Wireless quality prediction system, wireless quality prediction device, wireless quality prediction method, and program

Abstract

In order to improve the prediction accuracy of wireless quality in a wireless quality prediction system for predicting the wireless communication quality of a communication terminal, this wireless quality prediction system comprises: an acquisition unit that acquires position information indicating the position of a communication terminal and information relating to a base station to which the communication terminal is connected; and a prediction unit that predicts the wireless quality of the communication terminal using a learning device corresponding to the information relating to the base station among a plurality of learning devices trained in advance to predict the wireless quality of the communication terminal on the basis of the position information.

Description

Wireless quality prediction system, wireless quality prediction device, wireless quality prediction method, and program

[0001] The present invention relates to a wireless quality prediction system, a wireless quality prediction device, a wireless quality prediction method, and a program.

[0002] For the stable use of wireless communication, there are wireless quality prediction techniques that predict wireless communication quality. For example, there are known techniques that use previously measured performance data and machine learning using a neural network to predict wireless communication quality at a current location or a future location from location information and environmental information (see, for example, Non-Patent Documents 1 and 2).

[0003] Naoki Shibuya et al., "A Study on Wireless Communication Quality Prediction Considering Time Variation," IEICE University, B-6-28, Sep, 2022 Naoki Shibuya et al., "A Study on Wireless Communication Quality Prediction Using Machine Learning," IEICE University, B-6-74, Mar, 2022

[0004] However, in conventional techniques, learning is performed from past performance for each communication terminal, and for example, when the base station to which the communication terminal is connected changes, the accuracy of predicting wireless quality may deteriorate.

[0005] The embodiments of the present invention have been made in view of the above-mentioned problems, and improve the prediction accuracy of wireless quality in a wireless quality prediction system that predicts the wireless quality of a communication terminal.

[0006] In order to solve the above problem, a wireless quality prediction system according to an embodiment of the present invention includes an acquisition unit that acquires location information indicating the location of a communication terminal and information about a base station to which the communication terminal is connected, and a prediction unit that predicts the wireless quality of the communication terminal using a learning unit corresponding to the information about the base station, among a plurality of learning units that have been trained in advance to predict the wireless quality of the communication terminal based on the location information.

[0007] According to an embodiment of the present invention, it is possible to improve the accuracy of predicting wireless quality in a wireless quality prediction system that predicts the wireless quality of a communication terminal.

[0008] FIG. 1 is a diagram illustrating an example of a system configuration of a wireless quality prediction system according to the present embodiment. FIG. 2 is a diagram for explaining an overview of processing according to Example 1. FIG. 3 is a flowchart illustrating an example of processing during learning according to Example 1. FIG. 4 is a diagram illustrating an image of a table T according to Example 1. FIG. 5 is a flowchart illustrating an example of processing during prediction according to Example 1. FIG. 6 is a diagram illustrating the creation of different learning devices according to Example 2. FIG. 7 is a diagram illustrating an example of a hardware configuration of a computer according to the present embodiment. FIG. 8 is a diagram illustrating an example of a hardware configuration of a communication terminal according to the present embodiment. FIG. 9 is a diagram illustrating an example of a wireless quality prediction process. FIG. 10 is a diagram illustrating problems with the prior art.

[0009] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The embodiment described below is merely an example, and the embodiment to which the present invention is applied is not limited to the following embodiment.

[0010] <Regarding Wireless Quality Prediction Technology> Before describing the system configuration of the wireless quality prediction system according to this embodiment, an example of a conventional wireless quality prediction technology will be described. The wireless quality prediction technology is a technology for predicting the wireless quality of a communication terminal (wireless terminal) that performs wireless communication with a base station (wireless base station), for example, in order to ensure stable use of wireless communication.

[0011] Fig. 9 is a diagram for explaining an example of a wireless quality prediction process. This diagram shows an overview of the wireless quality prediction process disclosed in Non-Patent Document 2. In the example of Fig. 9, a wireless quality prediction server 901 acquires performance data, such as location information, radio wave intensity, and throughput, from multiple communication terminals 902 (step S01). Furthermore, the wireless quality prediction server 901 uses the acquired performance data to perform a learning process to train a neural network 904 by machine learning so that wireless qualities, such as radio wave intensity and throughput, of the communication terminals 902 and 903 can be predicted based on, for example, the location information of the communication terminals 902 and 903 (step S02).

[0012] After the wireless quality prediction server 901 has completed the learning process, it receives an inquiry about a predicted value of wireless quality from, for example, a communication terminal 903 having a mobile function such as a vehicle (step S03). In response, the wireless quality prediction server 901 inputs the location information of the communication terminal 903 included in the wireless quality inquiry into the trained neural network 904, thereby executing an inference process (prediction process) to predict wireless qualities such as radio wave intensity and throughput of the communication terminal 903 (step S04). The wireless quality prediction server 901 also notifies the communication terminal 903 of the predicted value of wireless quality of the communication terminal 903 (step S05). This allows the communication terminal 903 to more appropriately perform communication control, such as network switching, based on the notified predicted value of wireless quality.

[0013] By the above process, for example, when the communication terminal 903 wants to know the wireless quality after a predetermined time has elapsed (n seconds), the communication terminal 903 estimates the position of the communication terminal 903 after the predetermined time has elapsed, and transmits a wireless quality inquiry including the estimated position of the communication terminal 903 to the wireless quality prediction server 901. This allows the communication terminal 903 to acquire wireless quality such as radio wave intensity and throughput of the communication terminal 903 after the predetermined time has elapsed from the wireless quality prediction server 901.

[0014] 10 is a diagram for explaining the problem of the conventional technology. In the conventional technology, a neural network 904 that has undergone machine learning using performance data measured in advance predicts the wireless quality of a communication terminal 903 based on, for example, terminal position A of the communication terminal 903.

[0015] However, for example, the base station to which communication terminal 903 connects may change depending on the situation, even if communication terminal 903 is at the same terminal position A. For example, assume that, during learning of neural network 904, communication terminal 903 received a lot of data from base station 1001b on frequency Y, and good communication quality was obtained at terminal position A.

[0016] However, when predicting the wireless quality at terminal position A, the communication terminal 903 may be connected to base station 1001a, which is different from base station 1001b. Here, it is assumed that base station 1001a performs wireless communication with communication terminal 903 at frequency X, which is lower than frequency Y. Furthermore, it is assumed that frequency X has stable received power but is more likely to experience base station congestion and therefore has a higher tendency for wireless quality to deteriorate than frequency Y.

[0017] In such a case, with the conventional technology, even if the terminal position is the same A, the communication terminal 903 is connected to a base station different from that at the time of learning, which poses a problem that the predicted wireless quality results will be degraded.

[0018] Therefore, in order to improve the accuracy of predicting wireless quality, the wireless quality prediction system 1 according to this embodiment has a system configuration as shown in FIG. 1, for example.

[0019] 1 is a diagram showing an example of the system configuration of a wireless quality prediction system according to this embodiment. The wireless quality prediction system 1 includes, for example, a communication terminal 10 which is a wireless station that performs wireless communication with a base station, and a prediction server 100 which can communicate with the communication terminal 10.

[0020] The prediction server (wireless quality prediction device) 100 is, for example, an information processing device having a computer configuration or a system including multiple computers. The prediction server 100 predicts the wireless quality of the communication terminal 10 based on location information indicating the location of the communication terminal 10 and information about the base station to which the communication terminal 10 is connected, and provides the prediction result to the communication terminal 10. The prediction server 100 is an example of a wireless quality prediction device according to the present embodiment.

[0021] Here, the location information includes, for example, information indicating the current location of the communication terminal 10 and / or information indicating an estimated location of the communication terminal 10 after a predetermined time has elapsed. The base station information includes, for example, identification information for identifying the base station (or a cell formed by the base station), such as a PCI (Physical Cell Identifier) ​​or a BSSID (Basic Service Set Identifier). The wireless quality includes, for example, throughput (uplink, downlink), received power, or delay amount.

[0022] The communication terminal 10 is a wireless station that connects to a base station among a plurality of base stations that has better wireless quality than the other base stations via a predetermined wireless communication such as 5G (5th Generation), LTE (Long Term Evolution), or wireless LAN (Local Area Network). The communication terminal 10 can communicate with the prediction server 100 via the base station, for example, and queries the prediction server 100 about the wireless quality at a location indicated by the location information. Based on the predicted wireless quality result notified by the prediction server 100, the communication terminal 10 performs communication control such as network switching.

[0023] <Functional Configuration> (Functional Configuration of Communication Terminal) Communication terminal 10 realizes, for example, each functional configuration as shown in Fig. 1 by executing a predetermined program on a computer provided in communication terminal 10. In the example of Fig. 1, communication terminal 10 has an information management unit 11, an inquiry unit 12, a communication control unit 13, a wireless communication unit 14, a wireless quality acquisition unit 15, a location information acquisition unit 16, a calculation unit 17, a base station information acquisition unit 18, and a storage unit 19. Note that at least a portion of each of the above functional configurations may be realized by hardware.

[0024] The information management unit 11 executes information management processing to acquire and manage the wireless quality acquired by the wireless quality acquisition unit 15, the location information acquired by the location information acquisition unit 16, the information on the base station acquired by the base station information acquisition unit 18, etc. For example, the information management unit 11 transmits learning data such as the location information of the communication terminal 10, the information on the base station, and the wireless quality information to the prediction server 100. The information management unit 11 also provides the inquiry unit 12 with information (e.g., location information, information on the base station, etc.) necessary for the inquiry unit 12 to inquire about the predicted value of wireless quality.

[0025] The inquiry unit 12 executes an inquiry process of transmitting an inquiry about the predicted value of wireless quality, including the location information of the communication terminal 10, the information about the base station, etc., to the prediction server. For example, the inquiry unit 12 inquires of the prediction server 100 about the predicted value of communication quality of the communication terminal 10 after a predetermined time has elapsed. The inquiry unit 12 also notifies the communication control unit 13 and the like of the predicted value of wireless quality notified by the prediction server 100.

[0026] The communication control unit 13 executes a communication control process for controlling wireless communication by the wireless communication unit 14. For example, the communication control unit 13 determines whether to switch networks based on the predicted value of wireless quality notified by the inquiry unit 12, and if the network is to be switched, controls the wireless communication unit 14 to switch the base station to which the communication terminal 10 is connected.

[0027] The wireless communication unit 14 connects to one or more base stations under the control of the communication control unit 13 and performs predetermined wireless communication.

[0028] The wireless quality acquisition unit 15 executes a wireless quality acquisition process to acquire wireless quality information such as throughput, etc. The wireless quality acquired by the wireless quality acquisition unit 15 may include, in addition to the throughput, for example, received power or delay amount.

[0029] The location information acquisition unit 16 executes a location information acquisition process to acquire location information indicating the location of the communication terminal 10. The location information acquisition unit 16 acquires the current location of the communication terminal 10, for example, using a GPS (Global Positioning System) device or the like. Preferably, the location information acquisition unit 16 also uses the calculation unit 17 to acquire an estimated location of the communication terminal after a predetermined time has elapsed from the current location of the communication terminal 10.

[0030] Calculation unit 17 executes a calculation process to calculate an estimated position of communication terminal 10 after a predetermined time has elapsed, based on the current position of communication terminal 10 acquired by position information acquisition unit 16. For example, calculation unit 17 calculates an estimated position of communication terminal 10 after a predetermined time (n seconds) has elapsed, based on the current position of communication terminal 10 acquired by position information acquisition unit 16 and the movement of communication terminal 10 measured by sensors such as an acceleration sensor and an angle sensor. Alternatively, calculation unit 17 may store a history of the current position of communication terminal 10 acquired by position information acquisition unit 16 in storage unit 19, and estimate an estimated position of communication terminal 10 after a predetermined time (n seconds) has elapsed, based on this history.

[0031] 1, the communication terminal 10 has the calculation unit 17, but this is just one example. For example, the calculation unit 17 may be included in the prediction server 100. In this case, the prediction server 100 may use the calculation unit 17 to calculate an estimated position of the communication terminal 10 after a predetermined time has elapsed from the current position of the communication terminal 10 acquired from the communication terminal 10.

[0032] The base station information acquisition unit 18 executes a base station information acquisition process to acquire information about the base station to which the communication terminal 10 is connected. The base station information acquired by the base station information acquisition unit 18 includes, for example, information specific to the base station, such as PCI, and information about the frequency used (used band, center frequency, bandwidth, etc.).

[0033] The memory unit 19 stores various information and data, including, for example, information managed by the information management unit 11, predicted values ​​of wireless quality notified by the prediction server 100, and history of the current location acquired by the location information acquisition unit 16.

[0034] (Functional Configuration of Prediction Server) The prediction server 100 realizes, for example, each functional configuration as shown in Fig. 1 by executing a predetermined program by a computer included in the prediction server 100. In the example of Fig. 1, the prediction server 100 includes an information acquisition unit 110, a response unit 120, a learning device management unit 130, a plurality of learning devices 101-1, 101-2, ... 101-n (n is an integer of 2 or more), a storage unit 140, and the like.

[0035] At least some of the above functional configurations may be realized by hardware. Furthermore, at least some of the above functional configurations are not limited to being realized by physical machines, and may be realized, for example, by a program executed by a virtual machine on the cloud. In the following description, the term "learning device 101" will be used to refer to any learning device among the multiple learning devices 101-1, 101-2, ..., 101-n.

[0036] The information acquisition unit 110 acquires learning data transmitted by multiple communication terminals 10, including location information of the communication terminals 10, base station information, and wireless quality information, and performs an information acquisition process to store the data in a memory unit 140, etc.

[0037] The responding unit 120 executes a response process to respond to an inquiry about a predicted value of wireless quality from the communication terminal 10. The responding unit 120 includes, for example, an acquiring unit 121 and a notifying unit 122.

[0038] The acquisition unit 121 executes an acquisition process to acquire location information indicating the location of the communication terminal 10, information on the base station to which the communication terminal 10 is connected, etc., which are included in the inquiry for the predicted value of wireless quality transmitted by the inquiry unit 12 of the communication terminal 10. As a specific example, the acquisition unit 121 acquires location information indicating an estimated location of the communication terminal 10 after a predetermined time has elapsed, which are included in the inquiry for the predicted value of wireless quality, and the PCI (an example of base station information) of the base station to which the communication terminal 10 is connected.

[0039] As another example, if the prediction server 100 includes the calculation unit 17, the acquisition unit 121 may acquire location information indicating the current location of the communication terminal 10 and the PCI of the base station to which the communication terminal 10 is connected, which are included in the inquiry about the predicted value of wireless quality. In this case, the calculation unit 17 calculates an estimated location of the communication terminal 10 after a predetermined time has elapsed, based on the current location of the communication terminal acquired by the acquisition unit 121.

[0040] The notification unit 122 executes a notification process to notify the communication terminal 10 of the predicted value of the wireless quality of the communication terminal 10 after a predetermined time has elapsed, as predicted by the learning device management unit 130, based on the information acquired by the acquisition unit 121.

[0041] The learning device management unit 130 executes a learning device management process for managing a plurality of learning devices 101-1, 101-2, ..., 101-n that predict the wireless quality of the communication terminal 10, based on the location information of the communication terminal 10. The learning device management unit 130 includes, for example, a creation unit 131, a prediction unit 132, and a determination unit 133.

[0042] The creation unit 131 executes a creation process to create multiple learning devices 101 for each network characteristic based on, for example, the location information of the communication terminal 10, base station information, and the value of wireless quality at the location indicated by the location information stored in the memory unit 140 by the information acquisition unit 110.

[0043] For example, the creation unit 131 may create multiple learning devices 101 for each piece of base station-specific information (PCI, Cell ID, BSSID, etc.), or may create multiple learning devices 101 for each frequency used (band, center frequency, bandwidth, etc.).

[0044] In a cellular network such as 5G or LTE, creating a learning device 101 for each PCI increases the number of learning devices 101, and a huge amount of data is required for learning each learning device 101. Therefore, in a cellular network, it is desirable for the creation unit 131 to create multiple learning devices 101 for each piece of band information (an example of a frequency used). In this case, the creation unit 131 stores and manages the correspondence between PCIs and band information in correspondence information 141, for example.

[0045] As another example, the creating unit 131 may create multiple learning devices 101 further based on current environmental information.

[0046] The prediction unit 132 executes a prediction process for predicting the wireless quality of the communication terminal 10, using a learner 101 corresponding to the information about the base station, from among a plurality of learners 101 that have been trained in advance to predict the wireless quality of the communication terminal 10, based on location information indicating the location of the communication terminal 10. For example, the prediction unit 132 predicts a predicted value of the wireless quality (e.g., throughput) of the communication terminal 10, using a learner 101 determined by the determination unit 133 from among the plurality of learners 101 created by the creation unit 131.

[0047] The determination unit 133 executes a determination process to determine, from among the multiple learning devices 101 created by the creation unit 131, a learning device 101 that corresponds to the information on the base station acquired by the acquisition unit 121. As a specific example, the determination unit 133 identifies a usage band that corresponds to the information on the base station acquired by the acquisition unit 121, based on correspondence information 141 that indicates a correspondence relationship between the information on the base station and the usage frequency and that is stored in the storage unit 140 by the creation unit 131. Furthermore, the determination unit 133 determines, from among the multiple learning devices 101 created for each usage band by the creation unit 131, the learning device 101 that corresponds to the identified usage band as the learning device 101 to be used by the prediction unit 132.

[0048] The multiple learners 101-1, 101-2, 101-n are learners 101 created for each network characteristic by the creation unit 131. Each learner 101 is, for example, a neural network that predicts the wireless quality (e.g., throughput) of the communication terminal 10 by inputting location information of the communication terminal 10. For example, the creation unit 131 performs machine learning on the multiple learners 101 for each used band so as to predict the wireless quality of the communication terminal 10 from learning data including the location information and wireless quality information of the communication terminal 10 acquired by the information acquisition unit 110.

[0049] As a specific example, the creation unit 131 identifies the band used by the communication terminal 10 from the information of the base station acquired by the information acquisition unit 110, and performs machine learning to create a learning device 101 corresponding to the identified band used, using the location information and wireless quality information of the communication terminal 10 acquired by the information acquisition unit 110.

[0050] In addition, the prediction unit 132 identifies the band used by the communication terminal 10 from the base station information acquired by the acquisition unit 121, and predicts the communication quality of the communication terminal 10 by inputting the location information acquired by the acquisition unit 121 into the learning device 101 corresponding to the identified band used.

[0051] The storage unit 140 stores various information or data including, for example, the above-mentioned correspondence information 141 and the learning data acquired by the information acquisition unit 110 .

[0052] 10 , for example, when the communication terminal 903 is connected to the base station 1001a, the wireless quality prediction system 1 uses the learning device 101 for frequency X corresponding to the base station 1001a to predict the wireless quality of the communication terminal 903. Also, when the communication terminal 903 is connected to the base station 1001b, the wireless quality prediction system 1 uses the learning device 101 for frequency Y corresponding to the base station 1001b to predict the wireless quality of the communication terminal 903.

[0053] As a result, the wireless quality prediction system 1 according to this embodiment can improve the accuracy of predicting wireless quality.

[0054] 1 is an example. For example, at least some of the functional components of the prediction server 100 may be included in the communication terminal 10. For example, the communication terminal 10 may include an acquisition unit 121, a prediction unit 132, a determination unit 133, and multiple learning devices 101-1, 101-2, ... In this case, the communication terminal 10 serves as a wireless quality prediction device.

[0055] <Processing Flow> Next, the processing flow of the wireless quality prediction method according to this embodiment will be described.

[0056] 2 is a diagram illustrating an overview of a process according to Example 1. In Fig. 2, a wireless communication system 200 includes base stations 211a and 211b, a communication terminal 10 capable of wireless communication with the base stations 211a and 211b within a target area 210, and a prediction server 100 that predicts wireless quality of the communication terminal 10.

[0057] (Processing during learning) The communication terminal 10 collects, for example, location information indicating the location of the communication terminal 10, information about base stations, and wireless quality measurements while moving within the target area 210 (step S1). The communication terminal 10 also transmits the collected data to the prediction server 100 (step S2).

[0058] The prediction server 100 uses the collected data notified from the plurality of communication terminals 10 as learning data to train a learning device 101 corresponding to the information on the base station so that wireless quality can be predicted from the location information.

[0059] As an example, when the notified collected data includes information (e.g., PCI) of the base station 211a, the prediction server 100 uses the location information, wireless quality measurement values, etc. included in the collected data to learn the learner 101-1 corresponding to the information of the base station 211a. Similarly, when the notified collected data includes information of the base station 211b, the prediction server 100 uses the location information, wireless quality measurement values, etc. included in the collected data to learn the learner 101-2 corresponding to the information of the base station 211b.

[0060] At this time, the prediction server 100 may use the correspondence information 141 to obtain additional information corresponding to the base station information, and based on the obtained information, determine the learning device 101 corresponding to the base station information.

[0061] (Processing during prediction) The communication terminal 10 collects, for example, location information indicating the location of the communication terminal 10, information on base stations, etc. (Step S4) while moving within the target area 210. Furthermore, the communication terminal 10 uses the collected location information and information on base stations to inquire of the prediction server 100 about a predicted value of wireless quality after a predetermined time has elapsed (Step S5).

[0062] When the prediction server 100 receives an inquiry about the predicted value of wireless quality from the communication terminal 10, it uses the learning device 101 corresponding to the base station information included in the inquiry to predict the wireless quality of the communication terminal 10 after a predetermined time has passed (step S6).

[0063] As an example, when an inquiry from the communication terminal 10 includes information (e.g., PCI) of the base station 211a, the prediction server 100 predicts the wireless quality of the communication terminal 10 after a predetermined time has elapsed using the learner 101-1 corresponding to the information of the base station 211a. For example, when the location information included in the inquiry from the communication terminal 10 includes an estimated location of the communication terminal 10 after a predetermined time has elapsed, calculated by the calculation unit 17, the prediction server 100 inputs the estimated location of the communication terminal 10 to the learned learner 101-1. As a result, the learner 101-1 calculates a predicted value of the wireless quality of the communication terminal 10 after the predetermined time has elapsed.

[0064] In addition, if the location information included in the inquiry from the communication terminal 10 does not include the estimated location of the communication terminal 10 after a predetermined time has passed, the prediction server 100 may calculate the estimated location of the communication terminal 10 after a predetermined time has passed based on the current location of the communication terminal 10.

[0065] Similarly, when the inquiry from the communication terminal 10 includes information about the base station 211b, the prediction server 100 uses the learning device 101-2 corresponding to the information about the base station 211b to predict the wireless quality of the communication terminal 10 after a predetermined time has elapsed. In addition, the prediction server 100 notifies the inquiry source communication terminal 10 of the predicted value (prediction result) of the wireless quality of the communication terminal 10 (step S7).

[0066] The communication terminal 10 controls wireless communication (switching networks, etc.) based on the predicted value of wireless quality notified by the prediction server 100 (step S8).

[0067] As described above, the conventional technology has a problem that even if the terminal location is the same, if the communication terminal 10 is connected to a base station different from that at the time of learning, the predicted wireless quality results will be degraded.

[0068] 2, for example, when the communication terminal 10 is connected to the base station 211a, the wireless quality prediction system 1 according to this embodiment can predict the wireless quality of the communication terminal 10 using the learning device 101-1 corresponding to the information of the base station 211a. Similarly, when the communication terminal 10 is connected to the base station 211b, the wireless quality prediction system 1 can predict the wireless quality of the communication terminal 10 using the learning device 101-2 corresponding to the information of the base station 211b.

[0069] Therefore, according to this embodiment, in the wireless quality prediction system 1 that predicts the wireless quality of the communication terminal 10, it is possible to improve the prediction accuracy of the wireless quality.

[0070] (Learning Process) Fig. 3 is a flowchart showing an example of learning process according to Example 1. This process shows an example of specific process corresponding to, for example, steps S1 to S3 in Fig. 2.

[0071] In step S301, the information management unit 11 acquires location information, PCI, band information, etc. Here, the location information includes information indicating the current location of the communication terminal 10 acquired by the location information acquisition unit 16. The PCI and band information are examples of base station information acquired by the base station information acquisition unit 18.

[0072] In step S302, the wireless quality acquisition unit 15 measures the throughput of the communication terminal 10, and the information management unit 11 stores collected data in a database (DB) that associates the measured throughput value with the information acquired in step S301.

[0073] For example, the information management unit 11 transmits the collected data to the prediction server 100. As a result, the information acquisition unit 110 of the prediction server 100 stores the collected data acquired from the communication terminal 10 in a DB such as the storage unit 140.

[0074] In step S303, the creation unit 131 uses a plurality of collected data stored in the DB to create learning devices 101-1, 101-2, ... for each piece of band information that have been trained to be able to predict throughput from position information for each piece of band information.

[0075] In step S304, the creation unit 131 saves the created plurality of learning devices 101-1, 101-2, .... The storage destination of the plurality of learning devices 101-1, 101-2, ... may be the prediction server 100 or the communication terminal 10.

[0076] In the example of FIG. 1, the creation unit 131 stores a plurality of learning devices 101-1, 101-2, . . . in the prediction server 100.

[0077] Furthermore, the creation unit 131 executes the process of step S305 in parallel with the processes of steps S303 and S304. In step S305, the creation unit 131 creates a table T that can identify band information from PCI.

[0078] FIG. 4 is a diagram illustrating an image of a table T according to the first embodiment. In the example of FIG. 4, table T stores PCIs and operating bands in association with each other. PCIs are identification information that identify physical cells formed by a base station. Operating bands are an example of band information that indicates the frequency bands used by the base station. In the example of FIG. 4, table T indicates that base stations with PCIs ID_A and ID_C use operating band "BAND_X," and that a base station with PCI ID_B uses operating band "BAND_Y."

[0079] Table T is an example of the correspondence information 141 according to the present embodiment. Table T shown in Fig. 4 is also an example. For example, table T may further store information about the learning device 101 corresponding to the used band.

[0080] 3, the wireless quality prediction system 1 can create, for each piece of band information, the learning devices 101-1, 101-2, . . . that have been trained in advance, and a table T that is an example of the correspondence information 141.

[0081] (Processing During Prediction) Fig. 3 is a flowchart showing an example of processing during learning according to Example 1. This processing shows an example of specific processing corresponding to, for example, the processing of steps S4 to S8 in Fig. 2.

[0082] In step S501, the acquisition unit 121 acquires location information and PCI included in the inquiry about the predicted value of wireless quality transmitted by the communication terminal 10. The location information acquired by the acquisition unit 121 includes the current location p １ , or the estimated position p of the communication terminal 10 after a predetermined time t has elapsed ｔ Includes:

[0083] In step S502, the acquisition unit 121 acquires the current position p １ From this, the estimated position p of the communication terminal 10 after a predetermined time t has elapsed is calculated. ｔ The estimated position p of the communication terminal 10 is calculated based on the position information acquired in step S501. ｔ If the value is included, the acquisition unit 121 can omit the process of step S502.

[0084] The acquisition unit 121 acquires the estimated position p ｔ The method for calculating the estimated position p of the communication terminal 10 may be, for example, a calculation method that assumes uniform linear motion, or another calculation method. For example, the prediction server 100 includes the calculation unit 17 described above, and the acquisition unit 121 uses the calculation unit 17 to calculate the estimated position p of the communication terminal 10. ｔ may be calculated.

[0085] In step S503, the determination unit 133 identifies band information corresponding to the acquired PCI from, for example, table T as shown in Fig. 4. Furthermore, in step S504, the determination unit 133 determines the learning device 101 corresponding to the identified band information.

[0086] In step S505, the determined learning device 101 is configured to estimate the position p ｔ and calculates a predicted value of the throughput of the communication terminal 10.

[0087] In step S506, the communication control unit 13 of the communication terminal 10 executes control using the predicted value of the throughput predicted by the prediction unit 132. For example, the communication control unit 13 performs control such as switching of the network or changing of the video rate based on the predicted value of the throughput of the communication terminal 10 acquired by the inquiry unit 12 from the prediction server 100.

[0088] In step S507, the wireless quality prediction system 1 determines whether to end the prediction (inference) operation. If the prediction operation is not to be ended, the wireless quality prediction system 1 shifts the process to step S501. On the other hand, if the prediction operation is to be ended, the wireless quality prediction system 1 ends the process of FIG. 5.

[0089] By the processing of Figure 5, the wireless quality prediction system 1 can predict the wireless quality of the communication terminal 10 using a learning device 101 corresponding to the information of the base station, out of multiple learning devices 101 that have been pre-trained to predict the wireless quality of the communication terminal 10 based on location information.

[0090] [Example 2] In Example 1, an example has been described in which the wireless quality prediction system 1 creates a learning device 101 for each operating frequency (operating band, center frequency, bandwidth, etc.). However, this is just one example. The wireless quality prediction system 1 may create multiple learning devices 101-1, 101-2, ... based on, for example, current environment information.

[0091] Fig. 6 is a diagram for explaining how to create different learning modules according to Example 2. The wireless quality prediction system 1 according to Example 2 creates learning modules 101-1, 101-2, ... that predict wireless quality for each of environmental information 601 and base station information 602 (step S11). In the example of Fig. 6, the environmental information 601 includes information such as the city size (e.g., urban area, rural area, etc.) and the current time (clock).

[0092] In this case, the learning devices 101 can be created differently, for example: Learning device 101-1: "City size: large, time: 9:00, base station: A" Learning device 101-2: "City size: large, time: 8:00, base station: B" Learning device 101-3: "City size: small, time: 8:00, base station: C" and so on.

[0093] It is advisable to register in advance the city-scale information corresponding to the PCI of the base station in, for example, a table T (correspondence information 141) as shown in FIG.

[0094] Furthermore, in the above example, for example, "time" is subdivided into smaller parts to create different learning devices 101, but the granularity of the subdivision can be freely set by an administrator or the like.

[0095] During prediction, the wireless quality prediction system 1 according to the second embodiment predicts the wireless quality of the communication terminal 10 using the learning device 101 corresponding to the current environment information and the information of the base station (step S12). For example, when the prediction server 100 receives a query for a wireless quality prediction value including location information and information of the base station (e.g., PCI) from the communication terminal 10, the prediction server 100 acquires city-size information corresponding to the information of the base station from the correspondence information 141. The prediction server 100 also acquires the current time, determines the learning device 101 corresponding to the acquired city size, current time, and information of the base station, and inputs the location information to the determined learning device 101 to predict the wireless quality of the communication terminal 10. For example, if the acquired city size is "large," the current time is in the "8:00" range, and the base station is "B," the prediction server 100 predicts the wireless quality of the communication terminal 10 using the learning device 101-2.

[0096] In this way, the wireless quality prediction system 1 may create a different learning device 101 for each piece of base station information and environmental information, and at the time of prediction, may predict the wireless quality of the communication terminal 10 using the learning device 101 corresponding to the base station information and environmental information obtained from the communication terminal 10.

[0097] <Hardware Configuration Example> (Hardware Configuration of Prediction Server) The prediction server 100 has, for example, the hardware configuration of a computer 700 as shown in Fig. 7. Alternatively, the prediction server 100 is realized by a plurality of computers 700.

[0098] 7 is a diagram showing the hardware configuration of a computer according to this embodiment. In the example of FIG. 7, a computer 700 includes a processor 701, a memory 702, a storage device 703, a communication device 704, an input device 705, an output device 706, and a bus B.

[0099] The processor 701 is, for example, an arithmetic unit such as a CPU (Central Processing Unit) that executes predetermined programs to realize various functions. The memory 702 is a storage medium readable by the computer 700, and includes, for example, a RAM (Random Access Memory) and a ROM (Read Only Memory). The storage device 703 is a computer-readable storage medium, and may include, for example, a HDD (Hard Disk Drive), an SSD (Solid State Drive), various optical disks, and magneto-optical disks.

[0100] The communication device 704 includes one or more pieces of hardware (communication devices) for communicating with other devices via a wireless or wired network. The input device 705 is an input device (e.g., a keyboard, a mouse, a microphone, a switch, a button, a sensor, etc.) that accepts input from the outside. The output device 706 is an output device (e.g., a display, a speaker, an LED lamp, etc.) that outputs to the outside.

[0101] The bus B is commonly connected to the above components and transmits, for example, address signals, data signals, and various control signals. The processor 701 is not limited to a CPU, and may be, for example, a DSP (Digital Signal Processor), a PLD (Programmable Logic Device), or an FPGA (Field Programmable Gate Array).

[0102] (Hardware Configuration of Communication Terminal) Fig. 8 is a diagram showing an example of the hardware configuration of a communication terminal according to this embodiment. The communication terminal 10 includes, for example, a GPS device 801, a sensor 802, and the like in addition to the hardware configuration of the computer 700 described in Fig. 7.

[0103] The GPS device 801 is a positioning device that receives positioning signals transmitted by GPS satellites and outputs position information indicating the current position of the communication terminal 10. The sensor 802 is a detection device that detects the movement of the communication terminal 10, such as an acceleration sensor or an angle sensor.

[0104] (Supplementary Note) The prediction server 100 in this embodiment is not limited to being realized by a dedicated device, but may be realized by a general-purpose computer. In this case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read into a computer system and executed to realize the function. Note that the term "computer system" here includes hardware such as an OS and peripheral devices.

[0105] Furthermore, "computer-readable recording media" includes various storage devices such as portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, as well as hard disks built into computer systems. Furthermore, "computer-readable recording media" may also include devices that dynamically store programs for a short period of time, such as communication lines when transmitting programs via networks such as the Internet or communication lines such as telephone lines, and devices that store programs for a certain period of time, such as volatile memory within computer systems that serve as servers or clients in such cases.

[0106] Furthermore, the above program may be one that realizes part of the above-mentioned functions, or may be one that can realize the above-mentioned functions in combination with a program already recorded in a computer system, or may be one that is realized using hardware such as a PLD (Programmable Logic Device) or FPGA (Field Programmable Gate Array).

[0107] <Effects of the embodiment> According to the present embodiment, in the wireless quality prediction system 1 that predicts the wireless quality of the communication terminal 10, it is possible to improve the prediction accuracy of the wireless quality.

[0108] For example, the wireless quality prediction system 1 according to this embodiment makes it possible to predict wireless quality taking into account network characteristics including base station information and / or environmental information, thereby improving the accuracy of wireless quality prediction.

[0109] Furthermore, base station information not only improves the accuracy of radio wave quality predictions based on frequency characteristics, but also makes it possible to make predictions that take congestion into account to some extent, since congestion susceptibility varies depending on the assigned band. For example, the platinum band is more likely to be assigned to terminals, making it possible to predict radio quality that takes into account characteristics such as congestion.

[0110] Furthermore, the wireless quality prediction system 1 according to this embodiment has correspondence information 141 indicating the correspondence between the information of the base station (such as PCI) and other information. As a result, even if, for example, an inquiry from the communication terminal 10 includes only PCI, it is possible to predict the wireless quality of the communication terminal 10 using, for example, the learning device 101 for each used band, or the learning device 101 for each base station information and environmental information.

[0111] Summary of Embodiments This specification discloses at least the wireless quality prediction system, wireless quality prediction device, wireless quality prediction method, and program described in the following paragraphs. (Item 1) A wireless quality prediction system comprising: an acquisition unit that acquires location information indicating the location of a communication terminal and information on a base station to which the communication terminal is connected; and a prediction unit that predicts the wireless quality of the communication terminal using a learning unit corresponding to the information on the base station among multiple learning units that have been trained in advance to predict the wireless quality of the communication terminal based on the location information. (Item 2) The wireless quality prediction system described in paragraph 1, comprising: a creation unit that creates the multiple learning units for each network characteristic based on the location information, information on the base station, and a value of the wireless quality at the location indicated by the location information. (Item 3) The wireless quality prediction system described in paragraph 2, wherein the creation unit creates the multiple learning units for each operating frequency, and the wireless quality prediction system comprises: a determination unit that determines a learning unit corresponding to the information on the base station based on correspondence information indicating a correspondence relationship between the information on the base station and the operating frequency. (Clause 4) The wireless quality prediction system according to Clause 2, wherein the creation unit creates the plurality of learning devices further based on current environmental information, and the system includes a determination unit that determines a learning device that corresponds to information about the base station based on the environmental information and information about the base station. (Clause 5) The wireless quality prediction system according to any of Clauses 1 to 4, wherein the creation unit calculates an estimated position of the communication terminal after a predetermined time has elapsed based on a current position of the communication terminal, and the prediction unit predicts communication quality of the communication terminal at the estimated position of the communication terminal. (Clause 6) A wireless quality prediction device comprising: an acquisition unit that acquires location information that indicates the position of the communication terminal and information about a base station to which the communication terminal is connected, and a prediction unit that predicts wireless quality of the communication terminal using a learning device that corresponds to the information about the base station acquired by the acquisition unit, from among a plurality of learning devices that have been trained in advance to predict wireless quality at the position of the communication terminal.(Clause 7) A wireless quality prediction method in which a computer executes: an acquisition process of acquiring location information indicating the location of a communication terminal and information on a base station to which the communication terminal is connected; and a prediction process of predicting the wireless quality of the communication terminal using a learning device corresponding to the information on the base station acquired in the acquisition process, among a plurality of learning devices that have been trained in advance to predict wireless quality at the location of the communication terminal. (Clause 8) A program, or a storage medium storing a program, that causes a computer to execute the wireless quality prediction method described in Clause 7.

[0112] Although the present embodiment has been described above, the present invention is not limited to such a specific embodiment, and various modifications and changes are possible within the scope of the gist of the present invention described in the claims.

[0113] REFERENCE SIGNS LIST 1 Wireless quality prediction system 10 Communication terminal 17 Calculation unit 101, 101-1, 101-2, ..., 101-n Learning device 121 Acquisition unit 131 Creation unit 132 Prediction unit 133 Determination unit 141 Correspondence information 700 Computer

Claims

1. An acquisition unit that acquires position information indicating the position of a communication terminal and information on a base station to which the communication terminal is connected; and a prediction unit that predicts the radio quality of the communication terminal by using a learning device corresponding to the information on the base station among a plurality of learning devices that have been pre-learned to predict the radio quality of the communication terminal based on the position information. A radio quality prediction system.

2. The radio quality prediction system according to claim 1, further comprising a creation unit that creates the plurality of learning devices for each network characteristic based on the position information, the information on the base station, and the value of the radio quality at the position indicated by the position information.

3. The creation unit creates the plurality of learning devices for each operating frequency, and the radio quality prediction system has a determination unit that determines a learning device corresponding to the information on the base station based on correspondence information indicating the correspondence between the information on the base station and the operating frequency. The radio quality prediction system according to claim 2.

4. The creation unit further creates the plurality of learning devices based on current environmental information, and the radio quality prediction system has a determination unit that determines a learning device corresponding to the information on the base station based on the environmental information and the information on the base station. The radio quality prediction system according to claim 2.

5. The radio quality prediction system according to claim 1, further comprising a calculation unit that calculates an estimated position of the communication terminal after a predetermined time has elapsed based on the current position of the communication terminal, and the prediction unit predicts the communication quality of the communication terminal at the estimated position of the communication terminal.

6. An acquisition unit that acquires position information indicating the position of a communication terminal and information on a base station to which the communication terminal is connected; and a prediction unit that predicts the radio quality of the communication terminal by using a learning device corresponding to the information on the base station acquired by the acquisition unit among a plurality of learning devices that have been pre-learned to predict the radio quality at the position of the communication terminal. A radio quality prediction device.

7. A method for predicting radio quality, in which a computer executes an acquisition process of acquiring position information indicating the position of a communication terminal and information on a base station to which the communication terminal is connected, and a prediction process of predicting the radio quality of the communication terminal by using a learning device corresponding to the information on the base station acquired in the acquisition process among a plurality of learning devices that have been pre-learned to predict the radio quality at the position of the communication terminal.

8. A program for causing a computer to execute the radio quality prediction method according to claim 7.

Images