Downlink rate balance detection method and device, electronic equipment and storage medium

Abstract

The embodiment of the application relates to the field of network detection, and discloses a downlink rate balance detection method and device, electronic equipment and a storage medium, comprising: acquiring KPI data of a cell in a detection period, the KPI data comprising downlink rates of a first operator and a second operator; calculating a feature index according to the downlink rates of the first operator and the second operator, the feature index representing a difference amplitude and a difference direction between the downlink rates of the first operator and the second operator; acquiring a yesterday baseline and a yesterday allowable deviation corresponding to the detection period; and determining a balance detection result of the cell in the detection period according to the feature index, the yesterday baseline and the yesterday allowable deviation. Downlink rate balance abnormalities of a shared cell are detected in a timely and accurate manner.

Description

Technical Field

[0001] This application relates to the field of network detection technology, and in particular to a method, apparatus, electronic device, and storage medium for detecting downlink rate equalization. Background Technology

[0002] With the large-scale deployment of 5G networks, telecom operators are widely adopting a co-construction and sharing model, where two operators share the same set of 5G NR (5G New Radio) physical base station equipment and provide services to their respective users through a radio resource management mechanism. Under this architecture, the same physical cell simultaneously carries the user traffic of two operators (hereinafter referred to as the first operator and the second operator), and the downlink speed experience of users on both sides should remain basically balanced.

[0003] However, in actual operation, due to differences in wireless resource scheduling strategies, unequal user numbers, and equipment configuration deviations, persistent rate imbalances can easily occur between the two operators. This directly leads to a decline in user experience for one party, triggering user complaints and affecting network quality assessments. Therefore, timely and accurate detection of downlink rate imbalances in shared cells has become one of the core requirements for 5G co-construction and sharing operation and maintenance management. Summary of the Invention

[0004] The purpose of this application is to provide at least one method, apparatus, electronic device and storage medium for detecting downlink rate balance, and at least one scheme for detecting downlink rate balance anomalies in shared cells, which can at least achieve the effect of timely and accurate detection of downlink rate balance anomalies in shared cells.

[0005] To address the aforementioned technical problems, at least one embodiment of this application provides a downlink rate equalization detection method, comprising: acquiring KPI (Key Performance Indicator) data of a cell within a detection period, the KPI data including downlink rates of a first operator and a second operator; calculating a feature index based on the downlink rates of the first operator and the second operator, the feature index representing the magnitude and direction of the difference between the downlink rates of the first operator and the second operator; acquiring the yesterday's baseline and yesterday's allowable deviation corresponding to the detection period, the yesterday's baseline being updated by weighting the previous day's baseline and yesterday's feature difference according to a preset weight, the yesterday's feature difference being equal to the yesterday's feature index minus the previous day's baseline, and the yesterday's allowable deviation being determined based on the previous day's allowable deviation and the yesterday's feature difference; and determining the cell's equalization detection result for the detection period based on the feature index, the yesterday's baseline, and the yesterday's allowable deviation.

[0006] The above method provides a way to detect downlink rate imbalance anomalies among different operators in a cell in a timely and accurate manner.

[0007] Furthermore, by constructing feature indicators that simultaneously include the magnitude and direction of the difference, the final balance detection result can not only determine whether there is an imbalance, but also clearly indicate which operator's downlink rate is too low, so that maintenance personnel can quickly locate the problem.

[0008] Furthermore, yesterday's baseline is not a fixed threshold, but is obtained by weighted updating of the feature differences between the previous day's baseline and yesterday's baseline. It can automatically adjust as the network status of the cell slowly drifts, avoiding the increase in false alarms / false negatives in the balance detection results caused by the static baseline over time.

[0009] Furthermore, the allowable deviation (for the current day) is derived recursively from the historical deviation (for the previous day) and the current feature difference (for the current day), which allows the boundary for anomaly judgment to adapt to data fluctuations, making it more reasonable than a fixed allowable deviation.

[0010] In some optional embodiments, it further includes: when the yesterday's baseline exists, according to the formula Calculate the characteristic differences for the day, where, The feature index, For the aforementioned yesterday's baseline; and, according to the formula Calculate the baseline for the day, where, For the weighting coefficients; and, according to the formula and Calculate the daily allowable deviation and the square of the daily allowable deviation, where, This is the allowable deviation for the day. It is the square of the allowable deviation for the day. The value is the square of yesterday's allowable deviation; the baseline of the current day, the allowable deviation of the current day, the square of the allowable deviation of the current day, the identifier of the cell, and the identifier of the detection period are stored accordingly.

[0011] pass and It implements an exponentially weighted moving average update of the baseline, which preserves historical information in a decaying manner, thereby achieving the effect of both tracking trend changes and suppressing short-term fluctuations.

[0012] In some optional embodiments, the method further includes: when the yesterday's baseline does not exist, following the formula... Calculate the baseline for the day; and, according to and Calculate the daily allowable deviation and the square of the daily allowable deviation; store the daily baseline, the daily allowable deviation, the square of the daily allowable deviation, the cell identifier, and the detection period identifier accordingly, and perform a cold start to skip downlink rate equalization detection in the detection period as a cold start process.

[0013] The cold start method, which involves performing a cold start when the baseline does not exist yesterday, allows newly built cells or cells without historical data to automatically complete initialization without the need for manual baseline setting, and then naturally converge as data accumulates.

[0014] In some optional embodiments, calculating the characteristic index based on the downlink rates of the first operator and the second operator includes: calculating the coefficient of variation according to the following formula: , And, based on the coefficient of variation, the characteristic index is calculated according to the following formula: Alternatively, the characteristic index can be calculated based on the coefficient of variation using the following formula: ;in, This represents the downlink speed of the first operator. This refers to the downlink speed of the second operator. for and The average value, for and coefficient of variation, The feature index is referred to here.

[0015] We obtain the result by normalizing by dividing by the mean. In this way, It is not affected by the absolute speed of the cell. The positive or negative sign can directly indicate which operator's downlink speed is dominant, making it easier for maintenance personnel to quickly locate the degraded operator and shorten the fault location time.

[0016] In some optional embodiments, determining the cell balance detection result for the detection period based on the feature index, the yesterday's baseline, and the yesterday's allowable deviation includes: determining the magnitude relationship between the feature index and a first threshold and a second threshold, where the first threshold = yesterday's baseline - a preset multiple × yesterday's allowable deviation, and the second threshold = yesterday's baseline + a preset multiple × yesterday's allowable deviation; and determining the cell balance detection result for the detection period based on the magnitude relationship.

[0017] In some optional embodiments, the method further includes: storing the balance detection results and the identifiers of the detection periods in the detection records of the cell, wherein any balance detection result is one of the following: downlink rate anomaly of the first operator, downlink rate anomaly of the second operator, or downlink rate balance between the first operator and the second operator; performing anomaly event identification on the detection records, and when it is identified that the balance detection results for multiple consecutive detection periods are all downlink rate anomalies of the first operator, or all downlink rate anomalies of the second operator, and the number of the multiple detection periods exceeds a preset value, determining that there are anomalies in the multiple detection periods, and generating anomaly event information to describe the anomalies.

[0018] An abnormal event is triggered only when multiple consecutive anomalies occur on the same operator side, thus avoiding alarm storms caused by single-point fluctuations and reducing the burden of ineffective handling by operation and maintenance personnel.

[0019] At least one embodiment of this application also provides a downlink rate equalization detection device, comprising: a first acquisition module, configured to acquire KPI data of a cell within a detection period, the KPI data including downlink rates of a first operator and a second operator; a calculation module, configured to calculate a feature index based on the downlink rates of the first operator and the second operator, the feature index representing the magnitude and direction of the difference between the downlink rates of the first operator and the second operator; a second acquisition module, configured to acquire the yesterday's baseline and yesterday's allowable deviation corresponding to the detection period, the yesterday's baseline being updated by weighting the previous day's baseline and yesterday's feature difference according to a preset weight, the yesterday's feature difference being equal to the yesterday's feature index minus the previous day's baseline, the yesterday's allowable deviation being determined based on the previous day's allowable deviation and the yesterday's feature difference; and a determination module, configured to determine the cell's equalization detection result under the detection period based on the feature index, the yesterday's baseline, and the yesterday's allowable deviation.

[0020] In some optional embodiments, it further includes: a first storage module, configured to, when the yesterday's baseline exists, perform a calculation according to the formula... Calculate the characteristic differences for the day, where, The feature index, For the aforementioned yesterday's baseline; and, according to the formula Calculate the baseline for the day, where, For the weighting coefficients; and, according to the formula and Calculate the daily allowable deviation and the square of the daily allowable deviation, where, This is the allowable deviation for the day. It is the square of the allowable deviation for the day. The value is the square of yesterday's allowable deviation; the baseline of the current day, the allowable deviation of the current day, the square of the allowable deviation of the current day, the identifier of the cell, and the identifier of the detection period are stored accordingly.

[0021] In some optional embodiments, the first storage module is further configured to, when the yesterday's baseline does not exist, perform a calculation according to the formula... Calculate the baseline for the day; and, according to and Calculate the daily allowable deviation and the square of the daily allowable deviation; store the daily baseline, the daily allowable deviation, the square of the daily allowable deviation, the cell identifier, and the detection period identifier accordingly, and perform a cold start to skip downlink rate equalization detection in the detection period as a cold start process.

[0022] In some optional embodiments, the calculation module is configured to calculate the coefficient of variation according to the following formula: , And, based on the coefficient of variation, the characteristic index is calculated according to the following formula: Alternatively, the characteristic index can be calculated based on the coefficient of variation using the following formula: ;in, This represents the downlink speed of the first operator. This refers to the downlink speed of the second operator. for and The average value, for and coefficient of variation, The feature index is referred to here.

[0023] In some optional embodiments, the determining module is used to determine the magnitude relationship between the feature index and a first threshold and a second threshold, respectively, wherein the first threshold = yesterday's baseline - a preset multiple × yesterday's allowable deviation, and the second threshold = yesterday's baseline + a preset multiple × yesterday's allowable deviation; and based on the magnitude relationship, the cell's balance detection result under the detection period is determined.

[0024] In some optional embodiments, the system further includes: a second storage module, configured to store the balance detection results and the identifiers of the detection periods in the detection records of the cell, wherein any balance detection result is one of the following: downlink rate anomaly of the first operator, downlink rate anomaly of the second operator, or downlink rate balance between the first operator and the second operator; and a generation module, configured to perform abnormal event identification on the detection records, and when it is identified that the balance detection results for multiple consecutive detection periods are all downlink rate anomalies of the first operator, or all downlink rate anomalies of the second operator, and the number of the multiple detection periods exceeds a preset value, it is determined that there are abnormal events in the multiple detection periods, and abnormal event information describing the abnormal events is generated.

[0025] At least one embodiment of this application also provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the downlink rate equalization detection method described above.

[0026] At least one embodiment of this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the downlink rate equalization detection method described above.

[0027] At least one embodiment of this application also provides a computer program product, including a computer program that, when executed by a processor, implements the downlink rate equalization detection method described above. Attached Figure Description

[0028] One or more embodiments are illustrated by way of example with reference to the accompanying drawings, and these illustrative descriptions do not constitute a limitation on the embodiments.

[0029] Figure 1 This is a flowchart of a downlink rate equalization detection method provided in one embodiment of this application; Figure 2 This is a schematic diagram of a balance detection result provided in one embodiment of this application; Figure 3 This is a schematic diagram of a downlink rate equalization detection device provided in another embodiment of this application. Detailed Implementation

[0030] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided to make this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0031] Furthermore, the accompanying drawings are merely illustrative of this application and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0032] To facilitate understanding of the embodiments of this application, the relevant content regarding downlink rate equalization detection will be introduced first.

[0033] With the large-scale deployment of 5G networks, telecom operators are widely adopting a co-construction and sharing model, where two operators share the same set of 5G NR physical base station equipment and provide services to their respective users through a radio resource management mechanism. Under this architecture, the same physical cell simultaneously carries the user traffic of two operators (hereinafter referred to as the first operator and the second operator), and the downlink speed experience of users on both sides should remain basically balanced.

[0034] However, in actual operation, due to differences in wireless resource scheduling strategies, unequal user numbers, and equipment configuration deviations, persistent rate imbalances can easily occur between the two operators. This directly leads to a decline in user experience for one party, triggering user complaints and affecting network quality assessments. Therefore, timely and accurate detection of downlink rate imbalances in shared cells has become one of the core requirements for 5G co-construction and sharing operation and maintenance management.

[0035] Several solutions have been proposed in the industry for detecting downlink rate imbalances in co-constructed and shared cells, but they have the following drawbacks: 1. Inability to simultaneously eliminate differences in rate dimensions and perceive the direction of imbalance. The common fixed difference threshold method directly compares the absolute difference in downlink rates between the two sides (two operators). However, using the same threshold for low-speed cells (such as edge coverage cells with rates of only 10-20 Mbps) and high-speed cells (such as hotspot cells with rates of 200-500 Mbps) leads to a large number of missed reports in low-speed scenarios and a large number of false alarms in high-speed scenarios. While the fixed ratio threshold method can partially eliminate the influence of dimensions, it cannot reflect the direction of the imbalance (i.e., which side has a lower downlink rate), making it difficult for maintenance personnel to quickly locate the degraded operator side, thus prolonging fault location time.

[0036] 2. Baselines based on static or mixed-time modeling methods cannot adapt to the slow drift and periodic changes in cell status. On the one hand, statistical methods based on historical averages mix downlink rate data from all time periods throughout the day for modeling. However, the downlink rate balance of shared cells differs significantly between morning and evening peak hours and low-load periods at night. Mixed modeling causes regular imbalances within specific time periods to be masked by mean smoothing, failing to detect periodic imbalance anomalies that only occur in certain fixed time periods. On the other hand, network environment, user distribution, and equipment configuration change slowly over time. Fixed statistical baselines cannot reflect the natural evolution of cell status, and both false alarm and false negative rates increase over time, requiring manual periodic recalibration.

[0037] 3. High engineering complexity or lack of continuous filtering mechanisms. Some solutions attempt to introduce machine learning models for anomaly detection, but these rely on large amounts of labeled data and complex model training and inference environments, making them difficult to implement in resource-constrained operations support systems. Furthermore, the model output lacks interpretability. In addition, such solutions typically judge anomalies based on a single point in time, triggering alarms even for instantaneous fluctuations, leading to alarm storms and increasing the workload of operations and maintenance personnel.

[0038] In summary, how to provide a downlink rate imbalance detection method that can eliminate the difference in rate dimensions and sense the direction of imbalance, adaptively track changes in cell state, and has a simple engineering approach is a technical problem that urgently needs to be solved in this field.

[0039] To address the aforementioned technical problems, this application proposes a downlink rate balance detection method. The implementation details of the downlink rate balance detection method in this embodiment are described below. The following implementation details are provided for ease of understanding and are not essential for implementing this solution.

[0040] Example 1: The downlink rate equalization detection method of this embodiment can be applied to electronic devices with communication, computing, and data storage capabilities. Its specific process can be as follows: Figure 1 As shown, it includes: S101-S104.

[0041] S101, Obtain the KPI data of the cell within the detection period. The KPI data includes the downlink rates of the first operator and the second operator.

[0042] The embodiments of this application do not limit how the detection cycle is divided. For example, the detection cycle is 15 minutes.

[0043] In some examples, obtaining the KPI data of a cell within a detection period may include retrieving the KPI data from a server that stores the KPI data of that cell.

[0044] In other examples, obtaining the cell's KPI data within the detection period may include: obtaining the cell's raw KPI data within the detection period from a server storing the cell's KPI data. This raw KPI data includes: downlink throughput of the first operator and the second operator, downlink RLC (Radio Link Control) layer traffic of the first operator and the second operator, number of RRC (Radio Resource Control) connected users of the first operator and the second operator, and downlink rate of the first operator and the second operator; filtering the raw KPI data to obtain the KPI data.

[0045] In some examples, the filtering conditions for this raw KPI data can be as follows: The downlink throughput on the first operator's side is >0; Downlink throughput on the second operator's side >0; Downlink RLC layer traffic on the first operator's side > 0; Downlink RLC layer traffic on the second operator's side > 0; The number of RRC connection users on the first operator's side is >0; The number of RRC connection users on the second operator's side is >0.

[0046] This method ensures that only KPI records showing actual online users and real traffic from both operators are included in the calculation, thus avoiding interference from invalid rate data during idle periods.

[0047] In some examples, KPI data may also include: the identifier of the community, the identifier of the detection period, the date of the detection period, and optionally the identifier of the province where the community is located.

[0048] S102, based on the downlink rates of the first operator and the second operator, calculate a characteristic index, which represents the magnitude and direction of the difference between the downlink rates of the first operator and the second operator.

[0049] The embodiments of this application do not limit how the characteristic indicators are calculated based on the downlink rates of the first operator and the second operator.

[0050] In some cases, the characteristic indicators calculated based on the downlink rates of the first and second operators may include: calculating the coefficient of variation according to the following formula 1-2: (1) (2) Furthermore, based on this coefficient of variation, the characteristic index is calculated according to the following formula 3: (3) Alternatively, based on the coefficient of variation, the characteristic index can be calculated according to the following formula 4: (4) in, This refers to the downlink speed of the first operator. This refers to the downlink speed of the second operator. for and The average value, for and coefficient of variation, This is the characteristic indicator.

[0051] If we calculate the characteristic index using Formula 3, then This indicates that the second operator has a superior downlink speed. This indicates that the downlink speeds of the first and second operators are completely balanced. This indicates that the first operator has a superior downlink speed.

[0052] If we calculate the characteristic index using Formula 4, then This indicates that the first operator has a superior downlink speed. This indicates that the downlink speeds of the first and second operators are completely balanced. This indicates that the second operator has a superior downlink speed.

[0053] absolute value The larger the value, the greater the difference in downlink speed between the first and second operators.

[0054] In other examples, it is also possible to do so without changing Under the premise of achieving normalization, for The calculation formula undergoes some mathematical transformations. And without changing... The positive and negative signs can distinguish the downlink speeds of different operators, while... The calculation formula is modified mathematically.

[0055] We obtain the result by normalizing by dividing by the mean. In this way, It is not affected by the absolute speed of the cell. The positive or negative sign can directly indicate which operator's downlink speed is dominant, making it easier for maintenance personnel to quickly locate the degraded operator and shorten the fault location time.

[0056] S103, obtain the baseline of yesterday and the allowable deviation of yesterday for the detection period. The baseline of yesterday is obtained by weighting the difference between the baseline of the previous day and the feature difference of yesterday according to a preset weight. The difference between the feature of yesterday is equal to the feature index of yesterday minus the baseline of the previous day. The allowable deviation of yesterday is determined based on the allowable deviation of the previous day and the difference between the feature of yesterday.

[0057] In some cases, the baseline and allowable deviation for each detection period on each date for a given cell are recorded in the baseline data for that cell.

[0058] In some examples, obtaining the yesterday's baseline and yesterday's allowable deviation for the detection period may include: obtaining the yesterday's baseline and yesterday's allowable deviation for the detection period from the baseline data corresponding to the cell.

[0059] For example, if the date of the current detection period is taken as today, then the day before today is yesterday, and the detection period of the day before yesterday, which falls within the same time slot, is yesterday's detection period. The baseline and allowable deviation under yesterday's detection period are called yesterday's baseline and yesterday's allowable deviation. For example, taking a detection period of 15 minutes as an example, the time period corresponding to this detection period is the 09:15 time slot (9:00-9:15) under January 2, 2026, and the yesterday's detection period is the 09:15 time slot under January 1, 2026. That is to say, each detection period corresponds to a baseline and allowable deviation.

[0060] In some cases, the detection cycle identifier does not include the date, but only indicates the time slot. For example, 0915 (9:00-9:15) or 1200 (11:46-12:00). The calculation method for yesterday's baseline and yesterday's allowable deviation is similar to that for today's baseline and today's allowable deviation. The calculation method for today's baseline and today's allowable deviation will be introduced below, and the calculation method for yesterday's baseline and yesterday's allowable deviation will not be repeated hereafter.

[0061] In some examples, this also includes: when the previous day's baseline exists, calculating the daily characteristic difference according to Formula 5: (5) in, This is the characteristic indicator. This is yesterday's baseline.

[0062] And, calculate the baseline for the day according to Formula 6: (6) in, For example, weighting coefficients. It is 0.25, or It can take any other value greater than 0 and less than 1.

[0063] And, calculate the daily allowable deviation and the square of the daily allowable deviation according to formula 7-8: (7) (8) in, This is the allowable deviation for the day. It is the square of the allowable deviation for the day. This is the square of yesterday's allowable deviation.

[0064] In some examples, when the previous day's baseline exists, the current day's baseline, the current day's allowable deviation, and the square of the current day's allowable deviation are calculated, and then the current day's baseline, the current day's allowable deviation, the square of the current day's allowable deviation, the cell's identifier, and the identifier of the detection period are stored accordingly. For example, the baseline data corresponding to the cell is stored.

[0065] pass and It implements an exponentially weighted moving average update of the baseline, which preserves historical information in a decaying manner, thereby achieving the effect of both tracking trend changes and suppressing short-term fluctuations.

[0066] In other examples, the baseline for the current day is calculated according to Formula 9 when the previous day’s baseline does not exist.

[0067] (9) And, calculate the daily allowable deviation and the square of the daily allowable deviation according to formula 10-11: (10) (11) When the baseline of the previous day does not exist, the baseline of the current day, the allowable deviation of the current day, and the square of the allowable deviation of the current day are calculated. The baseline of the current day, the allowable deviation of the current day, the square of the allowable deviation of the current day, the cell identifier, and the identifier of the detection period are stored accordingly. A cold start is performed to skip the downlink rate equalization detection in the detection period as a cold start process.

[0068] The cold start method, which involves performing a cold start when the baseline does not exist yesterday, allows newly built cells or cells without historical data to automatically complete initialization without the need for manual baseline setting, and then naturally converge as data accumulates.

[0069] In some examples, storing the daily baseline, daily allowable deviation, square of daily allowable deviation, cell identifier, and detection period identifier in a corresponding manner can include: left-joining the daily baseline, daily allowable deviation, square of daily allowable deviation, cell identifier, and detection period identifier with the data from the previous day's detection period, with the join key being the cell identifier and the detection period identifier.

[0070] S104. Based on the characteristic index, the baseline of yesterday, and the allowable deviation of yesterday, determine the cell's balance detection result for this detection period.

[0071] In some examples, the cell balance detection result for the detection period is determined based on the characteristic indicator, the previous day's baseline, and the previous day's allowable deviation. This includes: determining the magnitude relationship between the characteristic indicator and a first threshold and a second threshold, where the first threshold = the previous day's baseline - a preset multiple × the previous day's allowable deviation, and the second threshold = the previous day's baseline + a preset multiple × the previous day's allowable deviation; and determining the cell balance detection result for the detection period based on this magnitude relationship.

[0072] The embodiments of this application do not limit the specific value of the preset multiplier. For example, the preset multiplier may be 2, 3, etc.

[0073] In some examples, the balance detection result of the cell under this detection period is determined based on this size relationship. This may include: if the characteristic index is calculated according to Formula 3, such as... Figure 2 As shown, when the characteristic index of the size relationship is less than the first threshold, it indicates that the downlink rate of the second operator is greater, that is, the downlink rate of the first operator is smaller. In this detection period, the balance detection result of the cell is that the downlink rate of the first operator is abnormal. When the characteristic index of the size relationship is greater than the second threshold, it indicates that the downlink rate of the first operator is greater, that is, the downlink rate of the second operator is smaller. In this detection period, the balance detection result of the cell is that the downlink rate of the second operator is abnormal.

[0074] If the characteristic index is calculated according to Formula 4, then when the characteristic index is less than the first threshold, it indicates that the downlink rate of the first operator is greater, which means that the downlink rate of the second operator is smaller. In this detection period, the cell balance detection result is that the downlink rate of the second operator is abnormal. When the characteristic index is greater than the second threshold, it indicates that the downlink rate of the second operator is greater, which means that the downlink rate of the first operator is smaller. In this detection period, the cell balance detection result is that the downlink rate of the first operator is abnormal.

[0075] If the magnitude relationship is such that the characteristic index is between the first threshold and the second threshold, it indicates that the downlink rates of the first operator and the second operator are relatively balanced and there are no abnormalities.

[0076] The above technical solution provides a method for timely and accurate detection of downlink rate imbalance between different operators in a cell.

[0077] Furthermore, by constructing feature indicators that simultaneously include the magnitude and direction of the difference, the final balance detection result can not only determine whether there is an imbalance, but also clearly indicate which operator's downlink rate is too low, so that maintenance personnel can quickly locate the problem.

[0078] Furthermore, yesterday's baseline is not a fixed threshold, but is obtained by weighted updating of the feature differences between the previous day's baseline and yesterday's baseline. It can automatically adjust as the network status of the cell slowly drifts, avoiding the increase in false alarms / false negatives in the balance detection results caused by the static baseline over time.

[0079] Furthermore, the allowable deviation (for the current day) is derived recursively from the historical deviation (for the previous day) and the current feature difference (for the current day), which allows the boundary for anomaly judgment to adapt to data fluctuations, making it more reasonable than a fixed allowable deviation.

[0080] In some examples, the method further includes: storing the balance detection result and the identifier of the detection period in the detection record of the cell, wherein any balance detection result is one of the following: downlink rate anomaly of the first operator, downlink rate anomaly of the second operator, or downlink rate balance between the first operator and the second operator; performing anomaly event identification on the detection record, and when it is identified that the balance detection results for multiple consecutive detection periods are all downlink rate anomalies of the first operator, or all downlink rate anomalies of the second operator, and the number of such multiple detection periods exceeds a preset value, determining that there are anomalies in such multiple detection periods, and generating anomaly event information to describe such anomalies.

[0081] The embodiments of this application do not limit the specific value of the preset value. For example, the preset value is 1.

[0082] The embodiments of this application do not limit the specific information included in the abnormal event information. For example, the abnormal event information may include at least one of the following: the identifier of the cell, the abnormal start time (the start time of the first detection period in the multiple consecutive detection periods), the abnormal end time (the end time of the last period in the multiple consecutive detection periods), the number of detection periods included in the multiple detection periods, and the average characteristic index during the abnormal period.

[0083] An abnormal event is triggered only when an anomaly occurs consecutively for multiple cycles on the same operator side, thus avoiding an alarm storm caused by single-point fluctuations and reducing the burden of ineffective handling by operation and maintenance personnel.

[0084] Example 2: Another embodiment of this application relates to a downlink rate equalization detection device. The implementation details of the downlink rate equalization detection device in this embodiment are described below. The following implementation details are provided for ease of understanding and are not essential for implementing this solution. A schematic diagram of the downlink rate equalization detection device in this embodiment can be seen as follows: Figure 3 As shown, it includes: a first acquisition module 31, used to acquire KPI data of the cell within the detection period, the KPI data including the downlink rates of the first operator and the second operator; a calculation module 32, used to calculate a feature index based on the downlink rates of the first operator and the second operator, the feature index representing the magnitude and direction of the difference between the downlink rates of the first operator and the second operator; a second acquisition module 33, used to acquire the yesterday's baseline and yesterday's allowable deviation corresponding to the detection period, the yesterday's baseline being updated by weighting the feature difference between the previous day's baseline and yesterday according to a preset weight, the yesterday's feature difference being equal to the yesterday's feature index minus the previous day's baseline, the yesterday's allowable deviation being determined based on the previous day's allowable deviation and the yesterday's feature difference; and a determination module 34, used to determine the cell's balance detection result under the detection period based on the feature index, the yesterday's baseline, and the yesterday's allowable deviation.

[0085] In some optional embodiments, the system further includes: a first storage module 35, configured to calculate the daily feature difference according to Formula 1 when the yesterday baseline exists; calculate the daily baseline according to Formula 2; calculate the daily allowable deviation and the square of the daily allowable deviation according to Formulas 3 and 4; and store the daily baseline, the daily allowable deviation, the square of the daily allowable deviation, the identifier of the cell, and the identifier of the detection period accordingly.

[0086] In some optional embodiments, the first storage module 35 is further configured to calculate the current day's baseline according to Formula 5 when the previous day's baseline does not exist; and to calculate the current day's allowable deviation and the square of the current day's allowable deviation according to Formulas 6 and 7; to store the current day's baseline, the current day's allowable deviation, the square of the current day's allowable deviation, the identifier of the cell, and the identifier of the detection period accordingly, and to perform a cold start, so as to skip the downlink rate equalization detection in the detection period as a cold start process.

[0087] In some optional embodiments, the calculation module 32 is used to calculate the coefficient of variation 8-9 according to the following formula; and, based on the coefficient of variation, calculate the feature index 10 according to the following formula; or, based on the coefficient of variation, calculate the feature index according to the following formula 11.

[0088] In some optional embodiments, the determining module 34 is used to determine the magnitude relationship between the feature index and a first threshold and a second threshold, respectively, wherein the first threshold = yesterday's baseline - a preset multiple × yesterday's allowable deviation, and the second threshold = yesterday's baseline + a preset multiple × yesterday's allowable deviation; and based on the magnitude relationship, the cell's balance detection result under the detection period is determined.

[0089] In some optional embodiments, the system further includes: a second storage module 36, configured to store the balance detection result and the identifier of the detection period in the detection record of the cell, wherein any balance detection result is one of the following: downlink rate anomaly of the first operator, downlink rate anomaly of the second operator, or downlink rate balance between the first operator and the second operator; and a generation module 37, configured to identify abnormal events in the detection record, and when it is identified that the balance detection results in multiple consecutive detection periods are all downlink rate anomalies of the first operator, or all downlink rate anomalies of the second operator, and the number of the multiple detection periods exceeds a preset value, it is determined that there are abnormal events in the multiple detection periods, and abnormal event information describing the abnormal events is generated.

[0090] The above method provides a way to detect downlink rate imbalance anomalies among different operators in a cell in a timely and accurate manner.

[0091] Furthermore, by constructing feature indicators that simultaneously include the magnitude and direction of the difference, the final balance detection result can not only determine whether there is an imbalance, but also clearly indicate which operator's downlink rate is too low, so that maintenance personnel can quickly locate the problem.

[0092] Furthermore, yesterday's baseline is not a fixed threshold, but is obtained by weighted updating of the feature differences between the previous day's baseline and yesterday's baseline. It can automatically adjust as the network status of the cell slowly drifts, avoiding the increase in false alarms / false negatives in the balance detection results caused by the static baseline over time.

[0093] Furthermore, the allowable deviation (for the current day) is derived recursively from the historical deviation (for the previous day) and the current feature difference (for the current day), which allows the boundary for anomaly judgment to adapt to data fluctuations, making it more reasonable than a fixed allowable deviation.

[0094] It is worth mentioning that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this application, this embodiment does not introduce units that are not closely related to solving the technical problems proposed in this application; however, this does not mean that other units are absent in this embodiment.

[0095] Example 3: Another embodiment of this application relates to an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the downlink rate equalization detection method in the above embodiments.

[0096] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.

[0097] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.

[0098] Example 4: Another embodiment of this application relates to a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the method embodiments described above.

[0099] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0100] Example 5: Another embodiment of this application relates to a computer program product, including a computer program. When executed by a processor, the computer program implements the method embodiments described above.

[0101] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0102] Furthermore, although the steps of the method in this application are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0103] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware.

[0104] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing this application, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of this application.

Claims

1. A method for detecting downlink rate equalization, characterized in that, include: Acquire key performance indicator (KPI) data of the cell during the detection period, including downlink rates of the first operator and the second operator; Based on the downlink rates of the first operator and the second operator, a characteristic index is calculated, which represents the magnitude and direction of the difference between the downlink rates of the first operator and the second operator. Obtain the yesterday's baseline and yesterday's allowable deviation corresponding to the detection period. The yesterday's baseline is obtained by weighting and updating the feature difference between the previous day's baseline and yesterday according to a preset weight. The feature difference of yesterday is equal to the feature index of yesterday minus the previous day's baseline. The allowable deviation of yesterday is determined based on the allowable deviation of the previous day and the feature difference of yesterday. Based on the aforementioned characteristic indicators, the previous day's baseline, and the previous day's allowable deviation, the cell's balance detection result for the detection period is determined.

2. The method according to claim 1, characterized in that, Also includes: When the aforementioned yesterday's baseline exists, according to the formula Calculate the characteristic differences for the day, where, The feature index, For the aforementioned yesterday's baseline; and, according to the formula Calculate the baseline for the day, where, For the weighting coefficients; and, according to the formula and Calculate the daily allowable deviation and the square of the daily allowable deviation, where, This is the allowable deviation for the day. It is the square of the allowable deviation for the day. This is the square of yesterday's allowable deviation; The daily baseline, the daily allowable deviation, the square of the daily allowable deviation, the cell identifier, and the detection cycle identifier are stored accordingly.

3. The method according to claim 2, characterized in that, Also includes: When the aforementioned yesterday's baseline does not exist, according to the formula Calculate the baseline for the day; and, according to and Calculate the daily allowable deviation and the square of the daily allowable deviation; The daily baseline, the daily allowable deviation, the square of the daily allowable deviation, the cell identifier, and the detection period identifier are stored accordingly, and a cold start is performed to skip downlink rate equalization detection in the detection period as a cold start process.

4. The method according to claim 1, characterized in that, The step of calculating characteristic indicators based on the downlink rates of the first operator and the second operator includes: The coefficient of variation is calculated using the following formula: And, based on the coefficient of variation, the characteristic index is calculated according to the following formula: Alternatively, the characteristic index can be calculated based on the coefficient of variation using the following formula: in, This represents the downlink speed of the first operator. This refers to the downlink speed of the second operator. for and The average value, for and coefficient of variation, The feature index is referred to here.

5. The method according to claim 1, characterized in that, The step of determining the cell balance detection result for the detection period based on the characteristic indicators, the yesterday's baseline, and the yesterday's allowable deviation includes: The relationship between the feature index and the first threshold and the second threshold is determined, wherein the first threshold = yesterday's baseline - preset multiple × yesterday's allowable deviation, and the second threshold = yesterday's baseline + preset multiple × yesterday's allowable deviation; Based on the aforementioned size relationship, the cell balance detection result is determined under the aforementioned detection period.

6. The method according to claim 1, characterized in that, Also includes: The balance detection results and the identifier of the detection period are stored in the detection record of the cell. Any balance detection result is one of the following: the downlink rate of the first operator is abnormal, the downlink rate of the second operator is abnormal, or the downlink rates of the first operator and the second operator are balanced. The detection records are used to identify abnormal events. When the balance detection results for multiple consecutive detection cycles are all abnormal downlink rates of the first operator or all abnormal downlink rates of the second operator, and the number of the multiple detection cycles exceeds a preset value, it is determined that there are abnormal events in the multiple detection cycles, and abnormal event information describing the abnormal events is generated.

7. A downlink rate equalization detection device, characterized in that, include: The first acquisition module is used to acquire the key performance indicator (KPI) data of the cell during the detection period, and the KPI data includes the downlink rates of the first operator and the second operator. The calculation module is used to calculate a characteristic index based on the downlink rates of the first operator and the second operator, wherein the characteristic index represents the magnitude and direction of the difference between the downlink rates of the first operator and the second operator. The second acquisition module is used to acquire the yesterday's baseline and yesterday's allowable deviation corresponding to the detection period. The yesterday's baseline is obtained by weighting and updating the feature difference between the previous day's baseline and yesterday according to a preset weight. The feature difference of yesterday is equal to the feature index of yesterday minus the previous day's baseline. The allowable deviation of yesterday is determined based on the allowable deviation of the previous day and the feature difference of yesterday. The determination module is used to determine the cell balance detection result under the detection period based on the feature indicators, the yesterday's baseline, and the yesterday's allowable deviation.

8. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the downlink rate equalization detection method as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the downlink rate equalization detection method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the downlink rate equalization detection method according to any one of claims 1 to 6.

Images