Wireless communication device and wireless communication method

Abstract

This wireless communication device comprises: a wireless communication unit (11) that acquires wireless communication information about a communication terminal (2); a first acquisition unit (12) that acquires terminal space information, which is space information about the communication terminal (2); a second acquisition unit (13) that acquires peripheral space information, which is space information about a surrounding object around the communication terminal (2); a data generation unit (14) that generates training data on the basis of the wireless communication information, the terminal space information, and a plurality of pieces of the peripheral space information; a model generation unit (15) that executes machine learning using the training data and constructs a prediction model that outputs a target parameter indicating the communication quality of the communication terminal (2); and a use unit (16) that inputs the terminal space information and the peripheral space information to the prediction model and acquires the target parameter to execute wireless communication.

Description

Wireless communication device and wireless communication method

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

[0002] When wireless communication is performed using a communication device such as a wireless communication terminal, communication quality changes depending on changes in the environment, such as the movement of objects in the vicinity. Communication quality refers to the communication capacity, throughput, etc. required by an application. Changes in the environment may cause the service provided by the communication device or the communication quality required by the system to be insufficient.

[0003] For example, fifth-generation communications (5G communications) such as IEEE802.11ad and cellular communications use high frequencies in the millimeter range, so blocking caused by obstacles between the transmitter and receiver in wireless communications has a significant impact on communication quality.

[0004] If the communication quality of a communication device can be predicted in advance, it may be possible to take measures before a service or system is affected. Non-Patent Document 1 discloses a method for predicting communication quality (obstruction state prediction) using spatial information obtained by a camera or LiDAR in a situation where the surrounding environment, such as obstructions, is fixed and a robot carrying a communication terminal is moving.

[0005] Non-Patent Documents 2 and 3 disclose a method for predicting communication quality (RSSI prediction) 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 people are walking between the wireless communication terminal and the AP, blocking each other.

[0006] 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.

[0007] The communication quality of a communication terminal varies depending on the number of surrounding objects and the movement conditions (speed, direction) of the communication terminal. However, Non-Patent Documents 1 to 3 do not mention generating parameters related to wireless communication of the communication terminal using a prediction model that takes into account changes in the number of surrounding objects and changes in the movement conditions.

[0008] The present disclosure has been made in consideration of the above circumstances, and its purpose is to provide a wireless communication device and a wireless communication method that are capable of performing wireless communication by setting an appropriate communication environment in response to changes in the number of objects around the communication terminal and the movement conditions, even when these changes cause changes in communication quality.

[0009] A wireless communication device of one aspect of the present disclosure includes a wireless communication unit that acquires wireless communication information of a communication terminal, a first acquisition unit that acquires terminal space information, which is spatial information of the communication terminal, a second acquisition unit that acquires surrounding space information, which is spatial information of surrounding objects present around the communication terminal, a data generation unit that generates training data based on the wireless communication information, the terminal space information, and a plurality of pieces of the surrounding space information, a model generation unit that performs machine learning using the training data and constructs a predictive model that outputs target parameters that indicate the communication quality of the communication terminal, and a utilization unit that inputs the terminal space information and the surrounding space information into the predictive model, acquires the target parameters, and performs wireless communication.

[0010] A wireless communication method of one aspect of the present disclosure includes a wireless communication unit acquiring wireless communication information of a communication terminal, a first acquisition unit acquiring terminal space information which is spatial information of the communication terminal, a second acquisition unit acquiring surrounding space information which is spatial information of surrounding objects present around the communication terminal, a data generation unit generating training data based on the wireless communication information, the terminal space information, and a plurality of pieces of the surrounding space information, a model generation unit performing machine learning using the training data to construct a predictive model that outputs a target parameter indicating the communication quality of the communication terminal, and a utilization unit inputting the terminal space information and the surrounding space information into the predictive model, acquiring the target parameter, and performing wireless communication.

[0011] According to the present disclosure, even if the number of objects around the communication terminal and the movement conditions change and the communication quality changes, it is possible to set an appropriate communication environment in response to these changes and perform wireless communication.

[0012] FIG. 1 is a block diagram showing the configuration of a wireless communication device and its peripheral devices according to an embodiment. FIG. 2 is a flowchart showing the processing procedure of the wireless communication device according to an embodiment. FIG. 3 is an explanatory diagram showing a situation in which a robot A equipped with a communication terminal and two robots B and C as peripheral objects move within a movement area. FIG. 4 is an explanatory diagram showing the prediction results of communication quality according to a conventional example and a first embodiment. FIG. 5 is an explanatory diagram showing the prediction results of communication quality according to a first embodiment and a second embodiment. FIG. 6 is a block diagram showing the hardware configuration of this embodiment.

[0013] 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 wireless communication device 1 is also capable of wireless or wired communication with a 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 wireless communication unit 11, a first acquisition unit 12, a second acquisition unit 13, a data generation unit 14, a model generation unit 15, and a utilization unit 16.

[0015] The wireless communication unit 11 performs wireless communication with the communication terminal 2. The communication terminal 2 is, for example, a wireless communication device such as a smartphone. The wireless communication unit 11 acquires wireless communication information of the communication terminal 2. As described below, the wireless communication unit 11 may be installed in the communication terminal 2. The wireless communication information refers to the throughput of the communication terminal 2, received signal power, signal-to-noise power ratio, signal-to-interference-plus-noise power ratio, received signal strength indication (RSSI), received signal reference 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 resource used, and setting items of the communication terminal 2 that affect these indicators. The wireless communication unit 11 outputs the acquired wireless communication information to the data generation unit 14.

[0016] The first acquisition unit 12 acquires spatial information of the communication terminal 2 (hereinafter referred to as "terminal spatial information"). The terminal spatial information is information such as the position, movement speed, and movement direction of the communication terminal 2. The first acquisition unit 12 acquires the terminal spatial information of the communication terminal 2 to be predicted from, for example, a Global Positioning System (GPS) mounted on the communication terminal 2. The first acquisition unit 12 can also acquire the terminal spatial information of the communication terminal 2 to be predicted from image information, point cloud information, etc. acquired by a sensor 3 (camera, LiDAR, etc.). The first acquisition unit 12 outputs the acquired various types of terminal spatial information to the data generation unit 14.

[0017] The second acquisition unit 13 acquires spatial information (hereinafter referred to as "surrounding space information") of objects (hereinafter referred to as "surrounding objects") present in the vicinity of the communication terminal 2 from the image information and point cloud information detected by the sensor 3. Surrounding objects refer to, for example, moving objects such as people, automobiles, and automated guided robots, as well as stationary objects such as fixedly installed buildings, utility poles, and shelves. The surrounding space information includes the position, movement speed, and movement direction of the surrounding objects. The second acquisition unit 13 outputs the acquired surrounding space information to the data generation unit 14. Note that, hereinafter, the terminal space information and surrounding space information may be collectively referred to as "spatial information."

[0018] The data generation unit 14 acquires the terminal space information output from the first acquisition unit 12, the surrounding space information output from the second acquisition unit 13, and the wireless communication information output from the wireless communication unit 11. The data generation unit 14 generates training data based on the terminal space information, the surrounding space information, and the wireless communication information. The "training data" is data used when performing machine learning to build a prediction model for predicting the communication quality of the communication terminal 2. The training data refers to, for example, the position, speed, and direction of the communication terminal 2, the position, speed, and direction of surrounding objects, and the throughput, RSSI, etc. of the communication terminal 2.

[0019] The data generation unit 14 sets in advance the maximum number of peripheral objects present around the communication terminal 2. For example, if the communication terminal 2 is mounted on a robot that moves within a predetermined area, the user inputs the maximum number of objects present within the area to the data generation unit 14. The data generation unit 14 generates training data by setting each object as a peripheral object and setting the maximum number of peripheral space information as input. For example, if there is a possibility that (N-1) peripheral objects exist within the area in which the communication terminal 2 moves, the maximum number of peripheral objects is set to (N-1). In other words, the total maximum number of communication terminal 2 and peripheral objects is set to N. The data generation unit 14 creates in advance training data that ensures the input of spatial information when no peripheral objects exist (terminal spatial information) and spatial information when there are 1 to (N-1) peripheral objects (terminal spatial information and peripheral spatial information), i.e., a total of N pieces of spatial information.

[0020] Furthermore, the data generation unit 14 generates training data in which the order of the surrounding space information to be output to the model generation unit 15 is changed. For example, if there are (N-1) surrounding objects around the communication terminal 2, there are "(N-1)! = (N-1) * (N-2) * ... * 1" possible orders. The data generation unit generates training data in (N-1)! possible orders and outputs it to the model generation unit 15.

[0021] For example, suppose that a mobile robot A is equipped with a communication terminal 2, and two other robots B and C exist in the area where robot A moves. That is, N=3. In this case, for example, training data is generated in the order of robot A's position → robot B's position → robot C's position based on the spatial information of robots A to C. Furthermore, the order of robots B and C is swapped, and training data is generated in the order of robot A's position → robot C's position → robot B's position. That is, training data with an order of "(N-1)!=2!" is generated. By creating (N-1)! ways of training data with the order of surrounding objects swapped, it becomes possible to input all surrounding objects without distinguishing them.

[0022] Furthermore, it is not necessary to generate training data for all "(N-1)!" possible sequences. For example, suppose that three other robots, B, C, and D, exist in the area in which robot A equipped with communication terminal 2 moves. In other words, N=4. In this case, if you want to input only robots B and C without distinguishing between them, you can generate and learn from two different sequences of training data: robot A → B → C → D, and robot A → C → B → D.

[0023] The model generation unit 15 learns the relationship between the terminal space information and the surrounding space information and the target parameters, which are indicators of the communication quality of wireless communication in the communication terminal 2 that is the prediction target or the control target, based on the training data generated by the data generation unit 14. The contents of the target parameters can be selected arbitrarily. The target parameters include the communication capacity and throughput required by the application. The target parameters include, for example, information on the movement of the base station and the remote terminal, control information in the OSI reference model from the physical layer to the application layer (e.g., mode, communication speed, communication destination, communication path, communication method), and control information on structures, metamaterials, and dielectrics that affect the propagation environment (e.g., control information related to position, movement, and settings).

[0024] The model generation unit 15 uses an arbitrary machine learning algorithm based on the training data to learn the relationship between the terminal space information, surrounding space information, and wireless communication information by machine learning, and constructs a prediction model that predicts target parameters of wireless communication based on the terminal space information and surrounding space information.

[0025] As described above, the number of surrounding objects is set to (N-1), and machine learning is performed by inputting (N-1)! sets of training data, so a prediction model can be constructed that can handle cases where the number of surrounding objects is "0 to (N-1)." For example, as in the example described above, if robot A is equipped with communication terminal 2 and two robots B and C exist in the area where robot A moves, the communication quality of communication terminal 2 can be predicted with high accuracy even if the order of inputting the spatial information of robot B and robot C is reversed.

[0026] For example, consider a case where a camera (sensor 3) installed in an area where robot A moves captures images of robots B and C, which are assumed to have identical appearances, and the positions of robots B and C are obtained by methods such as extracting the bounding boxes of each robot B and C. In this case, robots B and C may cross paths, making it impossible to distinguish between robots B and C from the image captured by the camera. Even in such a case, a predictive model is constructed by performing machine learning using training data in which the spatial information of robots B and C is reversed, making it possible to predict the communication quality of the communication terminal 2 possessed by robot A with high accuracy. Note that in the above example, the spatial information is the positions of robots A, B, and C, but the spatial information may also include the movement speed and movement direction of each robot A, B, and C.

[0027] That is, the model generation unit 15 constructs a prediction model that predicts the communication quality of the communication terminal 2, covering cases where there are no peripheral objects around the communication terminal 2 and cases where there are 1 to (N-1) peripheral objects around the communication terminal 2. Furthermore, if the number of peripheral objects is less than (N-1), the missing data can be input as "0" (null data). For example, if N is set to 10 and the number of peripheral objects is 6, the total number of communication terminal 2 and peripheral objects is 7, so the remaining 3 can be input as "0" to the prediction model.

[0028] The utilization unit 16 inputs the terminal space information and the surrounding space information into the prediction model generated by the model generation unit 15, thereby performing wireless communication.

[0029] 1 may be mounted on the communication terminal 2. In this case, no communication occurs between the wireless communication unit 11 and the communication terminal 2. Furthermore, in addition to the wireless communication unit 11, at least one of the first acquisition unit 12, the second acquisition unit 13, the data generation unit 14, and the model generation unit 15 may be mounted on the communication terminal 2. In such a configuration, when a user is wirelessly communicating using a video viewing service on the communication terminal 2 (e.g., a smartphone), communication quality information can be acquired in the communication terminal 2, and a model for predicting communication quality can be constructed within the communication terminal 2.

[0030] Furthermore, at least one of the data generating unit 14 and the model generating unit 15 shown in FIG. 1 may be configured to be installed on a server or in the cloud.

[0031] Next, the operation of the wireless communication device 1 according to this embodiment will be described with reference to a flowchart shown in FIG.

[0032] 2, the data generation unit 14 sets the number (N-1) of peripheral objects present in the area in which the communication terminal 2 moves. The data generation unit 14 reserves N storage areas corresponding to the communication terminal 2 and the (N-1) peripheral objects in a memory (not shown) or the like.

[0033] In step S12, the wireless communication unit 11 communicates with the communication terminal 2 whose communication quality is to be predicted, and acquires wireless communication information of the communication terminal 2. As described above, the wireless communication information includes the throughput of the communication terminal 2, the received signal power, and the like.

[0034] In step S13, the first acquisition unit 12 acquires terminal space information of the communication terminal 2 based on GPS information or detection data by a sensor 3 such as a camera or LiDAR mounted on the communication terminal 2. As described above, the terminal space information is information such as the position, movement speed, and movement direction of the communication terminal 2.

[0035] In step S14, the second acquisition unit 13 acquires surrounding space information based on the detection data by the sensor 3. As described above, the surrounding space information is information such as the position, movement speed, and movement direction of surrounding objects present around the communication terminal 2.

[0036] In step S15, the data generator 14 generates training data based on the wireless communication information, the terminal space information, and the plurality of pieces of surrounding space information. As described above, the data generator 14 generates training data in which the input order of the plurality of pieces of surrounding space information is changed. Specifically, (N-1)! sets of training data are generated.

[0037] In step S16 , the model generation unit 15 performs machine learning based on the training data generated by the data generation unit 14 and constructs a prediction model for predicting target parameters of wireless communication by the communication terminal 2 .

[0038] In step S17, the utilization unit 16 inputs terminal space information and surrounding space information related to the communication terminal 2 to be predicted into the prediction model constructed in the processing of step S16, and acquires target parameters.

[0039] In step S18, the utilization unit 16 executes wireless communication using the target parameters.

[0040] Next, the results of an experiment in which communication quality was predicted using the wireless communication device 1 according to this embodiment will be described. In this experiment, a situation is assumed in which a robot A equipped with a communication terminal and two robots B and C are moving along routes indicated by R1, R2, and R3 within a movement area D1, as shown in FIG. 3. The communication terminal equipped on robot A is also assumed to be in wireless communication with base station F1. The model generation unit 15 constructs a prediction model assuming the presence of robot A equipped with a communication terminal and two robots B and C as peripheral objects. That is, the prediction model is constructed with N=3.

[0041] Below, we will explain the results of predicting the communication quality of the communication terminal 2 using this prediction model. In the first example, a prediction model is constructed assuming that the number of surrounding objects is (N-1), and this prediction model is used to generate target parameters for the communication terminal 2. In the second example, the number of surrounding objects is (N-1), and further, the input order of the spatial information is changed to construct a prediction model using (N-1)! sets of training data, and this prediction model is used to generate target parameters for the communication terminal 2.

[0042] [Prediction results of the first embodiment] As shown in Figure 3, as three robots A, B, and C move within the movement area D1, the communication terminal mounted on robot A is affected by the position, movement direction, and movement speed of each robot A, B, and C when performing wireless communication with base station F1.

[0043] In the first embodiment, the number of surrounding objects is set to (N-1), so it is possible to generate parameters using one common prediction model for the case where there is one surrounding object (when there are a total of two robots, robot A and robot B), and for the case where there are two surrounding objects (when there are a total of three robots, robot A, robot B, and robot C).

[0044] When the communication terminal mounted on robot A is communicating with base station F1, a communication quality prediction (e.g., throughput prediction) of the communication terminal was performed using a prediction model constructed by the model generation unit 15 of the wireless communication device 1 according to this embodiment. Specifically, for a case in which robot A and one robot B present in its vicinity (i.e., two robots A and B) are moving, communication quality was predicted using a prediction model in which the maximum number of surrounding objects was two (i.e., N=3). The prediction results are shown in FIG. 4(c). Note that since there is only one robot B, a "0" (null data) was entered for the one surrounding object that was left blank.

[0045] Furthermore, the communication quality of the communication terminal was predicted using a conventional prediction model. Specifically, the communication quality was predicted using a prediction model with one surrounding object. The prediction results are shown in Figure 4(a).

[0046] As shown in Figure 4(a), when the conventional prediction model was used, the MAE (Mean Absolute Error) of the throughput prediction of the communication terminal was "13.43." Note that a smaller MAE is a numerical value indicating higher accuracy of the prediction. As shown in Figure 4(c), when the prediction model constructed by the model generation unit 15 according to this embodiment was used, the MAE of the throughput prediction of the communication terminal was "13.75," which is almost the same MAE as in (a).

[0047] Furthermore, in the case where robot A and two robots B and C located nearby (i.e., three robots A, B, and C) are moving, communication quality predictions were performed using a prediction model in which the maximum number of nearby objects is two (the prediction model of this embodiment) and a prediction model in which the number of nearby objects is two (a conventional prediction model). The prediction results are shown in Figures 4(d) and (b), respectively.

[0048] As shown in Fig. 4(b), when the conventional prediction model was used, the MAE of the throughput prediction of the communication terminal was 27.61. Also, as shown in Fig. 4(d), when the prediction model of this embodiment was used, the MAE of the throughput prediction of the communication terminal was 27.14, which was almost the same as (b).

[0049] That is, by predicting the communication quality of a communication terminal using a prediction model constructed using the model generation unit 15 according to this embodiment, it is possible to obtain prediction results that are almost identical to those obtained using prediction models constructed with the number of peripheral objects fixed at one or two. This confirms that the prediction model of this embodiment can generate appropriate parameters that respond to changes in the number of peripheral objects and movement conditions of the communication terminal, even when these changes occur. Furthermore, similar results can be obtained even when the number of peripheral robots is three or more (i.e., N = 4 or more). Thus, by setting the maximum number of peripheral objects in advance, it is possible to generate appropriate parameters that are independent of the number of peripheral objects.

[0050] [Prediction Results of the Second Example] Next, the prediction results of the second example will be described. In the first example described above, the prediction results were explained when a prediction model was constructed with the number of surrounding objects set to (N-1). In the second example, the prediction results were explained when a prediction model constructed by performing machine learning using training data in which the order of surrounding space information was changed. The environment in which each of robots A, B, and C moves is assumed to be the same as that shown in FIG. 3 described above.

[0051] 5A and 5B are diagrams showing prediction results using a prediction model when the order of the surrounding space information for robots B and C is not interchanged (first embodiment described above) and when it is interchanged (second embodiment). Figures 5A and 5B show prediction results when the prediction model shown in the first embodiment described above is used, with Figure 5A showing the prediction result when the spatial information for robots A, B, and C is input to the prediction model in the order of robots A, C, and B, respectively.

[0052] In Fig. 5(a), the MAE is "27.29", and in Fig. 5(b), the MAE is "33.12". In other words, by changing the input order of the spatial information, the MAE value increases, and the prediction accuracy of the communication quality decreases.

[0053] Figures 5(c) and (d) show the prediction results when the prediction model shown in the second embodiment is used. Figure 5(c) shows the prediction results when spatial information is input to the prediction model in the order of robots A → B → C, and Figure 5(d) shows the prediction results when spatial information is input to the prediction model in the order of robots A → C → B.

[0054] In Figure 5(c), the MAE is "26.42," and in Figure 5(d), the MAE is "26.33." Even when the input order of the spatial information is swapped, the MAE value remains almost the same. In other words, it was confirmed that the prediction accuracy of communication quality remains almost unchanged even when the input order of the spatial information of robots B and C is swapped. Therefore, by using a prediction model constructed by performing machine learning using training data in which the order of surrounding spatial information is swapped, it becomes possible to set appropriate target parameters that are independent of the number of surrounding objects and the input order.

[0055] As such, the wireless communication device 1 of this embodiment comprises a wireless communication unit 11 that acquires wireless communication information of the communication terminal 2, a first acquisition unit 12 that acquires terminal space information, which is spatial information of the communication terminal 2, a second acquisition unit 13 that acquires surrounding space information, which is spatial information of surrounding objects present around the communication terminal 2, a data generation unit 14 that generates training data based on the wireless communication information, terminal space information, and multiple pieces of surrounding space information, a model generation unit 15 that performs machine learning using the training data and constructs a predictive model that outputs target parameters that indicate the communication quality of the communication terminal 2, and a utilization unit 16 that inputs the terminal space information and surrounding space information into the predictive model, acquires the target parameters, and performs wireless communication.

[0056] According to this embodiment, in an environment where multiple AMRs (Autonomous Mobile Robots) are moving at a site such as a logistics center or a factory, it is possible to predict future fluctuations in communication quality of the communication terminals 2 carried by workers or robots and notify the users and systems. Furthermore, since highly accurate prediction results of communication quality can be obtained without depending on the number of surrounding objects, the user can set appropriate target parameters to be used in the communication terminal and perform wireless communication without considering the number of surrounding objects present around the communication terminal 2 that is the target of prediction.

[0057] Furthermore, the data generator 14 sets a maximum number (N-1) of surrounding objects that affect the communication quality of the communication terminal 2, and generates training data using surrounding space information for this maximum number as input. For example, the maximum number can be set based on information such as the maximum number of users in a conference room, the number of people expected to attend an event, and the number of past participants. This makes it possible to predict communication quality with high accuracy even when the number of surrounding objects changes within a range below the maximum number. In other words, there is no need to build a prediction model for each number of surrounding objects, simplifying the configuration.

[0058] The model generation unit 15 executes machine learning corresponding to the combination of input orders of the training data by changing the input order of the plurality of pieces of surrounding space information included in the training data, thereby constructing a prediction model. Therefore, even if the order of the surrounding space information input to the prediction model is changed, it is possible to set appropriate target parameters according to these changes and execute wireless communication.

[0059] According to this embodiment, even if the communication quality of a communication terminal carried by a user in a place where an unspecified number of people come and go, such as an event venue or a conference room, fluctuates due to the presence or behavior of people in the vicinity, it is possible to predict the future communication quality of the communication terminal and notify the user or the system. For example, if a deterioration in communication quality is predicted, stable communication quality can be ensured by performing beamforming at the access point and base station early.

[0060] When the future communication quality of the communication terminal 2 is predicted to be high, power consumption can be reduced by efficiently transferring to the server high-load data, such as video and log data acquired by a robot equipped with the communication terminal 2. When the quality is predicted to be low, proactive control can be performed, such as not transferring the data as described above, and ensuring essential communications, such as sending and receiving emergency stop signals to the robot, thereby enabling the stable use of robotized logistics centers and factories.

[0061] By using the wireless communication device 1 according to the present embodiment, when the communication quality fluctuates due to the presence or movement of people and objects in the vicinity, for example, in a situation where a user is remotely transferring skills such as assembly or inspection work, or in a situation where a user is using VR goggles for a game console, it is possible to predict the communication quality of the communication terminal and notify the user or the system. For example, by proactively controlling the system to increase the resolution and frame rate of the video displayed in VR when the quality is high, and to decrease the resolution and frame rate when the quality is low, stable use of the VR goggles can be ensured.

[0062] Furthermore, by adding a function that enables the user or the system to proactively respond to notifications of fluctuations in communication quality, it is possible to avoid problems caused by deterioration in communication quality in advance.

[0063] Furthermore, proactive control is possible, such as performing large-volume communications such as video images captured by the robot's camera and log data when it is predicted that the communication quality of the communication terminal 2 will improve, and not performing such large-volume communications when it is predicted that the communication quality will deteriorate.

[0064] The wireless communication device 1 of the present embodiment described above may 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. 6. 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.

[0065] 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.

[0066] 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.

[0067] The present disclosure is not limited to the above-described embodiments, and various modifications are possible within the scope of the present disclosure.

[0068] REFERENCE SIGNS LIST 1 wireless communication device 2 communication terminal 3 sensor 11 wireless communication unit 12 first acquisition unit 13 second acquisition unit 14 data generation unit 15 model generation unit 16 utilization unit D1 movement area

Claims

1. A wireless communication device comprising: a wireless communication unit that acquires wireless communication information of a communication terminal; a first acquisition unit that acquires terminal spatial information which is spatial information of the communication terminal; a second acquisition unit that acquires peripheral spatial information which is spatial information of peripheral objects existing around the communication terminal; a data generation unit that generates training data based on the wireless communication information, the terminal spatial information, and a plurality of the peripheral spatial information; a model generation unit that performs machine learning using the training data and constructs a prediction model that outputs a target parameter indicating the communication quality of the communication terminal; and a utilization unit that inputs the terminal spatial information and the peripheral spatial information to the prediction model, acquires the target parameter, and performs wireless communication.

2. The wireless communication device according to claim 1, wherein the communication terminal moves within a predetermined area, and when the maximum number of objects that can exist within the area is input, the data generation unit generates the training data with each object as the peripheral object and the maximum number of the peripheral spatial information as input.

3. The wireless communication device according to claim 1 or 2, wherein the model generation unit constructs the prediction model by performing machine learning corresponding to combinations of input orders of the training data by swapping the input orders of the plurality of the peripheral spatial information included in the training data.

4. A wireless communication method, comprising: a wireless communication unit acquiring wireless communication information of a communication terminal; a first acquisition unit acquiring terminal spatial information which is spatial information of the communication terminal; a second acquisition unit acquiring peripheral spatial information which is spatial information of peripheral objects existing around the communication terminal; a data generation unit generating training data based on the wireless communication information, the terminal spatial information, and a plurality of the peripheral spatial information; a model generation unit performing machine learning using the training data and constructing a prediction model that outputs a target parameter indicating the communication quality of the communication terminal; and a utilization unit inputting the terminal spatial information and the peripheral spatial information to the prediction model, acquiring the target parameter, and performing wireless communication.

Images