US-12627386-B2 - Model building apparatus and model building method, root cause estimation apparatus and root cause estimation method, and recording medium

US12627386B2US 12627386 B2US12627386 B2US 12627386B2US-12627386-B2

Abstract

A model building apparatus includes: an acquisition unit that obtains, from a first wireless station, time-series data of received power of wireless communication between the first wireless station and a second wireless station; a generation unit that generates input data including change data about a degree of a change in the received power per unit time, based on the time-series data; and a building unit that builds the estimation model for estimating the root cause of degradation from the input data, by machine learning using the input data.

Inventors

Tansheng LI
Eiji Takahashi
Nobuhiko Itoh

Assignees

NEC CORPORATION

Dates

Publication Date: 20260512
Application Date: 20210317

Claims (10)

1 . A root cause estimation apparatus that estimates at least one root cause of degradation of communication quality between a first wireless station and a second wireless station, the root cause estimation apparatus comprising: at least one memory configured to store instructions; and at least one processor configured to execute the instructions to: obtain, from the first wireless station, time-series data of received power of wireless communication between the first wireless station and the second wireless station, generate input data including change data about a degree of a change in the received power per unit time, based on the time-series data, and estimate the at least one root cause of degradation and that the communication quality is not degraded, by using the input data and an estimation model that is built by machine learning and that is trained by using training data that includes: the time-series data for training, and correct labels corresponding to the time-series data for training and indicating that the at least one root cause of degradation and the communication quality is not degraded, and that is for estimating, from the input data, the at least one root cause of degradation and that the communication quality is not degraded.
2 . The root cause estimation apparatus according to claim 1 , wherein the change data include data about at least one of a coefficient of an approximate expression for approximating the received power, a variance of the received power, and a standard deviation of the received power.
3 . The root cause estimation apparatus according to claim 1 , wherein the change data include data about a classification result obtained by classifying the received power in accordance with the degree of the change in the received power per unit time.
4 . The root cause estimation apparatus according to claim 1 , wherein the received power includes received power between the first wireless station and the second wireless station in which an influence of another first wireless station that is different from the first wireless station is eliminated.
5 . The root cause estimation apparatus according to claim 1 , wherein the received power includes at least one of RSSI (Received Signal Strength Indicator), RSRP (Reference Signal Received Power), RSRQ (Reference Signal Received Quality), and SINR (Signal to Interference plus Noise Ratio).
6 . The root cause estimation apparatus according to claim 1 , wherein the at least one processor is configured to execute the instructions to generate the input data including the change data about the degree of the change in the received power per unit time, in a variable predetermined period.
7 . The root cause estimation apparatus according to claim 1 , wherein the at least one processor is configured to execute the instructions to estimate whether the at least one root cause of degradation is at least one of such a first root cause that the communication quality is degraded by distance attenuation, such a second root cause that the communication quality is degraded by shielding, and such a third root cause that the communication quality is degraded by fading.
8 . The root cause estimation apparatus according to claim 1 , wherein: the change data are velocity data about a change velocity of the received power, and the at least one processor is configured to execute the instructions toto: estimate that the at least one root cause of degradation is such a first root cause that the communication quality is degraded by distance attenuation, when the velocity data indicate that the change velocity is less than a predetermined velocity, and estimate that the at least one root cause of degradation is such a second root cause that the communication quality is degraded by shielding, when the velocity data indicate that the change velocity is greater than the predetermined velocity.
9 . A root cause estimation method of estimating a at least one root cause of degradation of communication quality between a first wireless station and a second wireless station, the root cause estimation method comprising: obtaining, from the first wireless station, time-series data of received power of wireless communication between the first wireless station and the second wireless station; generating input data including change data about a degree of a change in the received power per unit time, based on the time-series data; and estimating the at least one root cause of degradation and that the communication quality is not degraded, by using the input data and an estimation model that is built by machine learning and that is trained by using training data that includes: the time-series data for training, and correct labels corresponding to the time-series data for training and indicating that the at least one root cause of degradation and the communication quality is not degraded, and that is for estimating, from the input data, the at least one root cause of degradation and that the communication quality is not degraded.
10 . A non-transitory recording medium that records thereon a computer program that allows a computer to execute a root cause estimation method of estimating at least one root cause of degradation of communication quality between a first wireless station and a second wireless station, the root cause estimation method including: obtaining, from the first wireless station, time-series data of received power of wireless communication between the first wireless station and the second wireless station; generating input data including change data about a degree of a change in the received power per unit time, based on the time-series data; and estimating the at least one root cause of degradation and that the communication quality is not degraded, by using the input data and an estimation model that is built by machine learning and that is trained by using training data that includes: the time-series data for training, and correct labels corresponding to the time-series data for training and indicating that the at least one root cause of degradation and the communication quality is not degraded, and that is for estimating, from the input data, the at least one root cause of degradation and that the communication quality is not degraded.

Description

This application is a National Stage Entry of PCT/JP2021/010777 filed on Mar. 17, 2021, the contents of all of which are incorporated herein by reference, in their entirety. TECHNICAL FIELD This disclosure relates, for example, to technical fields of a model building apparatus, a model building method, and a recording medium that are capable of building an estimation model for estimating a root cause of degradation of communication quality between a first wireless station and a second wireless station, a model building method and a recording medium, and a root cause estimation apparatus, a root cause estimation method, and a recording medium that are capable of estimating the root cause of degradation of the communication quality between the first wireless station and the second wireless station. BACKGROUND ART Patent Literature 1 describes a quality estimation apparatus that estimates communication quality in wireless communication (specifically, mobile communication), by using a neural network model. In addition, as prior art documents related to this disclosure, Patent Literature 2 to Patent Literature 4 are cited. CITATION LIST Patent Literature Patent Literature 1: Japanese Patent No. 6696859Patent Literature 2: International Publication No. WO2020/217457 pamphletPatent Literature 3: JP2017-123124APatent Literature 4: JP2008-278148A SUMMARY Technical Problem When the communication quality is degraded, it is desirable to take measures to restore the degraded communication quality. Here, the measures to restore the degraded communication quality vary in many cases depending on a root cause of the degradation of the communication quality. The quality estimation apparatus described in Patent Literature 1, however, has a technical problem that it is not capable of estimating the root cause of the degradation of the communication quality. It is an example object of this disclosure to provide a model building apparatus and a model building method, a root cause estimation apparatus and a root cause estimation method, and a recording medium that are capable of solving the above-described technical problem. As an example, it is an example object of this disclosure to provide a model building apparatus, a model building method, and a recording medium that are capable of building an estimation model for estimating a root cause of degradation of communication quality between a first wireless station and a second wireless station, a model building method and a recording medium, and a root cause estimation apparatus, a root cause estimation method, and a recording medium that are capable of estimating the root cause of degradation of the communication quality between the first wireless station and the second wireless station. Solution to Problem A model building apparatus according to an example aspect is a model building apparatus that builds an estimation model for estimating a root cause of degradation of communication quality between a first wireless station and a second wireless station, the model building apparatus including: an acquisition unit that obtains, from the first wireless station, time-series data of received power of wireless communication between the first wireless station and the second wireless station; a generation unit that generates input data including change data about a degree of a change in the received power per unit time, based on the time-series data; and a building unit that builds the estimation model for estimating the root cause of degradation from the input data, by machine learning using the input data. A root cause estimation apparatus according to an example aspect is a root cause estimation apparatus that estimates a root cause of degradation of communication quality between a first wireless station and a second wireless station, the root cause estimation apparatus including: an acquisition unit that obtains, from the first wireless station, time-series data of received power of wireless communication between the first wireless station and the second wireless station; a generation unit that generates input data including change data about a degree of a change in the received power per unit time, based on the time-series data; and an estimation unit that estimates the root cause of degradation, by using the input data and an estimation model that is built by machine learning and that is for estimating the root cause of degradation from the input data. A model building method according to an example aspect is a model building method of building an estimation model for estimating a root cause of degradation of communication quality between a first wireless station and a second wireless station, the model building method including: obtaining, from the first wireless station, time-series data of received power of wireless communication between the first wireless station and the second wireless station; generating input data including change data about a degree of a change in the rec