Wireless communication device and communication quality prediction method

Abstract

This wireless communication device comprises: a communication unit (11) that performs wireless communication with a communication terminal (2) and acquires the communication quality of the communication terminal (2); an acquisition unit (12) that acquires spatial information; a data generation unit (13) that classifies the spatial information into a plurality of pieces of spatial information in accordance with the environment information, and generates training data for each of the classified pieces of spatial information on the basis of each of the classified pieces of spatial information and the communication quality; and a model generation unit (14) that performs machine learning by using the training data and constructs a prediction model for the communication quality for each of the classified pieces of spatial information. The wireless communication device comprises: an estimation unit (15) that, during communication with the communication terminal (2), estimates environmental information within a prescribed region on the basis of the communication quality; a selection unit (16) that selects a prediction model corresponding to the estimated environmental information from the plurality of prediction models on the basis of the environmental information; and a use unit (17) that, during communication with the communication terminal (2), inputs the communication quality acquired by the communication unit (11) and the spatial information acquired by the acquisition unit (12) to the prediction model selected by the selection unit (16) and predicts the communication quality of the communication terminal (2).

Description

Wireless communication device and communication quality prediction method

[0001] The present disclosure relates to a wireless communication device and a communication quality prediction method.

[0002] When using a communication terminal equipped with wireless communication functions, communication quality may change due to changes in the surrounding environment, such as the movement of nearby objects. Changes in communication quality can cause the communication terminal's services and systems to fail to meet the communication quality requirements. In particular, the impact of the surrounding environment on communication quality is significant for IEEE 802.11ad, which uses millimeter wave bands, and 5G cellular communications. By predicting communication quality in advance, it may be possible to take measures before services and systems are affected.

[0003] Known methods for predicting communication quality include the techniques disclosed in Non-Patent Documents 1 to 3. Non-Patent Document 1 discloses a technique for predicting communication quality using spatial information acquired from a camera and LiDAR while a robot carrying a communication terminal is moving, with the surrounding environment, such as obstacles, fixed.

[0004] Non-Patent Documents 2 and 3 disclose a technique for predicting communication quality using spatial information obtained from a depth camera and LiDAR in a situation where a communication terminal and an AP (Access Point) are fixed and two users are walking so as to block the communication terminal and the AP.

[0005] T. Zhang, J. Liu and F. Gao, “Vision Aided Beam Tracking and Frequency Handofffor mmWave Communications,” Proc. IEEE Conf. Comput.Commun. Workshops, pp. 1-2, May 2022. T. Nishioet al., “Proactive ReceivedPower Prediction UsingMachine Learning and Depth Imagesfor mmWave Networks,” in IEEE Journalon Selected Areasin Communications, vol. 37, no. 11, pp. 2413-2427, Nov. 2019S. Ohta,T. Nishio, R. Kudo and K. Takahashi, “Millimeter-wave Received Power Prediction Using Point CloudData and Supervised Learning,” Proc. IEEE 95th Veh. Technol.Conf., pp. 1-5, Jun. 2022.

[0006] However, the techniques disclosed in the above-mentioned Non-Patent Documents 1 to 3 do not mention the influence of environmental changes on communication quality, such as the movement of surrounding objects, the number of surrounding objects, etc. Furthermore, a prediction model that addresses this is necessary, but predictions can only be used in specific environments, and no prediction model that takes into account environmental changes, such as an increase or decrease in the number of communication terminals or the number of moving objects in the vicinity, is mentioned.

[0007] The present disclosure has been made in consideration of the above circumstances, and its purpose is to provide a wireless communication device and a communication quality prediction method that are capable of predicting communication quality with high accuracy even when environmental changes occur.

[0008] A wireless communication device of one aspect of the present disclosure includes a communication unit that performs wireless communication with a communication terminal and acquires communication quality of the communication terminal; an acquisition unit that acquires spatial information including movement information of the communication terminal and surrounding objects within a specified area; a data generation unit that classifies the spatial information into multiple pieces of spatial information according to environmental information and generates training data for each piece of classified spatial information based on each piece of classified spatial information and the communication quality; a model generation unit that performs machine learning using the training data and constructs a prediction model of communication quality for each piece of classified spatial information; an estimation unit that, when communicating with the communication terminal, estimates environmental information within the specified area based on the communication quality; a selection unit that selects a prediction model corresponding to the estimated environmental information from multiple prediction models based on the estimated environmental information; and a utilization unit that, when communicating with the communication terminal, inputs the communication quality acquired by the communication unit and the spatial information acquired by the acquisition unit into the prediction model selected by the selection unit to predict the communication quality of the communication terminal.

[0009] A communication quality prediction method according to one aspect of the present disclosure is a communication quality prediction method performed by a wireless communication device, which performs wireless communication with a communication terminal, acquires the communication quality of the communication terminal, acquires spatial information including movement information of the communication terminal and surrounding objects within a specified area, classifies the spatial information into multiple pieces of spatial information according to environmental information, generates training data for each piece of classified spatial information based on each piece of classified spatial information and the communication quality, performs machine learning using the training data, and constructs a communication quality prediction model for each piece of classified spatial information, estimates environmental information within the specified area based on the communication quality during communication with the communication terminal, selects a prediction model corresponding to the estimated environmental information from multiple prediction models based on the estimated environmental information, and, during communication with the communication terminal, inputs the communication quality acquired by the communication unit and the spatial information acquired by the acquisition unit into the prediction model selected by the selection unit to predict the communication quality of the communication terminal.

[0010] According to the present disclosure, it is possible to predict communication quality with high accuracy even when an environmental change occurs.

[0011] FIG. 1 is a block diagram showing the configuration of a wireless communication device according to an embodiment. FIG. 2 is a flowchart showing a processing procedure of the wireless communication device according to an embodiment. FIG. 3A is an explanatory diagram showing a movement path when one robot moves within a space. FIG. 3B is a diagram showing a histogram of throughput when one robot moves within a space. FIG. 3C is a graph showing the relationship between the y coordinate of robot A and the throughput. FIG. 4A is an explanatory diagram showing movement paths when two robots move within a space. FIG. 4B is a diagram showing a histogram of throughput when two robots move within a space. FIG. 4C is a graph showing the relationship between the y coordinate of robot A and the throughput. FIG. 5A is an explanatory diagram showing movement paths when three robots move within a space. FIG. 5B is a diagram showing a histogram of throughput when three robots move within a space. FIG. 5C is a graph showing the relationship between the y coordinate of robot A and the throughput. FIG. 6A is an explanatory diagram showing movement paths when four robots move within a space. Fig. 6B is a diagram showing a histogram of throughput when four robots move in a space. Fig. 6C is a graph showing the relationship between the y coordinate of robot A and throughput. Fig. 7 is a graph showing the relationship between predicted error and CDF. Fig. 8 is a block diagram showing the hardware configuration of this embodiment.

[0012] Hereinafter, an embodiment will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a wireless communication device 1 and its peripheral devices according to an embodiment. As shown in FIG. 1, the wireless communication device 1 is capable of wireless communication with a communication terminal 2. The communication terminal 2 is a mobile device such as a smartphone that can be carried by a user. The communication terminal 2 may employ a communication method such as cellular communication "5G" or "IEEE 802.11ac" to reduce the influence of surrounding objects.

[0013] The wireless communication device 1 is capable of wireless or wired communication with the sensor 3. The sensor 3 may be a camera or LiDAR fixed within an area in which the communication terminal 2 moves.

[0014] As shown in FIG. 1, the wireless communication device 1 includes a communication unit 11, an acquisition unit 12, a data generation unit 13, a model generation unit 14, an estimation unit 15, a selection unit 16, and a utilization unit 17.

[0015] The communication unit 11 performs wireless communication with the communication terminal 2. The communication unit 11 acquires communication quality information through communication with the communication terminal 2. The communication quality refers to the communication terminal 2's throughput, received signal power, signal-to-noise power ratio, signal-to-interference-plus-noise power ratio, received signal strength indication (RSSI), reference signal received quality (RSRQ), packet error rate, number of arriving bits, bit error rate, number of arriving bits per unit time, modular code index (MCS), number of retransmissions, packet arrival delay time, error correction technology settings, communication terminal user contract information, differential information of these values, indicators calculated from these values ​​using a formula, frequency conditions such as the frequency of the communication terminal 2 and the bandwidth of the resources used, and setting items of the communication terminal 2 that affect these indicators. The communication unit 11 outputs the acquired communication quality information to the data generation unit 13.

[0016] The acquisition unit 12 acquires spatial information within a predetermined area in which the communication terminal 2 moves. The spatial information is, for example, information (movement information) such as the position, speed, and direction of the communication terminal 2 and surrounding objects present within the predetermined area in which the user carrying the communication terminal 2 moves. The surrounding objects refer to stationary objects and moving objects installed within the predetermined area. The position, speed, and direction information can be acquired, for example, from a GPS (Global Positioning System) receiver. In addition, this information can be acquired from images captured by an externally fixed camera or point cloud information detected by LiDAR.

[0017] The data generation unit 13 acquires the communication quality (such as the above-mentioned throughput and RSSI) output from the communication unit 11 and the spatial information acquired by the acquisition unit 12. The data generation unit 13 classifies the acquired spatial information based on at least one of the number of communication terminals and the number of surrounding objects (hereinafter referred to as "environmental information"). For example, the spatial information is classified according to the number of communication terminals 2 present in a predetermined area, such as 1, 2, 3, .... Note that the environmental information may be set based on conditions other than the number of communication terminals and the number of surrounding objects.

[0018] The data generation unit 13 generates training data to be used for model generation based on each piece of spatial information classified by environmental information. "Training data" is data used when executing machine learning to build a prediction model for predicting future communication quality of the communication terminal 2. The training data refers to, for example, data indicating the relationship between information on the position, speed, and direction of the communication terminal 2 and surrounding objects and communication quality.

[0019] That is, the data generating unit 13 classifies the spatial information into a plurality of pieces of spatial information according to the environmental information, and generates training data for each piece of classified spatial information based on each piece of classified spatial information and communication quality.

[0020] The data generator 13 also extracts communication quality features from the training data. The communication quality features include the distribution, average value, and variance of data indicating communication quality, such as the average throughput value. The data generator 13 outputs the communication quality features to the estimation unit 15.

[0021] The model generation unit 14 acquires the training data generated by the data generation unit 13, performs machine learning using the training data, and constructs a prediction model. The prediction model outputs target parameters used to predict the communication quality of the communication terminal 2 in the future (for example, 0.5 seconds later) by inputting the communication quality when the communication unit 11 communicates with the communication terminal 2 and spatial information at that time. The "target parameters" are indicators for evaluating the communication quality when the communication terminal 2 performs wireless communication.

[0022] The model generation unit 14 constructs a prediction model for each of the spatial information classified into multiple categories according to the above-mentioned environmental information. For example, a model is constructed according to the number of communication terminals 2 present in the space. That is, the model generation unit 14 performs machine learning using training data and constructs a prediction model that outputs a target parameter, which is an index of communication quality, for each of the classified spatial information. The model generation unit 14 constructs multiple prediction models. The model generation unit 14 provides the constructed prediction models to the selection unit 16.

[0023] The estimation unit 15 estimates spatial information and environmental information within a predetermined area around the communication terminal 2 based on the communication quality features extracted by the data generation unit 13. As described above, the "spatial information" is information such as the position, speed, and direction of the communication terminal 2 present within the predetermined area, and the position, speed, and direction of surrounding objects. The "environmental information" includes information on the number of communication terminals 2 present within the predetermined area and the number of surrounding objects.

[0024] That is, the communication terminal 2 and surrounding objects present within a predetermined area may not be recognized using sensor information from a camera or LiDAR, etc. For example, when the communication terminal 2 is moving in an area that is a blind spot of the camera, or when a clear image captured by the camera cannot be obtained due to changes in the amount of light between day and night, the communication terminal 2 and surrounding objects may not be recognized with high accuracy.

[0025] The estimation unit 15 estimates spatial information and environmental information within a predetermined area around the communication terminal 2 by analyzing the characteristics of communication quality acquired from the data generation unit 13. The estimation unit 15 estimates the presence of the communication terminal 2 and surrounding objects that cannot be acquired by the acquisition unit 12, based on the characteristics of communication quality.

[0026] That is, the characteristics of communication quality, such as the distribution, average value, and variance of wireless communication quality, vary depending on environmental changes such as the number of communication terminals 2 present in the space and the number of surrounding objects. The estimation unit 15 extracts these characteristics and estimates current environmental information based on the similarity with communication quality characteristics extracted in the past. For example, if the average wireless communication throughput is 200 Mbps, the number of communication terminals is estimated to be "1" and the number of surrounding objects is estimated to be "1".

[0027] That is, the estimation unit 15 estimates environmental information within a predetermined area based on the communication quality during communication with the communication terminal 2. The estimation unit 15 extracts communication quality features from the communication quality included in the training data, and estimates environmental information based on the features.

[0028] Based on the environmental information estimated by the estimation unit 15, the selection unit 16 selects a prediction model that matches the environmental information estimated by the estimation unit from among the multiple prediction models constructed by the model generation unit 14, i.e., from each prediction model constructed based on spatial information classified by environmental information.

[0029] For example, if the estimation unit 15 estimates that the number of communication terminals 2 present in a predetermined area is N, a prediction model constructed with the number of communication terminals 2 set to N is selected. A prediction model may be selected based on factors other than the number of communication terminals 2, such as the number of peripheral objects present in the predetermined area. That is, the selection unit 16 selects a prediction model that corresponds to the estimated environmental information from among multiple prediction models, based on the environmental information estimated by the estimation unit 15. By selecting an appropriate prediction model, a prediction model that is suited to the environmental information can be selected, making it possible to predict communication quality with high accuracy.

[0030] When the communication unit 11 performs wireless communication with the communication terminal 2, the utilization unit 17 inputs the communication quality information and spatial information output from the data generation unit 13 into the prediction model selected by the selection unit 16, and acquires target parameters. The utilization unit 17 predicts the communication quality of the communication terminal 2 in the future (for example, 0.5 seconds later) based on the target parameters.

[0031] That is, when communicating with the communication terminal 2, the utilization unit 17 inputs the communication quality acquired by the communication unit 11 and the spatial information acquired by the acquisition unit 12 into the prediction model selected by the selection unit 16 to acquire target parameters, and predicts the communication quality of the communication terminal.

[0032] The utilization unit 17 controls the application or the wireless communication device based on the predicted communication quality. The application control is, for example, bit rate control and buffer control in a video transmission application, and the wireless communication device control is, for example, handover and resource control. By performing these controls, the communication quality can be improved in advance before it deteriorates.

[0033] Next, the processing procedure of the communication quality prediction device according to this embodiment configured as above will be described with reference to the flowchart shown in FIG.

[0034] 2, the communication unit 11 performs wireless communication with the communication terminal 2 and acquires the communication quality at this time. As described above, the communication quality is, for example, the throughput, the received signal power, the received signal strength indicator (RSSI), etc.

[0035] In step S12, the acquisition unit 12 acquires spatial information within a predetermined area around the communication terminal 2 detected by the sensor 3 such as a camera or LiDAR.

[0036] In step S13, the data generation unit 13 classifies the spatial information into a plurality of categories according to the environmental information. As described above, the environmental information is information about the number of communication terminals 2 or peripheral objects present in a predetermined area. For example, the spatial information is classified as spatial information when there is one communication terminal 2 present in the predetermined area, spatial information when there are two communication terminals 2 present, and so on.

[0037] In step S14, the data generating unit 13 generates training data in which each of the classified spatial information is associated with communication quality.

[0038] In step S15, the model generation unit 14 performs machine learning based on the training data for each of the classified spatial information to construct multiple prediction models. For example, multiple prediction models are constructed, such as a prediction model when the number of communication terminals 2 in a predetermined area is one, a prediction model when the number is two, etc.

[0039] In step S16, the estimation unit 15 estimates environmental information based on the communication quality output from the data generation unit 13 when wireless communication is being performed between the communication terminal 2 and the communication unit 11. For example, the estimation unit 15 estimates the number of communication terminals or the number of surrounding objects within a predetermined area. As described above, the characteristics of communication quality when wireless communication is being performed change depending on environmental information such as the number of communication terminals 2 and the number of surrounding objects. The estimation unit 15 estimates the environmental information by extracting these characteristics.

[0040] In step S17, based on the result of estimation in the process of step S15, the selection unit 16 selects an appropriate prediction model from among the multiple prediction models generated by the model generation unit 14. For example, if the estimation unit 15 estimates that the number of communication terminals 2 present in a predetermined area is "two," the selection unit 16 selects a prediction model constructed using training data when the number of communication terminals 2 is "two."

[0041] In step S18, the utilization unit 17 inputs the communication quality and spatial information into the selected prediction model, and outputs target parameters to be used for predicting the communication quality.

[0042] In step S19, the utilization unit 17 predicts the communication quality of the communication terminal 2 in the future (for example, 0.5 seconds later) using the target parameters.

[0043] Next, with reference to Figures 3A to 3C, Figures 4A to 4C, Figures 5A to 5C, and Figures 6A to 6C, we will explain an example of estimating the number of communication terminals 2 (environmental information) within a specified area based on changes in communication quality (here, throughput).

[0044] Figure 3A is a diagram showing the path of movement of a single robot A carrying a communication terminal as it moves within an area, Figure 3B is a graph showing a histogram of throughput over a certain period of time, and Figure 3C is a graph showing the change in throughput versus the position (y coordinate) of robot A.

[0045] As shown in Figure 3A, a base station is set at point p1 within the area, and robot A carrying a communication terminal moves from point p2 to point p3. The average throughput measurement result at this time was 218.3 [Mbps], with a standard deviation of 80.1 [Mbps]. The throughput at this time changes as shown in the histogram in Figure 3B, and the throughput changes with respect to the y coordinate of robot A as shown in the graph in Figure 3C.

[0046] Figure 4A is a diagram showing the path of movement when robot A carries a communication terminal and robot B moves through space while blocking the space between robot A and base station p1, Figure 4B is a graph showing a histogram of throughput over a certain period of time, and Figure 4C is a graph showing the change in throughput versus the position (y coordinate) of robot A.

[0047] As shown in Figure 4A, robot A carrying a communication terminal moves from point p2 to p3, and robot B moves from point p4 to p5. The average throughput measurement result at this time was 212.8 [Mbps], with a standard deviation of 81.4 [Mbps]. The throughput at this time changes as shown in the histogram in Figure 4B, and the throughput changes with respect to the y coordinate of robot A as shown in the graph in Figure 4C.

[0048] Figure 5A shows the path of movement when robot A carries a communication terminal and two robots B and C move through space, each shielding robot A from base station p1; Figure 5B is a graph showing a histogram of throughput over a certain period of time; and Figure 5C is a graph showing the change in throughput versus the position (y coordinate) of robot A.

[0049] As shown in Figure 5A, robot A carrying a communication terminal moves from point p2 to p3, robot B moves from point p4 to p5, and robot C moves from point p6 to p7. The average throughput measurement result at this time was 182.4 [Mbps], with a standard deviation of 82.2 [Mbps]. The throughput at this time changes as shown in the histogram in Figure 5B, and the throughput changes with respect to the y coordinate of robot A as shown in the graph in Figure 5C.

[0050] Figure 6A is a diagram showing the path of movement when robot A carries a communication terminal and three robots B, C, and D move through space while blocking the space between robot A and base station p1; Figure 6B is a graph showing a histogram of throughput over a certain period of time; and Figure 6C is a graph showing the change in throughput versus the position (y coordinate) of robot A.

[0051] As shown in Figure 6A, robot A carrying a communication terminal moves from point p2 to p3, robot B moves from point p4 to p5, robot C moves from point p6 to p7, and robot D moves from point p8 to p9. The average throughput measurement result at this time was 128.8 [Mbps], with a standard deviation of 75.1 [Mbps]. The throughput at this time changes as shown in the histogram in Figure 6B, and changes with respect to the y coordinate of robot A as shown in the graph in Figure 6C.

[0052] As can be seen from the above figures, as the number of robots increases, the number of times throughput drops due to blockage between the communication terminal and the base station increases, and the average throughput tends to decrease. Furthermore, it can be seen that the number of times near 100 Mbps in the throughput histograms (see Figures 3B, 4B, 5B, and 6B) increases. Therefore, based on the average throughput and the data rate near 100 Mbps, it is possible to obtain environmental information including the number of robots present in a specified area, i.e., the number of communication terminals and surrounding objects.

[0053] 7 is a graph showing the relationship between the prediction error and the cumulative distribution function (CDF), where curve s1 represents the predicted value of throughput in the past, curve s2 represents the predicted value of throughput based on physical space information, and curve s3 represents the predicted value of throughput using both.

[0054] A prediction model appropriate for the environmental information was selected from the throughput information, and physical space information such as the robot's position was obtained and predictions were made based on data detected by sensors 3 such as cameras and LiDAR. Compared to the prediction results based on conventional throughput, a 57% improvement was observed in the median graph prediction error of each curve shown in Figure 7.

[0055] As described above, the wireless communication device 1 according to this embodiment includes a communication unit 11 that performs wireless communication with the communication terminal 2 and acquires the communication quality of the communication terminal 2; an acquisition unit 12 that acquires spatial information including movement information of the communication terminal 2 and surrounding objects within a predetermined area; a data generation unit 13 that classifies the spatial information into multiple pieces of spatial information according to environmental information and generates training data for each piece of classified spatial information based on each piece of classified spatial information and the communication quality; a model generation unit 14 that performs machine learning using the training data and constructs a prediction model of communication quality for each piece of classified spatial information; an estimation unit 15 that estimates environmental information within the predetermined area based on the communication quality during communication with the communication terminal 2; a selection unit 16 that selects a prediction model corresponding to the estimated environmental information from multiple prediction models based on the estimated environmental information; and a utilization unit 17 that inputs the communication quality acquired by the communication unit 11 and the spatial information acquired by the acquisition unit 12 into the prediction model selected by the selection unit 16 during communication with the communication terminal 2 to predict the communication quality of the communication terminal 2.

[0056] In this embodiment, multiple prediction models are constructed for each piece of environmental information, and when wireless communication is performed between the communication terminal 2 and the communication unit 11, the environmental information at that time is estimated, and an appropriate prediction model is selected from the multiple prediction models based on the estimation result. The selected prediction model is then used to predict future communication quality for the communication terminal 2, so that communication quality can be predicted with high accuracy even when surrounding objects move or the number of communication terminals changes.

[0057] 3A, 4A, 5A, and 6A, as the number of robots around robot A carrying a communication terminal 2 increases, communication quality changes significantly. In this embodiment, a prediction model is selected taking into account the number of communication terminals 2 (environmental information), making it possible to predict communication quality with high accuracy without being affected by environmental information such as the number of communication terminals 2 or surrounding objects. Therefore, service quality can be improved by executing application control and wireless system control based on the communication quality prediction results.

[0058] In this embodiment, by predicting communication quality only when necessary, it is possible to effectively utilize computational resources. Furthermore, in this embodiment, the estimation unit 15 extracts communication quality features from the communication quality included in the training data and estimates environmental information based on the features, so that an appropriate prediction model can be selected, enabling highly accurate prediction of communication quality.

[0059] In this embodiment, in places where an unspecified number of people enter and exit, such as an event venue or a conference room, it is possible to predict future fluctuations in the communication quality of a wireless terminal carried by a user, which may fluctuate depending on the presence and movements of people around the user, and notify the user or the system.

[0060] In this embodiment, since communication quality can be predicted with high accuracy, for example, when communication quality is low, control can be executed to switch access points and base stations. Furthermore, by switching to a beamforming method, stable communication quality can be ensured.

[0061] In this embodiment, in an environment where multiple automated guided robots (AMRs: Autonomous Mobile Robots) are moving around, such as at a site such as a logistics center or a factory, it is possible to predict future fluctuations in the communication quality of wireless terminals carried by workers and robots, and notify the user and the system.

[0062] In this embodiment, by predicting communication quality, for example, when communication quality is high, it is possible to reduce power consumption by efficiently transferring high-load data such as video and log data acquired by the robot to the server.When communication quality is low, the above-mentioned data transfer is not performed, and proactive control is performed such as ensuring essential communications such as sending and receiving emergency stop signals to the robot, thereby enabling stable use of robotized logistics centers or factories.

[0063] Furthermore, for example, when performing remote component assembly or inspection work, or when playing games using VR (Virtual Reality) goggles, communication quality may vary depending on people or objects in the vicinity. In this embodiment, even when communication quality varies due to people or objects in the vicinity, it is possible to predict the communication quality of the communication terminal and notify the user or the system.

[0064] By predicting communication quality and proactively controlling it, for example, by increasing the video resolution or frame rate when the quality is high and decreasing the resolution or frame rate when the quality is low, it becomes possible to use VR goggle services stably. Furthermore, by further providing a means for a person or system to proactively control when the predicted communication quality is notified, it becomes possible to avoid problems caused by deterioration of communication quality in advance.

[0065] The wireless communication device 1 of the present embodiment described above can be, for example, a general-purpose computer system including a CPU (Central Processing Unit, processor) 901, a memory 902, a storage 903 (HDD: Hard Disk Drive, SSD: Solid State Drive), a communication device 904, an input device 905, and an output device 906, as shown in Fig. 8. The memory 902 and the storage 903 are storage devices. In this computer system, the CPU 901 executes a predetermined program loaded on the memory 902, thereby realizing each function of the wireless communication device 1.

[0066] The wireless communication device 1 may be implemented by one computer or by multiple computers, and may also be a virtual machine implemented on a computer.

[0067] The program for the wireless communication device 1 can be stored in a computer-readable recording medium such as an HDD, an SSD, a Universal Serial Bus (USB) memory, a Compact Disc (CD), or a Digital Versatile Disc (DVD), or can be distributed via a network. The computer-readable recording medium is, for example, a non-transitory recording medium.

[0068] The present disclosure is not limited to the above-described embodiments, and various modifications are possible within the scope of the present disclosure.

[0069] REFERENCE SIGNS LIST 1 wireless communication device 2 communication terminal 3 sensor 11 communication unit 12 acquisition unit 13 data generation unit 14 model generation unit 15 estimation unit 16 selection unit 17 utilization unit

Claims

1. A wireless communication device comprising: a communication unit that performs wireless communication with a communication terminal and acquires communication quality of the communication terminal; an acquisition unit that acquires spatial information including movement information of the communication terminal and surrounding objects within a specified area; a data generation unit that classifies the spatial information into multiple pieces of spatial information according to environmental information and generates training data for each piece of classified spatial information based on each piece of classified spatial information and the communication quality; a model generation unit that performs machine learning using the training data and builds a prediction model of communication quality for each piece of classified spatial information; an estimation unit that, when communicating with the communication terminal, estimates environmental information within the specified area based on the communication quality; a selection unit that, based on the estimated environmental information, selects a prediction model corresponding to the estimated environmental information from multiple prediction models; and a utilization unit that, when communicating with the communication terminal, inputs the communication quality acquired by the communication unit and the spatial information acquired by the acquisition unit into the prediction model selected by the selection unit to predict the communication quality of the communication terminal.

2. The wireless communication device according to claim 1, wherein the estimation unit extracts communication quality characteristics from the communication quality included in the training data and estimates the environmental information based on the characteristics.

3. The wireless communication device according to claim 2, wherein the environmental information is at least one of the number of communication terminals present within the specified area and the number of surrounding objects present within the specified area.

4. A communication quality prediction method performed by a wireless communication device, comprising: performing wireless communication with a communication terminal, acquiring the communication quality of the communication terminal; acquiring spatial information including movement information of the communication terminal and surrounding objects within a specified area; classifying the spatial information into multiple pieces of spatial information according to environmental information; generating training data for each piece of classified spatial information based on each piece of classified spatial information and the communication quality; performing machine learning using the training data to construct a communication quality prediction model for each piece of classified spatial information; estimating environmental information within the specified area based on the communication quality during communication with the communication terminal; selecting a prediction model corresponding to the estimated environmental information from multiple prediction models based on the estimated environmental information; and inputting the communication quality acquired by the communication unit and the spatial information acquired by the acquisition unit into the prediction model selected by the selection unit during communication with the communication terminal to predict the communication quality of the communication terminal.

Images