power; obtain the excess noise component based on the difference between the residual power and the correction basis value.
[0026] Because the total received power at any given moment includes the superposition of service signals, thermal noise, and potential interference, a safe anchor point unaffected by service fluctuations needs to be calculated to prevent algorithm failure due to continuous high load during subsequent online monitoring. Therefore, historical data needs to be filtered during low-load periods at night (off-peak hours) to remove abnormal samples from peak service periods and establish a benchmark reference characterizing the equipment's inherent thermal noise level. Thus, based on the linear power domain distribution of the total received power at different times, an off-peak thermal noise benchmark is obtained. This value, as a quantitative representation of the thermal noise benchmark during off-peak conditions, can effectively prevent the failure and misjudgment of the minimum filter under continuous high load conditions during busy periods.
[0027] Since residual power includes the superposition of thermal noise drift and potential fault interference, if the residual power in the buffer area shows a continuously rising distribution without any decline, it indicates that the current operating condition is under continuous high load. Conventional minimum value filtering will fail and misjudge the fault. Therefore, based on the residual power distribution in the buffer area at different times, and combined with the idle-time thermal noise benchmark and residual power as constraints, a correction baseline value is obtained to eliminate deadlock misjudgment during busy times, accurately separate thermal noise, and improve robustness under extreme operating conditions.
[0028] Furthermore, the residual power is the component remaining after removing the service load. If it differs significantly from the correction base value, it indicates the presence of abnormal interference signals. The excess noise component is obtained based on the difference between the residual power and the correction base value. The larger this excess noise component, the more severe the link failure or interference.
[0029] Step S104: Obtain the feature sample set based on the distribution of the excess noise component; obtain the regression feature value through the numerical distribution of the elements in the feature sample set.
[0030] Since the excess noise component contains both dead-zone invalid data and real fault information, and invalid data can affect the judgment of link status, a feature sample set is obtained based on the distribution of the excess noise component to characterize the real fault information. Further, by performing logarithmic domain regression fitting on the numerical distribution of elements in the feature sample set, regression feature values characterizing the correlation between noise and transmit power can be obtained, including the regression sensitivity coefficient and the average fitted excess strength. The regression sensitivity coefficient characterizes the degree of dependence of noise on transmit power; a larger value indicates a stronger correlation between the fault and transmit power, and a greater tendency towards PIM linear faults. The average fitted excess strength characterizes the overall intensity level of the fault noise; a larger value indicates a more severe increase in link noise floor and a deeper fault degree.
[0031] Step S105: Determine the status of the radio frequency link based on the numerical distribution of the regression feature values.
[0032] Because the state of the radio frequency link depends on the coupling relationship between the fault intensity and the fault nature, the radio frequency link state discrimination result can be accurately obtained by first determining whether there is a fault by the average fitted excess intensity and then distinguishing the fault type by the regression sensitivity coefficient, based on the numerical distribution of the regression feature values.
[0033] In summary, this invention, based on the superposition characteristics of service load, thermal noise, and potential fault interference in the power domain of an RF link, constructs an expected load component characterizing normal service load by analyzing downlink transmit power, uplink utilization, and total received power at different times, thereby decoupling the residual power. Using data from idle, low-load periods, an idle-time thermal noise benchmark unaffected by service fluctuations is established, and the baseline value is dynamically corrected by combining the distribution of residual power in the buffer, effectively separating thermal noise drift caused by equipment temperature rise. Based on this, the excess noise component reflecting abnormal interference or faults is accurately extracted, and by analyzing its correlation with downlink transmit power, regression characteristic values such as regression sensitivity coefficient and average fitted excess strength are obtained. This method enables the link state discrimination process to adaptively isolate the influence of service noise and temperature rise interference, accurately identifying and distinguishing power-dependent faults such as passive intermodulation (PIM) from external independent interference. The final judgment result can effectively overcome the problems of misjudgment and missed judgment in the traditional static threshold method under the scenarios of high business load and equipment temperature rise during busy hours, and significantly improve the accuracy and robustness of monitoring the link status of radio frequency combiners.
[0034] Preferably, in some possible implementations of this embodiment, the method for obtaining the expected load component includes: Because there is a service load coupling relationship between downlink transmit power and uplink utilization, a significant linear deviation between the two under triggered conditions indicates that power-dependent fault components are mixed in the signal noise floor of the current link. Therefore, the difference between a preset non-triggered threshold and downlink transmit power can be used to filter valid data pairs under low-power conditions. Then, by fitting a linear function to all valid data pairs, the linear relationship between uplink utilization and total received power can be characterized. In this embodiment, the non-triggered threshold is set to 20% of the rated maximum output power of the RF channel. The linear fitting function uses least-squares linear fitting, and its calculation method is a well-known technique, which will not be elaborated here.
[0035] Substituting the current uplink utilization rate into the linear function yields the expected load component, which characterizes the current, uninterrupted load level of the RF combiner. In one specific implementation of this invention, the filtering method for historical data pairs using a non-triggered threshold is as follows: when the downlink transmit power of a historical time is less than or equal to the non-triggered threshold, the data pair for that historical time is extracted. This data pair includes the uplink utilization rate and total received power of that historical time. Then, the uplink utilization rate as the independent variable and the total received power as the dependent variable from all data pairs are input into the linear fitting function, and the output is the fitting function.
[0036] If the uplink utilization rate at the current moment is less than or equal to the non-triggered threshold, the system is determined to be in a low-power condition and there is no physical condition to trigger PIM. In this case, the data point is only used to store in the historical database and does not trigger subsequent decoupling calculations. If the uplink utilization rate at the current moment is greater than the non-triggered threshold, the system is determined to be in a high-power condition and there may be PIM interference. In this case, the uplink utilization rate at the current moment is substituted into the above fitting function to obtain the expected load component at the current moment.
[0037] Preferably, in some possible implementations of this embodiment, the method for obtaining residual power includes: The total power of the received channel is an incoherent superposition of service signals, thermal noise, and potential interference in the energy domain (mW). Directly performing subtraction in the logarithmic domain (dBm) violates physical laws and leads to mathematical distortion of the energy conservation relationship. Conversely, if the power is first converted to the linear power domain for differential operation, and then converted back to the logarithmic domain as needed, the physical nature of signal power superposition can be strictly followed. Therefore, for the power difference between the total received power and the expected load component, the normal service load can be stripped away, the actual noise floor retained, and the hidden abnormal interference signals exposed. Thus, the residual power is obtained by calculating the difference between the linear power values of the total received power and the expected load component. This residual power characterizes the degree to which the measured noise floor deviates from historical statistical expectations. When the residual power is positive, it indicates that the actual noise floor at the current moment is higher than the historical statistical expectations (possibly due to faults or temperature rise); when the residual power is negative, it indicates that the actual noise floor at the current moment is lower than the historical statistical expectations (possibly due to lower temperature or random fluctuations). As a specific example, in a specific implementation of this invention, the residual power can be expressed by the formula:
[0038]
[0039]
[0040] In the formula, This represents the linear power value indicating the total received power at the current moment. This indicates the total received power at the current moment; This represents the linear power value of the expected load component at the current moment; This represents the expected load component at the current moment; This represents the residual power at the current moment.
[0041] Preferably, in some possible implementations of this embodiment, the method for obtaining the idle-time thermal noise reference includes: Since the total received power of the RF combiner at any given time is essentially an incoherent superposition of the service signal, thermal noise floor, and potential interference components in the energy domain, during peak service periods, highly dynamic service noise can mask weak early fault signals. Simultaneously, the temperature rise caused by continuous high-power operation of the equipment can lead to significant drift in the thermal noise floor. Therefore, to prevent algorithm failure due to continuous high load in online monitoring, a safe anchor point unaffected by service fluctuations needs to be calculated. Thus, the total received power is filtered by pre-setting an idle time period; the linear power of the filtered total received power is calculated and averaged to obtain the idle thermal noise benchmark for the idle time period. It should be noted that the idle time period in this embodiment can be from 03:00 to 04:00 AM.
[0042] As a specific example, in one implementation of this invention, the idle-time thermal noise reference can be expressed by the formula:
[0043] In the formula, Indicates the thermal noise reference during idle time; This indicates the number of data items filtered during the off-peak period; This represents the i-th element in the idle time period, which is the total received power at the corresponding moment.
[0044] Preferably, in some possible implementations of this embodiment, the method for obtaining the correction basis value includes: Since residual power inherently contains the superposition of thermal noise drift and potential fault interference, under continuous high load conditions, all data in the buffer may contain high PIM interference components, causing the baseline value extracted by traditional minimum filtering to be abnormally raised. Simultaneously, the cumulative temperature rise caused by long-term equipment operation will cause the thermal noise baseline to continuously deviate from its initial state. Without an effective constraint mechanism, this can easily lead to busy-hour deadlock misjudgment or temperature drift accumulation distortion. Therefore, a dynamic safety boundary unaffected by fault interference needs to be constructed. Thus, based on the distribution of residual power in the buffer, a temperature rise baseline value is obtained. In this embodiment, the minimum value of the residual power is used as the temperature rise baseline value, representing the lower limit level of thermal noise and potential interference components within the buffer. The temperature rise baseline value is truncated to a minimum value using a temperature rise residual threshold to obtain a corrected baseline value, representing the pure temperature rise thermal noise reference level after idle-hour correction. It should be noted that the data stored in the buffer is the residual power of the previous W hours at the current time. In this embodiment, W can be 2. Data is collected once every 15 minutes, and the buffer stores a total of 8 data points.
[0045] As a specific example, in one implementation of this invention, the correction basis value can be expressed by the formula:
[0046]
[0047] In the formula, Indicates the base value of temperature rise; This represents the residual power data stored in the buffer. This represents a function that takes the minimum value. Indicates the correction base value; This represents the temperature rise residual threshold.
[0048] Preferably, in some possible implementations of this embodiment, the method for obtaining the temperature rise residual threshold includes: During long-term high-power operation, the temperature rise of the RF combiner causes a slow and continuous drift in the thermal noise floor, and this drift has a physical upper limit. If a reasonable boundary is not set for the noise increment caused by temperature rise, normal temperature drift may be misjudged as a fault signal, or the algorithm may lose its robustness due to the lack of reference under extreme operating conditions. Therefore, it is necessary to establish a temperature rise tolerance upper limit mechanism that is consistent with the physical characteristics of the equipment and can dynamically adapt to environmental changes.
[0049] Therefore, the maximum temperature rise tolerance is obtained by measuring the fluctuation of downlink transmit power, which characterizes the maximum reasonable rise range of thermal noise under the current environment and equipment conditions. Based on the addition and fusion of the standard state linear power and the maximum temperature rise tolerance, a mathematical mapping is performed in the linear power domain to obtain the first eigenvalue, which characterizes the theoretical upper limit of the thermal noise component in the residual domain after the maximum temperature rise. Furthermore, the first eigenvalue is offset and corrected by combining the linear power value of the expected load component at the current moment, thereby obtaining the temperature rise residual threshold. This threshold serves as a key constraint for determining whether the temperature rise substrate is contaminated by interference, ensuring that temperature drift false alarms can still be effectively suppressed and the minimum filter can be abnormally raised even in busy high interference scenarios.
[0050] As a specific example, in one implementation of this invention, the temperature rise residual threshold can be expressed by the formula:
[0051]
[0052] In the formula, For standard state linear power, denoted as the value of the idle thermal noise reference in the logarithmic domain; Represents the maximum temperature rise tolerance; log represents the logarithmic function; This represents the first eigenvalue. The maximum temperature rise tolerance is the sum of the mean of the downlink transmit power residuals during historical busy periods and the difference of three times the standard deviation. It should be noted that the busy period can be from 12:00 to 13:00, and the historical busy period can be the busy periods of the most recent 7 days.
[0053] Preferably, in some possible implementations of this embodiment, the method for obtaining the excess noise component includes: Since residual power may contain sudden interference or non-steady-state temperature drift components, directly using it for noise assessment can easily lead to inaccurate criteria. Therefore, it is necessary to correct the base value to eliminate sudden interference or non-steady-state temperature drift components in the residual power. Simultaneously, due to measurement errors or over-correction, the residual power after eliminating sudden interference or non-steady-state temperature drift components may be negative or extremely small, rendering the results meaningless. Therefore, the results need to be processed. Thus, the difference between the residual power and the corrected base value is truncated to suppress anomalous disturbances, and then transformed to the logarithmic domain to obtain the excess noise component. This excess noise component can characterize the level of abnormal thermal noise; the larger its value, the more likely the link has abnormal interference, non-steady-state temperature drift, or potential fault risk.
[0054] As a specific example, in one implementation of this invention, the excess noise component can be expressed by the formula:
[0055]
[0056] In the formula, This represents the excess noise at the current moment; max represents the function that takes the maximum value. This represents the excess noise component at the current moment; This represents a very small positive number to prevent overflow in logarithmic operations. In this embodiment, the value can be 0.0001.
[0057] Preferably, in some possible implementations of this embodiment, the method for obtaining the feature sample set includes: To effectively identify abnormal interference events, threshold constraints need to be applied to the excess noise component. Specifically, for the current detection period, a preset effective detection threshold is used to filter it: when the excess noise component is greater than the effective detection threshold, it is determined that there is a significant non-steady-state disturbance at the current moment, and the excess noise component and its corresponding downlink transmit power are obtained to form a valid sample data pair; then, all sample data pairs that meet the conditions are obtained to construct a feature sample set for the identification of the RF combiner link status.
[0058] It should be noted that the effective detection threshold is k times the maximum value of the excess noise component in the previous detection period. Here, k is determined by the significance level, and in this embodiment, 1.05 can be used; the detection period in this embodiment can be 1 day. The significance level is a well-known technique, and its specific calculation will not be elaborated here.
[0059] Preferably, in some possible implementations of this embodiment, the method for obtaining regression feature values includes: Based on the statistical correlation of each data pair in the feature sample set, a linear function is used to perform regression fitting to characterize the quantitative relationship between downlink transmit power and excess noise component. The slope of the obtained fitted line is defined as the regression sensitivity coefficient to characterize the sensitivity of the system to noise rise caused by power changes. At the same time, the mean of the fitted dependent variable (i.e., excess noise component) is recorded as the average fitted excess strength to reflect the average level of noise overshoot within the effective interference range.
[0060] Preferably, in some possible implementations of this embodiment, the method for determining the radio frequency link includes: Based on preset fault alarm thresholds and slope thresholds, the operating status of the RF link is classified and determined. When the average fitted excess strength is lower than the fault alarm threshold, it indicates that the system noise level is within a controllable range, and the RF link is judged to be in a normal state. When the average fitted excess strength reaches or exceeds the threshold, the system is considered to be in a fault state. In the fault state, if the regression sensitivity coefficient is greater than or equal to the preset slope threshold, it indicates that the excess noise component and the downlink transmit power show a significant positive correlation, and the fault cause can be classified as transmit power-dependent linear interference, such as intermodulation distortion or device nonlinear degradation. Conversely, if the regression sensitivity coefficient is less than the slope threshold, it indicates that the excess noise and the transmit power have no obvious linear correlation, and the fault is more likely to originate from external environmental interference, sudden impulse noise, or internal thermal noise anomalies independent of the power amplifier link.
[0061] It should be noted that the preset fault alarm threshold and slope threshold were obtained experimentally in this embodiment. The average excess strength of the fit under N normal conditions was calculated through the above steps. Then, the mean and standard deviation of the average excess strength of the N fits were calculated. The principle is to use the mean plus three times the standard value as the fault alarm threshold. The slope threshold is a preset empirical value, initially set to 0.5 in this embodiment, and can be iteratively updated based on fault samples in subsequent iterations.
[0062] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0063] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.
Claims
1. A method for monitoring the link status of an RF combiner, characterized in that, The method includes: Step S101: In radio frequency link monitoring, obtain downlink transmit power, uplink utilization, and total received power at different times; Step S102: Obtain the expected load component through the linear relationship between uplink utilization and downlink transmit power; obtain the residual power based on the difference between the total received power and the expected load component in the linear power domain. Step S103: Obtain the idle thermal noise reference based on the linear power domain distribution of the total received power at different times; obtain the correction basis value based on the residual power distribution state of the buffer area at different times, combined with the idle thermal noise reference and the residual power; obtain the excess noise component based on the difference between the residual power and the correction basis value. Step S104: Based on the distribution of excess noise components, obtain a feature sample set, which is used to characterize real fault information; obtain regression feature values through the numerical distribution of elements in the feature sample set, which are used to characterize the correlation between noise and transmission power, including regression sensitivity coefficient and average fitting excess strength. Step S105: Determine the status of the radio frequency link based on the numerical distribution of the regression feature values.
2. The link status monitoring method for an RF combiner according to claim 1, characterized in that, The method for obtaining the expected load component includes: Data pairs are filtered based on the difference between a preset non-triggered threshold and downlink transmit power. The data pairs include uplink utilization and total received power. A linear function is obtained based on the linear relationship of all data pairs. The uplink utilization at the current moment is substituted into the linear function to obtain the expected load component at the current moment.
3. The link status monitoring method for an RF combiner according to claim 1, characterized in that, The method for obtaining the residual power includes: Calculate the linear power values of the total received power and the expected load component respectively; obtain the residual power at different times based on the difference between the linear power values of the total received power and the expected load component.
4. The link status monitoring method for an RF combiner according to claim 1, characterized in that, The method for obtaining the idle-time thermal noise reference includes: The total received power is filtered by a preset idle time period; the linear power of the filtered total received power is calculated and the mean is obtained to obtain the idle thermal noise benchmark for the idle time period.
5. The link status monitoring method for an RF combiner according to claim 1, characterized in that, The method for obtaining the correction basis value includes: Based on the distribution of residual power in the buffer, the temperature rise baseline value is obtained; the temperature rise baseline value is then limited by the temperature rise residual threshold to obtain the correction baseline value.
6. The link status monitoring method for an RF combiner according to claim 5, characterized in that, The method for obtaining the temperature rise residual threshold includes: The maximum temperature rise tolerance is obtained by measuring the fluctuation of the downlink transmit power; a first feature value is obtained by performing a linear power domain transformation based on the feature fusion value of the standard state linear power and the maximum temperature rise tolerance; and a temperature rise residual threshold is obtained based on the first feature value and the linear power value of the expected load component.
7. The link status monitoring method for an RF combiner according to claim 1, characterized in that, The method for obtaining the excess noise component includes: The excess noise component is obtained by truncating the difference between the residual power and the correction basis value and converting it to the logarithmic domain; the power unit in the logarithmic domain is dBm.
8. The link status monitoring method for an RF combiner according to claim 1, characterized in that, The method for obtaining the feature sample set includes: Excess noise components are filtered by setting a preset effective detection threshold. The filtering method is as follows: when the excess noise component is greater than the preset effective detection threshold, the excess noise component and its corresponding downlink transmit power are used to form a sample data pair; a feature sample set is formed based on all sample data pairs.
9. The link status monitoring method for an RF combiner according to claim 1, characterized in that, The method for obtaining the regression feature values includes: By fitting a linear function to the elements in the feature sample set, regression feature values are obtained. The fitting coefficient is denoted as the regression sensitivity coefficient, and the mean of the fitted dependent variable is denoted as the average excess strength of the fit. The regression eigenvalues include the regression sensitivity coefficient and the average excess strength of the fit.
10. A link status monitoring method for an RF combiner according to claim 9, characterized in that, The method for identifying the radio frequency link includes: The fault level is determined based on preset fault alarm thresholds and slope thresholds. When the average fitted excess strength is less than the fault alarm threshold, the RF link is considered to be in a normal state; when the average fitted excess strength is greater than or equal to the fault alarm threshold, the RF link is considered to be in a fault state. Furthermore, in the fault state, when the regression sensitivity coefficient is greater than or equal to the slope threshold, the fault cause is determined to be linear interference dependent on transmit power; when the regression sensitivity coefficient is less than the slope threshold, the fault cause is determined to be external interference or independent noise.