KR102724600B1 - Method for monitoring wireless network quality and apparatus therefor - Google Patents

Abstract

The present invention relates to a method for monitoring quality deterioration of a base station cell and a device therefor. A method for monitoring wireless network quality by detecting whether a base station cell has deteriorated in a wireless network monitoring device according to an embodiment of the present invention comprises the steps of: collecting a plurality of cell data generated in the base station cell, and obtaining a sample vector from each cell data; counting the number of deviations in which the sample vector deviates from a deterioration reference value; estimating a local probability distribution using sample vectors obtained during a predetermined period of time if the counted number of deviations exceeds a threshold number set in advance; and checking the similarity between the estimated local probability distribution and the global probability distribution, and determining that quality deterioration has occurred in the base station cell if the similarity is less than or equal to a threshold value.

Description

Method for monitoring wireless network quality and apparatus therefor {Method for monitoring wireless network quality and apparatus therefor}

The present invention relates to a method for monitoring wireless network quality, and more specifically, to a method for monitoring quality deterioration of a base station cell and a device therefor.

In order to ensure the stability of mobile communication network services, each communication service provider is monitoring key equipment and key facilities. In particular, communication service providers collect data generated from the communication network and analyze this data to continuously check the status of the mobile communication system.

However, although the criteria for occurrence of failures such as equipment or facility failure or damage are clear, there is a problem that the criteria for occurrence of quality deterioration are not clear when monitoring wireless network quality. Accordingly, managers set the criteria for KPI (Key Performance Indicator) as reference data to evaluate the quality of wireless networks, and compare the KPI criteria with the data currently collected from the wireless network to evaluate the quality of the wireless network. However, there is a problem that the accuracy of the KPI criteria set directly by the manager varies depending on the manager's ability and know-how, and is set differently by country.

In addition, cells formed at base stations have different characteristics depending on the surrounding environment, and there is a problem that it is difficult to accurately monitor the quality of base station cells when setting a uniform KPI standard. In addition, there are too many base stations installed, so the existing method of managers individually monitoring the deterioration of base station cells also has physical limitations.

The present invention has been proposed to solve such conventional problems, and its purpose is to provide a wireless network quality monitoring method and a device therefor that automatically monitors the status of a base station cell by reflecting the environmental characteristics surrounding the base station cell.

Other objects and advantages of the present invention can be understood by the following description, and will be more clearly known by the embodiments of the present invention. In addition, it will be easily understood that the objects and advantages of the present invention can be realized by the means and combinations thereof indicated in the claims.

In order to achieve the above object, according to a first aspect of the present invention, a method for monitoring wireless network quality by detecting whether a base station cell is deteriorated in a wireless network monitoring device comprises the steps of: collecting a plurality of cell data generated in the base station cell, and obtaining a sample vector from each cell data; counting the number of deviations in which the sample vector deviates from a deterioration reference value; estimating a local probability distribution using sample vectors obtained during a predetermined period of time if the counted number of deviations exceeds a threshold number set in advance; and checking the similarity between the estimated local probability distribution and the global probability distribution, and determining that quality deterioration has occurred in the base station cell if the similarity is less than or equal to a threshold value.

The method further includes, prior to the step of obtaining the sample vector, a step of obtaining a sample vector from each of a plurality of cell data, applying the obtained sample vector to machine learning to estimate the global probability distribution; and a step of setting a sample vector corresponding to a predetermined percentage in the global probability distribution as the deterioration reference value.

The above global probability distribution can be updated to reflect additional sample vectors when additional sample vectors are acquired. When the above global probability distribution is updated, the degradation criterion value is also changed.

The step of estimating the above local probability distribution can estimate the above local probability distribution by applying the sample vectors acquired during the above period of time to machine learning.

The above local probability distribution has a shorter sample vector collection period than the above global probability distribution, and the number of sample vectors used as samples is smaller than that of the global probability distribution.

The above method may further include, after the above judgment step, a step of checking the direction and size of sample vectors that have deviated from the above deterioration reference value; a step of checking a deterioration area corresponding to the checked direction and size; and a step of checking a deterioration cause corresponding to the checked deterioration area from deterioration cause data for each deterioration area stored in advance.

The above method may further include a step of outputting or transmitting to an administrator's terminal a fault alarm message including identification information of a cell in which deterioration has occurred and the cause of the deterioration confirmed above.

The above-described method can be recorded on a computer-readable recording medium and used as a program.

According to a second aspect of the present invention for achieving the above object, a device for monitoring the quality status of a base station cell includes: a sample vector acquisition unit for collecting a plurality of cell data generated in the base station cell and acquiring a sample vector from each cell data; a probability distribution estimation unit for estimating a global probability distribution using the plurality of sample vectors acquired by the sample vector acquisition unit and selectively estimating a local probability distribution using the sample vectors acquired during a predetermined period of time; and a quality monitoring unit for counting the number of deviations in which the sample vector deviates from a deterioration reference value, estimating a local probability distribution through the probability distribution estimation unit when the counted number of deviations exceeds a threshold number set in advance, and checking the similarity between the estimated local probability distribution and the global probability distribution, and determining that quality deterioration has occurred in the base station cell when the similarity is less than or equal to a threshold value.

According to an embodiment of the present invention, there is an advantage in that different degradation criteria are applied to each base station cell by considering the surrounding environment of the base station cell, and quality degradation of each base station cell is accurately confirmed according to the different degradation criteria.

In addition, according to an embodiment of the present invention, there is an advantage in that the base station cell-specific deterioration criterion can be set more accurately by deriving the base station cell-specific deterioration criterion based on machine learning.

In addition, according to an embodiment of the present invention, there is an advantage of reducing the labor of the manager by automatically detecting whether or not each base station cell is deteriorated.

In addition, according to an embodiment of the present invention, there is an effect of allowing a failure of a base station cell to be quickly processed by estimating the cause of deterioration of the base station cell and providing it to the manager.

The following drawings attached to this specification illustrate preferred embodiments of the present invention and, together with specific details for carrying out the invention, serve to further understand the technical idea of the present invention; therefore, the present invention should not be interpreted as being limited to matters described in such drawings.
FIG. 1 is a diagram illustrating a system for monitoring wireless network quality according to one embodiment of the present invention.
FIG. 2 is a drawing showing a wireless network monitoring device according to one embodiment of the present invention.
FIG. 3 is a flowchart illustrating a method for estimating a global probability distribution and setting a degradation threshold in a wireless network monitoring device according to one embodiment of the present invention.
FIG. 4 is a flowchart illustrating a method for monitoring whether a base station cell is deteriorated in a wireless network monitoring device according to one embodiment of the present invention.
FIG. 5 is a drawing that two-dimensionally simplifies and expresses a deteriorated area according to one embodiment of the present invention.

The above-described objects, features and advantages will become more apparent through the following detailed description with reference to the attached drawings, so that those with ordinary skill in the art to which the present invention pertains can easily practice the technical idea of the present invention. In addition, when describing the present invention, if it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. Hereinafter, a preferred embodiment according to the present invention will be described in detail with reference to the attached drawings.

FIG. 1 is a diagram illustrating a system for monitoring wireless network quality according to one embodiment of the present invention.

As illustrated in FIG. 1, a system for monitoring wireless network quality includes a plurality of RUs (Radio Units) (100), gNB-DUs (next Generation NodeB-Digital Units) (300), gNB-CUs (ext Generation NodeB-Central Units) (400), and a wireless network monitoring device (200).

The RU (100) is a base station device forming a cell, which forms a wireless link with a mobile terminal and also transmits and receives data to and from the mobile terminal. The RU (100) can communicate with the mobile terminal through a certain frequency band. The RU (100) periodically measures one or more of a handover success rate, the number of RRE (RRC Connection Re-establishment) occurrences, the number of connected mobile terminals, uplink SINR (Signal to Interference Noise Ratio), and RSSI (Received Signal Strength Indicator), and periodically generates cell data including one or more of the measured number of RRE occurrences, the number of connected mobile terminals, uplink SINR, and RSSI, and then provides the cell data to a wireless network monitoring device (200). Identification information of the base station cell is recorded in the cell data. The RSSI is an RSSI generated in uplink communication.

The gNB-DU (300) is a unit distributed in a 5G RAN (Radio Access Network) and can be connected to multiple RUs (100). The gNB-DU (300) communicates with the RU (100) through a fronthaul interface defined by a manufacturer, etc., and also communicates with the gNB-CU (400) using the F1 interface defined by 3GPP.

The gNB-CU (400) is a unit that communicates with the gNB-DU (300) and performs call processing for each gNB-DU (300). In addition, the gNB-CU (400) can communicate with each communication unit included in the core network (e.g., MME AMF, SMF, etc.).

Meanwhile, the gNB-CU (400) and the gNB-DU (300) may be configured as an integrated unit, or the gNB-DU (300) and the RU (100) may be configured as an integrated unit.

The wireless network monitoring device (200) receives cell data of each RU (100) through the gNB-CU (400) and the gNB-DU (300), obtains a sample vector from the cell data, analyzes the sample vector through machine learning, sets a deterioration criterion for each RU (100), and monitors whether each RU (i.e., base station cell) is deteriorated based on the deterioration criterion.

FIG. 2 is a drawing showing a wireless network monitoring device according to one embodiment of the present invention.

As illustrated in FIG. 2, a wireless network monitoring device (200) according to one embodiment of the present invention includes a sample vector acquisition unit (210), a probability distribution estimation unit (220), a storage unit (230), a quality monitoring unit (240), and a deterioration cause identification unit (250), and these components may be implemented in hardware or software, or may be implemented through a combination of hardware and software.

In addition, the wireless network monitoring device (200) includes one or more processors and a memory, and the sample vector acquisition unit (210), the probability distribution estimation unit (220), the quality monitoring unit (240), and the deterioration cause identification unit (250) can be loaded (stored) in the memory in the form of a program executed by the processor.

The storage unit (230) is a storage means such as a memory or a disk device, and stores a plurality of cell data. In addition, the storage unit (230) stores one or more deterioration areas expressed as vectors, and also stores deterioration causes for each deterioration area. The deterioration causes for each deterioration area may be set by an administrator's input and stored in the storage unit (230), or the deterioration causes for each deterioration area may be learned through machine learning and stored in the storage unit (230).

FIG. 5 is a drawing that two-dimensionally simplifies and expresses a deteriorated area according to one embodiment of the present invention.

In Fig. 5, it is exemplified that there are a total of eight deterioration areas (A1 to A8). In addition, the boundary of the boundary area (51) having a distorted circular shape based on the center point is expressed by the deterioration reference value. That is, the area (N) within the boundary area (51) indicating the deterioration reference value is the area of the sample vector indicating normal quality, and the area (A1 to A8) beyond the boundary area (51) is the area where the sample vector indicating quality deterioration is located. Here, the sample vector is data generated in a mobile communication network that can be used to determine whether there is deterioration among cell data, and the size of the data value can be used as a scalar, and the type of the data can be used as the direction of the vector.

As the above sample vector, one or more of the number of RRE (RRC Connection Re-establishment) occurrences, the number of connected terminals, the uplink SINR (Signal to Interference Noise Ratio), the handover success rate, and the RSSI (Received Signal Strength Indicator) may be used, and if multiple data are used in the sample vector, the sample vector becomes multi-dimensional.

As illustrated in Fig. 5, a sample vector (52) having a short length is located at the deterioration reference value, and a sample vector (53) having a long length deviates from the deterioration reference value and is located in the A2 deterioration area.

Different causes of deterioration are mapped to each of the above deterioration areas and stored in the storage unit (230). For example, information on causes of deterioration, such as excessive traffic in the deterioration area of A1, excessive number of connected terminals in the deterioration area of A2, and occurrence of signal interference in the deterioration area of A3, can be recorded in the storage unit (230).

Referring again to FIG. 2, the sample vector acquisition unit (210) acquires a sample vector from cell data. When cell data is received from the RU (100), the sample vector acquisition unit (210) acquires one or more of the number of RRE occurrences, the number of connected mobile terminals, uplink SINR, and RSSI as a sample vector from the cell data. When the sample vector acquisition unit (210) acquires multiple sample vectors from the cell data, the sample vectors are multidimensional vectors.

The above sample vector acquisition unit (210) can accumulate and store a plurality of sample vectors that serve as a basis when generating a probability distribution in the storage unit (230). Specifically, the sample vector acquisition unit (210) acquires a plurality of sample vectors from each cell data collected during a probability distribution analysis period (e.g., over the past three months), and accumulates and stores the acquired plurality of sample vectors in the storage unit (230).

The probability distribution estimation unit (220) analyzes a plurality of sample vectors stored in the storage unit (230) through machine learning during the probability distribution analysis period, and estimates a global probability distribution for the sample vectors. In addition, the probability distribution estimation unit (220) analyzes a plurality of sample vectors stored in the storage unit (230) through machine learning from a pre-set past time (e.g., one week ago) to the time when the possibility of deterioration is detected, and estimates a local probability distribution for the sample vectors.

The above global probability distribution is estimated using sample vectors acquired during the probability distribution analysis period, and thus has a longer collection period of sample vectors than the local probability distribution, and furthermore has a larger number of samples (i.e., sample vectors) than the local probability distribution. In contrast, the local probability distribution is estimated based on sample vectors collected from a specific point in the past of RU (100) to the point at which the possibility of deterioration is detected, and thus has a shorter collection period of sample vectors than the global probability distribution, and furthermore has a smaller number of samples (i.e., sample vectors) than the global probability distribution.

The probability distribution estimation unit (220) can estimate the global/local probability distribution for the sample vectors by estimating the multivariate probability density using the plurality of sample vectors. The probability distribution estimation unit (220) can identify the type and characteristic parameters representing the probability distribution closest to the sample vector distribution by using one or more of the already known probability distribution functions such as Gaussian, Poisson, and Exponential based on the Parametric Density Estimation, and can estimate the local/global probability distribution for the sample vector based on the parameters and types. Alternatively, the probability distribution estimation unit (220) can identify the type and characteristic parameters representing the probability distribution closest to the sample vector distribution by using a known probability distribution function based on a Parametric Mixture Model, and can estimate the local/global probability distribution for the sample vector based on the parameters and types.

In addition, the probability distribution estimation unit (220) can additionally use non-parametric density estimation techniques, such as kernel density estimation (KDE) and k nearest neighbor (k-NN), to estimate local/global probability distributions for sample vectors.

The above probability distribution estimation unit (220) re-estimates the global probability distribution when a certain number of new sample vectors are acquired or when an update cycle arrives, thereby periodically updating the existing global probability distribution.

The probability distribution estimation unit (220) identifies a sample vector corresponding to a pre-set percentage (e.g., the lower 2%) in the estimated global probability distribution and sets this sample vector as the deterioration reference value of the RU (100).

The deterioration criterion can also be expressed using the following mathematical formula.

The probability distribution of the sample vector is estimated using the multivariate probability density function (M-PDF) estimated based on the sample vector. It is written as above The multivariate cumulative distribution function (M-CDF) obtained by integrating If expressed as, is a sample vector A symbol indicating the probability that the above value will be observed. Is It has the opposite meaning to .

The deterioration criterion of the above sample vector is the probability of occurrence of the sample vector observed only in the direction in which the quality deteriorates. can be defined as a value that starts to become less than or equal to the value below. Here, is a sample vector It means the unit vector of and is a symbol for defining the direction of the corresponding sample vector. In general, Deterioration criterion determined by value can be expressed as the mathematical formula below.

The above deterioration criterion is expressed in vector form, and when the probability distribution of the sample vector changes by the probability distribution estimation unit (220), the above deterioration criterion also changes.

When the degradation threshold of the RU (100) is set, the quality monitoring unit (240) analyzes the cell data received from the corresponding RU (100) to monitor whether degradation occurs in the base station cell (i.e., RU). Specifically, when the quality monitoring unit (240) periodically acquires a new sample vector for the RU (100) from the sample vector acquisition unit (210), the quality monitoring unit (240) counts the number of deviations in which the sample vector exceeds the degradation threshold during a specified period (e.g., for 1 day), and if the counted number of deviations exceeds a threshold, the quality monitoring unit (240) determines that the RU (100) has a high possibility of degradation and requests the probability distribution estimation unit (220) to estimate a local probability distribution for the sample vector of the RU (100). In addition, the quality monitoring unit (240) checks the similarity between the global probability distribution and the local probability distribution, and if the similarity is lower than a threshold value set in advance, the quality monitoring unit (240) finally determines that degradation has occurred in the RU (100).

The above quality monitoring unit (240) can check the similarity by calculating the distance difference between the global probability distribution and the local probability distribution.

The global probability distribution function estimated through machine learning , and the local probability distribution function is If you say so, and The distance between them can be calculated using the Kullback-Leibler Divergence (KLD) as follows.

The distance calculated through the above mathematical formula represents the similarity between the global probability distribution and the local probability distribution. The shorter the distance, the higher the similarity, and the longer the distance, the lower the similarity.

When the quality monitoring unit (240) determines that the RU (100) is deteriorated, the deterioration cause confirmation unit (250) confirms the cause of the deterioration, generates a fault notification message including the deterioration cause and cell identification information, and outputs the message to a display device (not shown in the drawing) or transmits the message to a manager's terminal. The deterioration cause confirmation unit (250) confirms the direction and size of sample vectors that have deviated from the deterioration reference value, and further confirms the deterioration area corresponding to the direction and size of the sample vectors in the storage unit (230), and then confirms the deterioration cause corresponding to the deterioration area in the storage unit (230), thereby confirming the deterioration cause of the RU (100).

FIG. 3 is a flowchart illustrating a method for estimating a global probability distribution and setting a degradation threshold in a wireless network monitoring device according to one embodiment of the present invention.

Referring to FIG. 3, the sample vector acquisition unit (210) periodically receives cell data from the RU (100) and accumulates and stores it in the storage unit (230) (S301).

Next, when the generation cycle of the global probability distribution arrives, the sample vector acquisition unit (210) accumulates in the storage unit (230) and acquires one or more of the number of RRE occurrences, the number of connected mobile terminals, uplink SINR, and RSSI as a sample vector from the cell data stored for a certain period of time (e.g., for 3 months) (S303).

Then, the probability distribution estimation unit (220) analyzes a plurality of sample vectors stored in the storage unit (230) through machine learning during the probability distribution analysis period to estimate the global probability distribution (S305).

The above probability distribution estimation unit (220) uses at least one of already known probability distribution functions, such as Gaussian, Poisson, and exponential, based on parametric density estimation, to identify a type and characteristic parameters representing a probability distribution closest to the sample vector distribution, and estimate a global probability distribution for the sample vector based on these parameters and types. Alternatively, the probability distribution estimation unit (220) uses a known probability distribution function based on a parametric mixture model to identify a type and characteristic parameters representing a probability distribution closest to the sample vector distribution, and estimate a global probability distribution for the sample vector based on these parameters and types.

Next, the probability distribution estimation unit (220) checks a sample vector corresponding to a pre-set percentage (e.g., the lower 2%) in the estimated global probability distribution, and sets this sample vector as a deterioration reference value of the RU (100) (S307).

When a certain number or more of additional cell data are newly stored in the storage unit (230), the probability distribution estimation unit (220) additionally obtains sample vectors from each of the newly stored plurality of cell data, analyzes the additionally obtained sample vectors again using machine learning, and updates the global probability distribution so that the additionally obtained sample vectors are reflected.

FIG. 4 is a flowchart illustrating a method for monitoring whether a base station cell is deteriorated in a wireless network monitoring device according to one embodiment of the present invention.

Referring to FIG. 4, when a global probability distribution is generated and a deterioration criterion value is set in the probability distribution estimation unit (220) as in FIG. 3, the sample vector acquisition unit (210) monitors whether cell data is received from the RU (100) (S401).

When cell data is received in the sample vector acquisition unit (210), one or more of the number of RRE occurrences, the number of connected mobile terminals, uplink SINR, and RSSI is acquired as a sample vector from the cell data, and analysis of the sample vector is requested to the quality monitoring unit (240) (S403).

Then, the quality monitoring unit (240) checks whether the sample vector exceeds the deterioration criterion (S405), and if so, increases the number of deviations by one (S407). The number of deviations is the number of times the sample vector exceeds the deterioration criterion, and is initially set to '0'. Next, the quality monitoring unit (240) determines whether the number of deviations exceeds a threshold number (e.g., 5) during a specified period (e.g., 1 day) (S409). If the number of deviations does not exceed the threshold number during the specified period, the number of deviations can be initialized to '0' again.

If the number of deviations does not exceed the threshold number during a specified period, the quality monitoring unit (240) determines that a sample vector temporarily showing deterioration due to an external environment is abnormally acquired, and restarts step S401 and thereafter to monitor the cell data of the RU (100).

On the other hand, if the number of departures exceeds a threshold number during the specified period, the quality monitoring unit (240) determines that there is a high possibility of deterioration of the RU (100) and requests a local probability distribution for the sample vector of the RU (100) from the probability distribution estimation unit (220).

Then, the probability distribution estimation unit (220) analyzes multiple sample vectors stored in the storage unit (230) from a preset past time (e.g., one week ago) to the present time through machine learning to estimate a local probability distribution (S411). The local probability distribution has a shorter collection period of sample vectors than the global probability distribution, and also has a smaller number of samples (i.e., sample vectors) than the global probability distribution. As described above, the probability distribution estimation unit (220) can estimate a local probability distribution using various known functions and machine learning.

Next, the quality monitoring unit (240) checks the similarity between the global probability distribution and the local probability distribution and checks whether this similarity is below a threshold value set in advance (S413, S415). At this time, the quality monitoring unit (240) calculates the distance between the global probability distribution and the local probability distribution through the Kullback-Leibler Divergence (KLD) and checks whether this distance is below the threshold distance, thereby verifying the similarity between the global probability distribution and the local probability distribution.

Next, the quality monitoring unit (240) determines that quality deterioration has occurred in the RU (100) if the similarity between the global probability distribution and the local probability distribution is below the threshold value (S417). Then, the deterioration cause confirmation unit (250) confirms the direction and size of the samples that exceed the deterioration criterion value, and confirms the deterioration area corresponding to the direction and size of these samples in the storage unit (230). As exemplified in Fig. 5, various deterioration areas can be defined, and the deterioration cause confirmation unit (250) confirms the deterioration area corresponding to the direction and size of the samples that exceed the deterioration criterion value in the storage unit (230).

And the deterioration cause confirmation unit (250) confirms the deterioration cause corresponding to the confirmed deterioration area in the storage unit (230) and estimates the deterioration cause (S419). Next, the deterioration cause confirmation unit (250) generates a failure notification message including the estimated deterioration cause and the identification information of the RU (100) where the deterioration occurred, and transmits the failure notification message to the administrator's terminal (S421).

The method for monitoring the quality of the RU (100) described above can be performed independently for each RU (100) in the wireless network monitoring device (200). That is, the wireless network monitoring device (200) generates different global probability distributions for each RU (100), sets different deterioration criteria for each RU (100), and monitors whether or not each RU (100) is deteriorated based on the different deterioration criteria set for each RU (100).

Meanwhile, although the above-described embodiment described monitoring a 5G-based wireless network, the present invention can be applied to wireless communication networks other than 5G that utilize a base station.

In addition, in the above-described embodiment, the RU (100) is described as measuring one or more of the handover success rate, the number of RRE occurrences, the number of connected mobile terminals, the SINR of the uplink, and the RSSI, and transmitting them to the wireless network monitoring device (200), but in addition, various KPI values can be collected and transmitted to the wireless network monitoring device (200). In this case, the wireless network monitoring device (200) can estimate the global probability distribution/local probability distribution and set the deterioration reference value using various KPI values.

While this specification includes many features, such features should not be construed as limiting the scope of the invention or the claims. Furthermore, features described in separate embodiments herein may be combined and implemented in a single embodiment. Conversely, various features described in a single embodiment herein may be implemented in multiple embodiments individually or in suitable combinations.

Although the operations are illustrated in the drawings in a particular order, it should not be understood that such operations are to be performed in the particular order illustrated, or in a series of sequential orders, or that all of the illustrated operations are to be performed to achieve a desired result. Multitasking and parallel processing may be advantageous in certain circumstances. Furthermore, it should be understood that the distinction between the various system components in the embodiments described above does not require such distinction in all embodiments. The program components and systems described above may generally be implemented as a package in a single software product or in multiple software products.

The method of the present invention as described above can be implemented as a program and stored in a computer-readable form on a recording medium (CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, etc.). Since this process can be easily performed by a person having ordinary skill in the art to which the present invention belongs, it will not be described in detail any further.

The present invention described above is not limited to the above-described embodiments and the attached drawings, as various substitutions, modifications, and changes can be made without departing from the technical spirit of the present invention by a person having ordinary skill in the art to which the present invention pertains.

100 : RU 200 : Wireless Network Monitoring Device
210: Sample vector acquisition section 220: Probability distribution estimation section
230: Storage section 240: Quality monitoring section
250: Deterioration Cause Identification Section 300: gNB-DU
400 : gNB-CU

Claims (12)

A method for monitoring wireless network quality by detecting whether a base station cell is deteriorated in a wireless network monitoring device,
A step of obtaining a sample vector from each of a plurality of cell data and estimating a global probability distribution using the obtained sample vector;
A step of collecting a plurality of cell data generated from the above base station cell and obtaining a sample vector from each cell data;
A step of counting the number of deviations in which the sample vector acquired in the above acquiring step deviates from the deterioration reference value;
If the counted number of departures exceeds a threshold number set in advance, a step of estimating a local probability distribution using sample vectors acquired during a certain period of time; and
A method for monitoring wireless network quality, comprising: a step of checking the similarity between the estimated local probability distribution and the global probability distribution, and determining that quality deterioration has occurred in the base station cell if the similarity is below a threshold value. In the first paragraph,
Before the step of obtaining the above sample vector,
A method for monitoring wireless network quality, further comprising: a step of setting a sample vector corresponding to a pre-set percentage in the above global probability distribution as the deterioration reference value. In the second paragraph,
The above global probability distribution is updated to reflect the additional sample vector when an additional sample vector is obtained.
A wireless network quality monitoring method, characterized in that when the above global probability distribution is updated, the above deterioration criterion value is also changed. In the second paragraph,
The step of estimating the above local probability distribution is to estimate the above local probability distribution by applying the sample vectors obtained during the above period to machine learning,
A wireless network quality monitoring method, characterized in that the local probability distribution has a shorter sample vector collection period than the global probability distribution and the number of sample vectors used as samples is smaller than that of the global probability distribution. In the first paragraph,
After the above judging step,
A step of checking the direction and size of sample vectors that deviate from the above deterioration criterion value;
A step of confirming a deterioration area corresponding to the direction and size confirmed above; and
A method for monitoring wireless network quality, further comprising: a step of confirming a cause of deterioration corresponding to the confirmed deterioration area from deterioration cause data for each deterioration area stored in advance. In paragraph 5,
A method for monitoring wireless network quality, further comprising: a step of outputting or transmitting to a manager's terminal a fault alarm message including identification information of a cell in which deterioration has occurred and the cause of the deterioration confirmed above. In the first paragraph,
The above sample vector is,
A wireless network quality monitoring method characterized by at least one of the number of RRE (RRC Connection Re-establishment) occurrences, the number of connected terminals, the uplink SINR (Signal to Interference Noise Ratio), the handover success rate, and the RSSI (Received Signal Strength Indicator). A computer program recorded on a computer-readable recording medium, executing any one of the methods according to any one of claims 1 to 7. In a device for monitoring the quality status of a base station cell,
A sample vector acquisition unit that collects a plurality of cell data generated from the above base station cell and acquires a sample vector from each cell data;
A probability distribution estimation unit that estimates a global probability distribution using a plurality of sample vectors acquired from the above sample vector acquisition unit and selectively estimates a local probability distribution using sample vectors acquired over a certain period of time; and
A wireless network monitoring device including a quality monitoring unit that counts the number of deviations in which sample vectors acquired during the above-described period of time deviate from a deterioration reference value, estimates a local probability distribution through the above-described probability distribution estimation unit when the counted number of deviations exceeds a threshold number set in advance, and determines that quality deterioration has occurred in the above-described base station cell when the similarity is less than or equal to a threshold value. In Article 9,
The above probability distribution estimation part is,
A wireless network monitoring device characterized in that a sample vector corresponding to a pre-set percentage in the above global probability distribution is set as the deterioration reference value. In Article 9,
The above probability distribution estimation unit estimates the global probability distribution and the local probability distribution by applying the sample vectors to machine learning, respectively.
A wireless network monitoring device, characterized in that the local probability distribution has a shorter sample vector collection period than the global probability distribution and the number of sample vectors used as samples is smaller than that of the global probability distribution. In Article 9,
A storage unit that stores the cause of deterioration by deterioration area; and
A wireless network monitoring device further comprising a deterioration cause confirmation unit that confirms the direction and size of sample vectors that deviate from the above deterioration reference value, confirms a deterioration area corresponding to the confirmed direction and size, and confirms the deterioration cause corresponding to the deterioration area in the storage unit.

Images