Precipitation amount estimation device, learning device, trained model, precipitation amount prediction device, base station, service provision device, and program

Abstract

The present invention provides a precipitation amount estimation device with which it is possible to instantaneously estimate the precipitation amount at a discretionary location within a range covered by a base station. This precipitation amount estimation device comprises: an acquisition unit that acquires, for each of a plurality of base stations of a mobile communication system, data relating to attenuation of radio waves transmitted and received between the base station and a plurality of terminal devices; and an estimation unit that estimates the precipitation amount on the basis of position information pertaining to base station antennas and / or the terminal devices and the data relating to attenuation of radio waves. The radio waves may be radio waves in a plurality of mutually different frequency bands, the radio waves being transmitted from the base station antennas. The data relating to attenuation of radio waves may be data pertaining to the attenuation rate, the attenuation amount, or the attenuation coefficient of the radio waves, the data being calculated on the basis of the transmission strength of radio waves transmitted from the terminal devices and the reception strength of the radio waves received by the base station antennas. The estimation unit may output the precipitation amount as objective-variable data by using a trained model. The precipitation amount estimation device may be mounted on a RAN Intelligent Controller (RIC).

Description

Precipitation estimation device, learning device, trained model, precipitation forecasting device, base station, service providing device and program

[0001] The present invention relates to estimating precipitation such as rainfall and snowfall.

[0002] Conventionally, methods and devices for estimating precipitation using radio waves transmitted from a satellite have been known. For example, a method for estimating precipitation using a microwave radar mounted on a satellite has been known. Patent Document 1 discloses a rainfall estimation device including: a receiving unit that acquires the reception strength of each of multiple radio waves in different frequency bands transmitted from the satellite and measures an attenuation rate from the transmission strength of each of the multiple radio waves; a feature extraction unit that extracts features from the measured attenuation rates of each of the multiple radio waves; and an estimation unit that uses the features as input and outputs an estimated value of rainfall using a machine learning model that has been trained using past attenuation rates and rainfall amounts as training data.

[0003] Japanese Patent Application Laid-Open No. 2023-021617

[0004] Conventional methods and devices for estimating precipitation have the problem that the area and time interval for measuring radio waves from a satellite used to estimate precipitation depend on the satellite's orbit and orbital period, making it impossible to instantly estimate precipitation at any point over a wide area.

[0005] According to one aspect of the present invention, there is provided a precipitation estimation device for estimating precipitation, the precipitation estimation device including: an acquisition unit configured to acquire, for each of one or more base stations in a mobile communication system, data relating to attenuation of radio waves transmitted between a base station antenna and one or more terminal devices; and an estimation unit configured to estimate precipitation based on location information of at least one of the base station antenna and the terminal devices and the data relating to attenuation of radio waves.

[0006] In the precipitation estimation device, radio waves transmitted and received between the base station antenna and the terminal device may be radio waves in a plurality of different frequency bands.

[0007] In the precipitation estimation device, the data regarding radio wave attenuation may be data on an attenuation rate, an attenuation amount, or an attenuation coefficient of the radio wave calculated based on the transmission strength or transmission power of the radio wave transmitted from the terminal device and the reception strength or reception power of the radio wave received by the base station antenna. Here, the radio wave transmitted from the terminal device and received by the base station antenna may be radio wave when an SRS (Sounding Reference Signal) is transmitted and received.

[0008] In the precipitation estimation device, the data regarding the attenuation of radio waves may be data on the attenuation rate, attenuation amount, or attenuation coefficient calculated based on the transmission strength or transmission power of the radio waves transmitted from the base station antenna and the reception strength or reception power of the radio waves received by the terminal device.

[0009] The precipitation estimation device may include a feature extraction unit that extracts feature values ​​from the data related to radio wave attenuation, and the estimation unit may output the precipitation as data of a dependent variable using a trained model to which location information of at least one of the base station antenna and the terminal device and the feature values ​​are input as data of explanatory variables, and the trained model may be a trained model that has been machine-learned using a plurality of sets of training data, each set including data of location information of at least one of the base station antenna and the terminal device that correspond to each other in the past, the feature values, and data of actual precipitation.Here, the precipitation estimation device may further include a learning unit that performs machine learning using a plurality of sets of training data, each set including data of location information of at least one of the base station antenna and the terminal device that correspond to each other in the past, the feature values, and data of actual precipitation, to generate a trained model to which location information of at least one of the base station antenna and the terminal device and the feature values ​​are input as data of explanatory variables and to which data of the precipitation amount is output as data of a dependent variable.

[0010] The precipitation estimation device may be provided in a base station of a mobile communication system, or in a radio access network (RAN) component device (e.g., a device constituting a substation of the RAN, or a device constituting a master station of the RAN) that constitutes a radio access network (RAN) including multiple base stations of the mobile communication system.

[0011] According to yet another aspect of the present invention, there is provided a learning device for generating a trained model, the learning device including: a precipitation data storage unit storing data on actual precipitation amounts in the past; and a learning unit configured to generate a trained model by machine learning using a plurality of sets of training data, each set including, for each of one or more base stations in a mobile communication system, features extracted from data on attenuation of radio waves transmitted and received between corresponding base station antennas and one or more terminal devices in the past, location information of at least one of the base station antennas and the terminal devices, and the actual precipitation data.

[0012] The learning device may further include an acquisition unit that acquires, for each of one or more base stations in the mobile communication system, data regarding the attenuation of radio waves transmitted and received between the base station antenna and one or more terminal devices, and a feature extraction unit that extracts features from the data regarding the attenuation of radio waves.

[0013] The learning device may be provided in a base station of a mobile communication system, or in a radio access network (RAN) component device (for example, a device constituting a substation of the RAN, or a device constituting a master station of the RAN) that constitutes a radio access network (RAN) including multiple base stations of the mobile communication system, or in an external network.

[0014] A base station of a mobile communication system according to yet another aspect of the present invention includes any one of the precipitation estimation devices or any one of the learning devices.

[0015] In yet another aspect of the present invention, a radio access network (RAN) component device (e.g., a device constituting a substation of the RAN, or a device constituting a master station of the RAN) constituting a radio access network (RAN) including multiple base stations of a mobile communication system is equipped with any of the precipitation estimation devices or any of the learning devices.

[0016] According to yet another aspect of the present invention, there is provided a trained model for causing a computer or processor to function to estimate precipitation. The trained model is generated by machine learning using a plurality of sets of training data, each set including, for one or more base stations in a mobile communication system, feature values ​​extracted from data relating to attenuation of radio waves transmitted and received between corresponding base station antennas and one or more terminal devices in the past, location information of at least one of the base station antennas and the terminal devices, and precipitation data. When the location information of at least one of the base station antennas and the terminal devices and the feature values ​​are input as explanatory variable data, the trained model outputs the precipitation data as objective variable data.

[0017] A further aspect of the present invention is a precipitation prediction device for predicting precipitation, comprising a prediction unit for predicting future precipitation for each region based on the regional correlation and time-series variation of the precipitation data estimated by any one of the precipitation estimation devices and on precipitation data actually measured in the past and present.

[0018] According to yet another aspect of the present invention, there is provided a service providing device for providing precipitation data, the service providing device including: a data aggregating unit that aggregates the precipitation data estimated by any one of the precipitation estimation devices; and a data providing unit that provides precipitation data for each region based on the precipitation data accumulated by the data aggregating unit.

[0019] According to yet another aspect of the present invention, there is provided a service providing device for providing precipitation data, the service providing device including a data providing unit for providing forecast data of the precipitation predicted by the precipitation forecasting device.

[0020] A program according to yet another aspect of the present invention is a program for causing a computer or a processor to function as any one of the precipitation estimation devices described above.

[0021] A program according to yet another aspect of the present invention is a program for causing a computer or a processor to function as any one of the learning devices described above.

[0022] A program according to yet another aspect of the present invention is a program for causing a computer or a processor to function as the precipitation prediction device.

[0023] The feature used to estimate precipitation using the trained model and to generate the trained model may be a vector whose elements are data related to the attenuation of the radio waves (e.g., data on the attenuation rate, attenuation amount, or attenuation coefficient of the radio waves) and whose direction corresponds to the propagation direction of the radio waves.

[0024] The precipitation estimation device, the learning device, and the precipitation prediction device may be mounted on or connected to a RAN Intelligent Controller (RIC).

[0025] According to the present invention, it is possible to instantly estimate the amount of precipitation at any point within the area covered by a base station.

[0026] Fig. 1 is a schematic diagram showing an example of the overall configuration of a system according to an embodiment. Fig. 2 is an explanatory diagram showing an example of the configuration of a radio access network (RAN) in which a precipitation estimation device according to an embodiment can be installed.

[0027] Hereinafter, embodiments of the present invention will be described with reference to the drawings. An example of a system according to the embodiment described herein includes a precipitation estimation device that calculates precipitation rates (e.g., rainfall rates) [mm / h] at or around tens of thousands of base stations nationwide in real time using a machine-learned model created using numerical values ​​related to attenuation of SRS (Sounding Reference Signals) transmitted to each base station from a terminal device, such as a mobile phone, in a mobile communication system, the location of the base station, and the location of the terminal device as causal variables, and past actual precipitation rates (e.g., rainfall rates) [mm / h] as training data. The system also includes a service provider that provides the calculated precipitation data as a service. This system enables instantaneous precipitation estimation at any point within the coverage area of ​​the base station. In particular, the system according to the present embodiment can be applied to a RAN Intelligent Controller (RIC).

[0028] As mentioned above, current precipitation is measured or estimated based on precipitation radar data from satellites, for example. However, the target area and time interval for precipitation measurement or estimation depend on the satellite's orbit and orbital period, making it impossible to determine the amount of precipitation (rainfall) at any given time or location. For example, precipitation radar satellites such as GPM measure precipitation every 95 minutes. This has the drawback of making it difficult to grasp the real-time situation of sudden events such as downbursts. Furthermore, 3D observation of cumulonimbus clouds using weather radar requires measurements every 5 minutes, even for localized areas. As such, precipitation radar satellites and weather radars lack the immediacy to respond to the onset of rain.

[0029] The system of this embodiment uses information about the attenuation of radio waves in wireless communications between a large number of base station antennas distributed across tens of thousands to hundreds of thousands of locations across the country and terminal devices, making it possible to instantly estimate the amount of precipitation at the locations where the base stations are located, the locations where the terminal devices are located, or the locations in their surrounding areas.

[0030] Fig. 1 is a schematic diagram showing an example of the overall configuration of a system including a precipitation estimation device according to an embodiment. In Fig. 1, the system of this embodiment utilizes data related to the attenuation of radio waves transmitted and received between antennas 11 of one or more base stations 10 (hereinafter also referred to as "base station antennas") in a mobile communication system and one or more terminal devices 20. In particular, this embodiment utilizes information related to the attenuation of radio waves in wireless communication between the terminal devices 20 and the base station antennas 11 of a large number of base stations 10 distributed throughout the country.

[0031] In this embodiment, the base station 10 is suitably a small cell base station that forms a small cell with a radius of approximately several meters to several hundred meters, but may also be a macro cell base station that forms a macro cell with a radius of approximately several hundred meters to several kilometers. The base station 10 may also be a base station called an eNodeB, gNodeB, or the like, depending on the generation of the specifications of the applied mobile communication system. Furthermore, the cell primarily formed by the base station 10 may be a two-dimensional cell or a three-dimensional cell. Furthermore, the terminal device 20 that wirelessly communicates with the base station 10 may be located on the ground, on the sea, or in the sky.

[0032] In this embodiment, the precipitation estimation device 30 has an acquisition unit 310 that acquires data regarding the attenuation of radio waves transmitted and received between the base station antenna 11 and multiple terminal devices 20 for each of multiple base stations 10 in the mobile communication system, and a precipitation estimation unit 330 that estimates the amount of precipitation based on the location information of at least one of the base station antenna 11 and the terminal devices 20 and the data regarding the attenuation of radio waves.

[0033] The location information used to estimate the amount of precipitation may be only the location information of the base station antenna 11, may be only the location information of the terminal device 20, or may be the location information of both the base station antenna 11 and the terminal device 20. The location information is, for example, latitude and longitude information. The location information may also include height information (for example, information on the altitude from a reference earth surface). For example, the location information of the base station antenna 11 and the terminal device 20 may be location information acquired by a receiver of a GSNN (Global Navigation Satellite System) such as a GPS (Global Positioning System).

[0034] The radio waves transmitted and received between the base station antenna 11 and the multiple terminal devices 20 may be radio waves in a single frequency band or radio waves in multiple different frequency bands. The frequency band may be, for example, a microwave band of 300 MHz to 30 GHz or a millimeter wave band higher than 30 GHz. The frequency band may also be a terahertz band (100 GHz to 10 THz or 100 GHz to 400 GHz).

[0035] 1 , the data on radio wave attenuation is data on the attenuation rate, amount of attenuation, or attenuation coefficient of radio waves calculated based on the transmission strength or transmission power of radio waves including an uplink SRS (Sounding Reference Signal) transmitted from the terminal device 20 and the reception strength or reception power of the radio waves received by the base station antenna 11. Here, the radio waves including the SRS (Sounding Reference Signal) are transmitted, for example, from the terminal device 20 periodically or aperiodically using predetermined radio resources.

[0036] In addition, the data regarding the attenuation of radio waves may also be data of the attenuation rate, attenuation amount, or attenuation coefficient calculated based on the transmission strength or transmission power of the downlink radio waves transmitted from the base station antenna 11 and the reception strength or reception power of the radio waves received by the terminal device 20.

[0037] 1 estimates precipitation using a trained model, and therefore includes a feature extraction unit 320 that extracts features used in the trained model. The feature extracted by the feature extraction unit 320 is, for example, a vector whose elements are data related to the attenuation of the radio waves (for example, data on the attenuation rate, attenuation amount, or attenuation coefficient of the radio waves) and whose direction corresponds to the propagation direction of the radio waves.

[0038] The trained model is, for example, a machine-learned model generated by machine learning using multiple sets of training data, each set including, for each of one or more base stations 10, feature values ​​extracted from data related to the attenuation of radio waves transmitted and received between corresponding base station antennas 11 and one or more terminal devices 20 in the past, location information of at least one of the base station antennas 11 and the terminal devices 20, and actual precipitation data (training teacher data). When the location information of at least one of the base station antennas 11 and the terminal devices 20 and the feature values ​​are input as explanatory variable (also referred to as "cause variable") data, the trained model outputs precipitation data as objective variable (also referred to as "result variable"). In this case, the precipitation estimation unit 330 outputs precipitation as objective variable data using the trained model to which the location information of at least one of the base station antennas 11 and the terminal devices 20 and the feature values ​​are input as explanatory variable data.

[0039] Furthermore, the trained model may be, for example, a machine-learned model generated by machine learning using multiple sets of training data (supervised learning data) for each of one or more base stations 10, each set including data on attenuation of radio waves transmitted and received between corresponding base station antennas 11 and one or more terminal devices 20 in the past, location information of at least one of the base station antennas 11 and the terminal devices 20, and actual precipitation data (teacher data). This trained model also outputs precipitation data as response variable data when the location information of at least one of the base station antennas 11 and the terminal devices 20 and the data on radio wave attenuation are input as explanatory variable data. In this case, the precipitation estimation unit 330 outputs precipitation as response variable data using the trained model to which the location information of at least one of the base station antennas 11 and the terminal devices 20 and the data on radio wave attenuation are input as explanatory variable data.

[0040] The algorithm used in the trained model of this embodiment is not limited to a specific algorithm. For example, as an algorithm of the machine-learned model that learns using training data (supervised learning data), SVR (Support Vector Regression), which is classified as "Regression" that learns numerical data and predicts numerical values, can be used. Instead of SVR, Linear (Ordinary) Regression, Bayesian Linear Regression, Random (Decision) Forest, Boosted Decision Tree, Fast Forest Quantile, Neural Network, Poisson Regression, Support Vector Ordinal Regression, Ridge Regression, Lasso Regression, etc. may also be used.

[0041] 1, the system of this embodiment includes a learning device 40 that generates the trained model. The learning device 40 includes a learning unit 410, a precipitation data storage unit 420, and a trained model storage unit 430. The learning unit 410 performs machine learning to generate a trained model for each of one or more base stations 10 in the mobile communication system using multiple sets of training data, each set including feature amounts extracted from data related to attenuation of radio waves transmitted and received between corresponding base station antennas 11 and one or more terminal devices 20 in the past, location information of at least one of the base station antennas 11 and the terminal devices, and actual precipitation data.

[0042] In the learning unit 410, the corresponding features from the past and the location information of at least one of the base station antenna 11 and the terminal device used to generate the learned model 435 can be obtained, for example, from the data accumulation unit 810 of the service providing device 80 described below.

[0043] The precipitation data storage unit 420 stores data on actual precipitation amounts acquired in the past from precipitation radar satellites, weather radars, and the like.

[0044] The learning unit 410 acquires, from the precipitation data storage unit 420, data on actual precipitation amounts corresponding to the feature amounts from the past and the location information of at least one of the base station antenna 11 and the terminal device.

[0045] The learning unit 410 stores information about the trained model 435 in the trained model storage unit 430. The information about the trained model 435 is, for example, parameter values ​​such as weights for feature quantities in a model formed by a predetermined algorithm used to estimate precipitation.

[0046] The learning device 40 may further include an acquisition unit that acquires data regarding the attenuation of radio waves transmitted and received between the base station antenna 11 and one or more terminal devices 20 for each of one or more base stations 10 in the mobile communication system, and a feature extraction unit that extracts features from the data regarding the attenuation of radio waves.

[0047] The learning unit 410 may generate a trained model without using the feature quantities. For example, the learning unit 410 may perform machine learning using multiple sets of training data, each set including data on attenuation of radio waves transmitted and received in the past between the base station antenna 11 and one or more terminal devices 20, location information of at least one of the base station antenna 11 and the terminal device 20, and data on actual precipitation, for each of one or more base stations 10 in the mobile communication system.

[0048] In this case, the data regarding the corresponding radio wave attenuation in the past and the location information of at least one of the base station antenna 11 and the terminal device used to generate the trained model can be obtained from the data accumulation unit 810 of the service providing device 80 described below.

[0049] The learning unit 410 also acquires from the precipitation data storage unit 420 data relating to past radio wave attenuation and data on actual precipitation corresponding to the location information of at least one of the base station antenna 11 and the terminal device.

[0050] Also in this case, the learning unit 410 stores information about the trained model in the trained model storage unit 430. The information about the trained model is, for example, parameter values ​​such as weights for feature quantities in a model formed by a predetermined algorithm used to estimate precipitation.

[0051] As shown by the two-dot chain line in FIG. 1, the precipitation estimation device 30' may be configured to incorporate the functions of the learning device 40 described above.

[0052] The estimated precipitation value estimated and output by the precipitation estimation unit 330 of the precipitation estimation device 30 (30') is sent to the data collection unit 810 of the service providing device 80, for example, via the core network 50 and an external network 60 such as the Internet.

[0053] 1, the system of this embodiment includes a precipitation prediction device 70 that predicts future precipitation. The precipitation prediction device 70 includes a prediction unit 710 that predicts future precipitation for each region based on the regional correlation and time-series changes in precipitation data estimated by the precipitation estimation device 30 (30') and precipitation data actually measured in the past and present. The predicted precipitation data output from the prediction unit 710 is sent to and accumulated in a data accumulation unit 810 of the service providing device 80.

[0054] 1, the system of this embodiment also includes a service providing device 80 that provides precipitation data. The service providing device 80 may be configured as a server consisting of a single computer device, or may be configured as a cloud system consisting of multiple computers that are arranged to cooperate with each other on a network.

[0055] The service providing device 80 includes a data aggregating unit 810 that aggregates data on precipitation estimated by the precipitation estimation device 30 (30'), and a data providing unit 820 that provides precipitation data for each region based on the precipitation data accumulated in the data aggregating unit 810. For example, in response to a request received from the terminal device 20 or another user device, the data providing unit 820 refers to the data aggregating unit 810 and transmits data on precipitation [mm / h] at a specific point and time to the terminal device 20 or another user device.

[0056] In the service providing device 80, the data collection unit 810 accumulates forecast data on precipitation predicted by the precipitation prediction device 70, and the data providing unit 820, for example, in response to a request received from the terminal device 20 or another user device, refers to the data collection unit 810 and transmits and provides forecast data on precipitation [mm / h] at a specific location and time to the terminal device 20 or another user device.

[0057] 2 is an explanatory diagram showing an example of the configuration of a radio access network (RAN) 100 that can be equipped with a precipitation estimation device 30 according to an embodiment. In FIG. 2, the RAN 100 includes a plurality of sets of base station antennas 11 and RUs (Radio Units) 110, a DU (Distributed Unit) 120 to which the plurality of RUs 110 are connected, and a CU (Central Unit) 130 connected to the DU 120. The RU 110 has a function of controlling the base station antenna 11 to transmit and receive radio waves to and from a terminal device 20, and also controls MIMO, beamforming, and the like. The DU 120 performs signal modulation, demodulation, encoding, decoding, and MAC layer communication control. The CU 130 has functions such as controlling the subordinate DUs 120 and RUs 110, connecting to the core network 50, and processing PDCP (Packet Data Convergence Protocol) for encrypting packets, and RRC (Radio Resource Control) for managing radio resources of the terminal device 20.

[0058] The base station 10 of this embodiment is configured, for example, with a slave station including a base station antenna 11 and an RU 110, and a master station including a DU 120 and a CU 130. The slave station RU 110 and the master station DU 120 are connected by a fronthaul network made up of optical fiber or the like. The CU 130 is connected to the core network 50 via a backhaul line.

[0059] 1, the precipitation estimation device 30 according to this embodiment can be provided in, for example, the base station 10, thereby making effective use of the surplus computational resources of the base station 10. The precipitation estimation device 30 may also be provided in, for example, the RU 110, DU 120, or CU 130, which are RAN component devices, thereby making effective use of the surplus computational resources of the RAN component devices.

[0060] 1, the learning device 40 according to this embodiment may be provided in, for example, the base station 10, to effectively utilize the surplus computational resources of the base station 10. The learning device 40 may also be provided in, for example, the RU 110, DU 120, or CU 130, which are RAN component devices, to effectively utilize the surplus computational resources of the RAN component devices.

[0061] As described above, according to this embodiment, by utilizing multiple base station antennas distributed at different locations in a mobile communication system, it is possible to instantly estimate and forecast precipitation at any point within the area covered by the base station. Moreover, since the surplus computational resources of the base station can be utilized for precipitation estimation or forecasting and for generating trained models, it is possible to reduce investment for precipitation estimation and forecasting and for services that provide information on estimated or forecasted precipitation.

[0062] Furthermore, according to this embodiment, the base stations 10 that are widely and abundantly present throughout the country can be used as observation points for the attenuation of the radio waves.

[0063] Furthermore, according to this embodiment, investment for services can be reduced by utilizing surplus computational resources in the base stations 10 and other RAN component devices in the RAN 100.

[0064] Furthermore, according to this embodiment, it is possible to estimate the amount of precipitation, such as rainfall or snowfall, that is not affected by the measurement period of the rainfall observation satellite.

[0065] Furthermore, according to this embodiment, by using data on the attenuation of radio waves such as SRS traveling from the terminal device 20 to the base station antenna 11 periodically or aperiodically at predetermined time intervals, the immediacy and accuracy of the precipitation estimation data can be improved, and the update frequency of the trained model used to estimate precipitation can be increased.

[0066] The present invention can provide devices and systems that can instantly estimate precipitation amounts, such as rainfall and snowfall, at any point within the coverage area of ​​a base station, thereby contributing to the achievement of Goal 9 of the Sustainable Development Goals (SDGs), which is to "build resilient infrastructure, promote inclusive and sustainable industrialization, and promote innovation and resilience."

[0067] The processing steps and components of the base station, terminal device, precipitation estimation device, learning device, precipitation prediction device, service providing device, etc. described in this specification can be implemented by various means. For example, these steps and components may be implemented by hardware, firmware, software, or a combination thereof.

[0068] For hardware implementation, means such as processing units used to realize the above steps and components in an entity (e.g., a base station, a base station device, a terminal device (UE: user equipment, mobile station, communication terminal), a RAN slave station, a master station, an RU, a DU, a CU, a server, a core network device, a hard disk drive device, or an optical disk drive device) may be implemented in one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, electronic devices, other electronic units designed to perform the functions described herein, computers, or combinations thereof.

[0069] Furthermore, with regard to firmware and / or software implementations, the means, such as a processing unit, used to realize the components may be implemented with a program (e.g., code, such as procedures, functions, modules, instructions, etc.) that performs the functions described herein. In general, any computer / processor-readable medium tangibly embodying firmware and / or software code may be used to implement the means, such as a processing unit, used to realize the steps and components described herein. For example, the firmware and / or software code may be stored in a memory and executed by a computer or processor, such as in a controller. The memory may be implemented within the computer or processor, or external to the processor. The firmware and / or software code may also be stored on a computer or processor readable medium such as, for example, random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), flash memory, floppy disk, compact disk (CD), digital versatile disk (DVD), magnetic or optical data storage device, etc. The code may be executed by one or more computers or processors and may cause the computers or processors to perform certain aspects of the functionality described herein.

[0070] The medium may be a non-transitory recording medium. The program code may be in any format as long as it can be read and executed by a computer, processor, or other device or machine. For example, the program code may be in any of source code, object code, and binary code, or may be a mixture of two or more of these codes.

[0071] Moreover, the description of the embodiments disclosed herein is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to the present disclosure will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of the present disclosure. Thus, the present disclosure is not intended to be limited to the examples and designs described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0072] 10: Base station 11: Base station antenna 20: Terminal device 30, 30': Precipitation estimation device 310: Acquisition unit 320: Feature extraction unit 330: Precipitation estimation unit 40: Learning device 410: Learning unit 420: Precipitation data storage unit 430: Trained model storage unit 50: Mobile communication network (core network) 60: External network such as the Internet 70: Precipitation prediction device 710: Precipitation prediction unit 80: Service providing device 810: Data collection unit 820: Data providing unit 100: Radio access network (RAN)

Claims

1. A precipitation estimation device that estimates precipitation, comprising: an acquisition unit that acquires, for each of one or more base stations in a mobile communication system, data relating to attenuation of radio waves transmitted and received between a base station antenna and one or more terminal devices; and an estimation unit that estimates precipitation based on location information of at least one of the base station antenna and the terminal device and the data relating to attenuation of the radio waves.

2. A precipitation estimation device according to claim 1, characterized in that the radio waves transmitted and received between the base station antenna and the terminal device are radio waves in a plurality of different frequency bands.

3. A precipitation estimation device according to claim 1, characterized in that the data relating to the attenuation of radio waves is data on the attenuation rate, amount of attenuation or attenuation coefficient of the radio waves calculated based on the transmission strength or transmission power of the radio waves transmitted from the terminal device and the reception strength or reception power of the radio waves received by the base station antenna.

4. A precipitation estimation device according to claim 1, characterized in that the data relating to radio wave attenuation is data on the attenuation rate, amount of attenuation or attenuation coefficient calculated based on the transmission strength or transmission power of the radio waves transmitted from the base station antenna and the reception strength or reception power of the radio waves received by the terminal device.

5. A precipitation estimation device according to claim 1, further comprising a feature extraction unit that extracts features from the data relating to radio wave attenuation, wherein the estimation unit outputs the precipitation as data for the objective variable using a trained model to which location information of at least one of the base station antenna and the terminal device and the features are input as data for the explanatory variables, and the trained model is a trained model that has been machine-learned using multiple sets of training data, each set consisting of corresponding past location information of at least one of the base station antenna and the terminal device, the features, and actual precipitation data.

6. The precipitation estimation device of claim 5, further comprising a learning unit that performs machine learning using multiple sets of training data, each set consisting of past location information of at least one of the base station antenna and the terminal device corresponding to each other, the feature, and actual precipitation data, and generates a trained model in which the location information and feature of at least one of the base station antenna and the terminal device are input as explanatory variable data and the precipitation data is output as target variable data.

7. A precipitation estimation device according to any one of claims 1 to 6, characterized in that the precipitation estimation device is provided in a base station of a mobile communication system, or in a radio access network (RAN) component device that constitutes a RAN including a plurality of base stations.

8. A learning device that generates a trained model, comprising: a precipitation data storage unit that stores data on actual precipitation amounts in the past; and a learning unit that generates a trained model by machine learning using multiple sets of training data, each set including, for each of one or more base stations in a mobile communication system, features extracted from data on attenuation of radio waves transmitted and received between corresponding base station antennas and one or more terminal devices in the past, location information of at least one of the base station antennas and the terminal devices, and the actual precipitation data.

9. A learning device according to claim 8, comprising: an acquisition unit that acquires, for each of one or more base stations in a mobile communication system, data relating to the attenuation of radio waves transmitted and received between a base station antenna and one or more terminal devices; and a feature extraction unit that extracts features from the data relating to the attenuation of radio waves.

10. A base station of a mobile communication system, comprising a precipitation estimation device according to any one of claims 1 to 6 or a learning device according to any one of claims 8 and 9.

11. A trained model for causing a computer or processor to function to estimate precipitation, the trained model being generated by machine learning using multiple sets of training data, each set consisting of features extracted from data relating to the attenuation of radio waves transmitted and received between corresponding base station antennas and one or more terminal devices in the past, location information of at least one of the base station antennas and the terminal devices, and precipitation data, for each of one or more base stations in a mobile communication system, and characterized in that when the location information of at least one of the base station antennas and the terminal devices and the features are input as explanatory variable data, the trained model outputs the precipitation data as target variable data.

12. A precipitation forecasting device for forecasting precipitation, characterized in that it has a prediction unit that predicts future precipitation for each region based on the regional correlation and time series changes in the precipitation data estimated by the precipitation estimation device of any one of claims 1 to 6 and precipitation data actually measured in the past and present.

13. A service providing device for providing precipitation data, comprising: a data collection unit that collects the precipitation data estimated by a precipitation estimation device according to any one of claims 1 to 6; and a data providing unit that provides precipitation data for each region based on the precipitation data collected in the data collection unit.

14. A service providing device that provides precipitation forecast data, characterized in that it comprises a data providing unit that provides the precipitation forecast data predicted by the precipitation forecasting device of claim 12.

15. A program for causing a computer or processor to function as the precipitation estimation device according to any one of claims 1 to 6.

16. A program for causing a computer or processor to function as the learning device of any of claims 8 and 9.

17. A program for causing a computer or processor to function as the precipitation forecasting device of claim 12.

Images