Information processing method, information processing device, and information processing system

Abstract

An information processing device 1 comprises: a plurality of first acquisition units 11 that respectively acquire, from a plurality of information collection means, physical world information of a wireless communication area; a second acquisition unit 12 that acquires, from a wireless communication apparatus in the wireless communication area, communication quality information about wireless communications; a plurality of generating / updating units 15 that respectively generate or update each model for respectively predicting, from the physical world information, the communication quality of wireless communications in the wireless communication area; a plurality of processing units 16 that respectively input the current physical world information acquired from each among the plurality of information collection means into each of the models and output a calculation result of each model; an evaluation unit 17 that evaluates the performance of each of the models by comparing the calculation result of each model with the current communication quality of wireless communications actually determined without using the models, and determines, on the basis of the performance evaluation of each of the models, a selection method or synthesis method for the calculation result of each of the models; and an integration unit 18 that integrates the calculation results of the models on the basis of the selection method or synthesis method.

Description

[Technical Field]

[0001] The present invention relates to an information processing method, an information processing device, and an information processing system. [Background technology]

[0002] The realization of IoT (Internet of Things), in which various devices are connected to the Internet, has been progressing, and various devices such as automobiles, drones, and construction machinery vehicles are being connected wirelessly. However, while wireless communication is used for various purposes, there is a problem in that wireless communication does not necessarily meet the communication quality requirements for some services.

[0003] In particular, the movement of wireless communication terminals and surrounding objects can change the direction of antenna directivity and the propagation environment, which can affect communication quality. Therefore, terminals equipped with wireless communication functions need to take measures to deal with fluctuations in communication quality due to their own state and the state of the surrounding area. Therefore, technologies that can predict communication quality and implement prescribed measures before services and systems are affected are being studied.

[0004] Non-Patent Document 1 discloses a technology that uses depth camera images to predict, through machine learning, the degradation of communication quality when a pedestrian blocks a wireless communication channel. It is known that communication quality can be predicted by acquiring terminal information and object information around the terminal from the camera in this way and inputting the information into a communication quality prediction model that has been generated in advance through machine learning, and the usefulness of information obtained from cameras is attracting attention.

[0005] Furthermore, Non-Patent Document 2 discloses a technique for improving the accuracy of predicting communication quality by using multiple sensors to supplement information on blind spots and the like. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] T. Nishio, et al., “Proactive received power prediction using machine learning and depth images for mmWave networks”, IEEE Journal on Selected Areas in Communications 37.11, 2019, p.2413-p.2427 [Non-patent document 2] Kahoko TAKAHASHI and 2 others, “Received Power Prediction using Multiple RGB cameras in Indoor Environment for Wireless LAN Systems”, 2020 International Conference on Emerging Technologies for Communications (ICETC,(2020)), p.1-p.3 Summary of the Invention [Problem to be solved by the invention]

[0007] However, because the vast amount of physical world information obtained from all multiple sensors was input into a single communication quality prediction model (a single neural network) to predict communication quality, the communication quality prediction model was large in scale, resulting in high training costs. In addition, there was a risk that the prediction results would be degraded if any of the input values contained abnormal values.

[0008] The present invention has been made in consideration of the above circumstances, and an object of the present invention is to provide a technology that can construct a small-scale model to which each piece of physical world information collected from multiple information collection means, such as cameras, can be input, and that can achieve performance equivalent to or better than that of a single model that inputs all of the physical world information from all of the multiple information collection means. [Means for solving the problem]

[0009] An information processing method according to one aspect of the present invention is an information processing method performed by an information processing device, comprising the steps of: a first step of acquiring physical world information, which is physical information of a wireless communication area, from a plurality of information collecting means installed at different locations; a second step of acquiring wireless communication quality information from wireless communication devices in the wireless communication area; a third step of using the respective physical world information and the communication quality information to generate or update models for predicting the communication quality of wireless communication in the wireless communication area from the respective physical world information, or models for controlling wireless communication or wireless communication devices in the wireless communication area from the respective physical world information; a fourth step of inputting the current physical world information acquired from the plurality of information collecting means into each of the models and outputting calculation results of each of the models; a fifth step of evaluating the performance of each of the models by comparing the calculation results of each of the models with the current communication quality of wireless communication or the effect of controlling the wireless communication or wireless communication device actually obtained without using each of the models, and determining a method of selecting or combining the calculation results of each of the models based on the performance evaluation of each of the models; and a sixth step of integrating the calculation results of each of the models based on the selection method or the combination method.

[0010] an information processing device according to one embodiment of the present invention, comprising: a plurality of first acquisition units that acquire physical world information, which is physical information of a wireless communication area, from a plurality of information collection means installed at different locations; a second acquisition unit that acquires wireless communication quality information from wireless communication devices in the wireless communication area; a plurality of generation / update units that use the physical world information and the communication quality information to generate or update models for predicting the communication quality of wireless communication in the wireless communication area from each piece of physical world information, or models for controlling wireless communication or wireless communication devices in the wireless communication area from each piece of physical world information; a plurality of processing units that input the current physical world information acquired from the plurality of information collection means to each of the models and output calculation results of each model; an evaluation unit that evaluates the performance of each model by comparing the calculation results of each model with the current wireless communication quality or the effect of controlling the wireless communication or wireless communication device actually obtained without using each of the models, and determines a selection method or combination method for the calculation results of each model based on the performance evaluation of each model; and an integration unit that integrates the calculation results of each model based on the selection method or combination method.

[0011] An information processing system according to one aspect of the present invention includes a plurality of information collection means, a wireless communication device, and an information processing device. The information processing device includes a plurality of first acquisition units that acquire physical world information, which is physical information of a wireless communication area, from a plurality of information collection means installed at different locations, a second acquisition unit that acquires communication quality information of wireless communication from wireless communication devices in the wireless communication area, and a model that uses the respective physical world information and the communication quality information to predict the communication quality of wireless communication in the wireless communication area from each piece of physical world information, or a model that predicts wireless communication or wireless communication devices in the wireless communication area from each piece of physical world information. a plurality of processing units that input each of the current physical world information acquired from the plurality of information collecting means to each of the models and output the calculation results of each of the models; an evaluation unit that evaluates the performance of each of the models by comparing the calculation results of each of the models with the communication quality of the current wireless communication or the effect of controlling the wireless communication or the wireless communication device that is actually obtained without using each of the models, and determines a method for selecting or combining the calculation results of each of the models based on the performance evaluation of each of the models; and an integration unit that integrates the calculation results of each of the models based on the selection method or the combination method. [Effects of the Invention]

[0012] According to the present invention, it is possible to construct a small-scale model that can input each piece of physical world information collected from multiple information collection means, and it is possible to provide a technology that can achieve performance that is equivalent to or better than that of a single model that inputs all of the physical world information from all of the multiple information collection means. [Brief explanation of the drawings]

[0013] [Figure 1] FIG. 1 is a diagram showing the overall configuration of an information processing system according to the first embodiment. [Figure 2] FIG. 2 is a diagram showing the processing flow of the model generation and update method. [Figure 3] FIG. 3 is a diagram showing a processing flow of the communication quality prediction and communication control method. [Figure 4] Figure 4 shows an aerial view of the indoor laboratory. [Figure 5] FIG. 5 is a diagram showing an error in the predicted value of communication quality. [Figure 6] FIG. 6 is a diagram showing a functional block configuration of an information processing device according to the second embodiment. [Figure 7] FIG. 7 is a diagram showing, for each sensor, a two-dimensional histogram of the number of selected models that output correct predicted values in each area. [Figure 8] FIG. 8 is a diagram showing an error in the predicted value of communication quality. [Figure 9] FIG. 9 is a diagram showing a functional block configuration of an information processing device according to the third embodiment. [Figure 10] FIG. 10 is a diagram illustrating a hardware configuration of an information processing device. DETAILED DESCRIPTION OF THE INVENTION

[0014] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the description of the drawings, the same parts are designated by the same reference numerals and the description thereof will be omitted.

[0015] [Summary of the Invention] The present invention is characterized by generating and updating a model for each information gathering means, such as a camera, and integrating the calculation results of each model based on the performance evaluation of each model. This makes it possible to build small-scale models that can input each piece of physical world information gathered from each information gathering means, and provides a technology that can achieve performance equivalent to or better than that of a single model that inputs all physical world information from all multiple information gathering means.

[0016] [First embodiment] 1 is a diagram showing the overall configuration of an information processing system according to the first embodiment. The information processing system includes an information processing device 1 that performs communication quality prediction and communication control using a learning model, a plurality of sensors 2, a base station 3, and a mobile communication terminal 4.

[0017] The multiple sensors 2 are devices installed at different positions near the wireless communication area AR, and collect physical world information, which is physical information of the wireless communication area AR, from different angles. The physical world information includes dynamic objects such as a person H or a mobile communication terminal 4, and static objects such as pillars. The sensors 2 are an example of information collection means. In addition to the sensors 2, RGB cameras, seismic intensity cameras, monochrome cameras, LiDAR (Light Detection and Ranging), radar, etc. may also be used.

[0018] The base station 3 is a device that forms a wireless communication area AR. The mobile communication terminal 4 is an autonomous communication device that performs wireless communication R with the base station 3 in the wireless communication area AR. The base station 3 performs wireless communication R with one or more mobile communication terminals 4. The base station 3 and the mobile communication terminal 4 are examples of wireless communication devices. The base station 3 may be a Wi-Fi router. The mobile communication terminal 4 may be a fixed terminal.

[0019] 1, the information processing device 1 includes a plurality of first acquisition units 11, a second acquisition unit 12, a storage unit 13, an extraction unit 14, a plurality of generation / update units 15, a plurality of processing units 16, an evaluation unit 17, and an integration unit 18. The first acquisition unit 11, the generation / update unit 15, and the processing unit 16 are provided for each sensor 2.

[0020] A first acquisition unit 11 is provided for each sensor 2. The multiple first acquisition units 11 each have a function of acquiring physical world information, which is physical information of the wireless communication area AR, from multiple sensors 2 installed at different positions. As described above, the physical world information is dynamic objects and static objects.

[0021] The second acquisition unit 12 has a function of acquiring, from the base station 3 within the wireless communication area AR, communication quality information of the wireless communication R between the base station 3 and the mobile communication terminal 4. The communication quality information includes throughput, received signal strength indicator (RSSI), etc.

[0022] The storage unit 13 has a function of storing each piece of physical world information acquired by each of the plurality of first acquisition units 11 and the communication quality information acquired by the second acquisition unit 12.

[0023] The extraction unit 14 has a function of reading out each piece of physical world information and communication quality information from the storage unit 13 and extracting information necessary for generating and updating a communication quality prediction model and a communication quality control model, which will be described later.

[0024] A generating and updating unit 15 is also provided for each sensor 2. The multiple generating and updating units 15 have a function of using the respective physical world information and communication quality information to generate and update respective communication quality prediction models for predicting the communication quality of wireless communication R in the wireless communication area AR from the respective physical world information. In this way, a communication quality prediction model is generated and updated for each sensor.

[0025] Furthermore, the plurality of generating / updating units 15 have a function of using the respective physical world information and communication quality information to generate / update each control model for controlling wireless communication R in the wireless communication area AR and the mobile communication terminal 4 from each piece of physical world information, thereby generating / updating a control model for each sensor.

[0026] A processing unit 16 is also provided for each sensor 2. The processing unit 16 is a functional unit that uses the communication quality prediction model and control model generated and updated by the generation and update unit 15. The multiple processing units 16 have the function of inputting the current physical world information acquired from each of the multiple sensors 2 to each communication quality prediction model, and outputting the calculation results of each communication quality prediction model. The processing unit 16 also has the function of performing similar processing on the control model. As a result, calculation results for each sensor (each model) are obtained.

[0027] The evaluation unit 17 has a function of evaluating the performance of each communication quality prediction model by comparing the calculation result of each communication quality prediction model with the communication quality of the current wireless communication R that is actually obtained without using each communication quality prediction model, and determining a method for selecting and combining the calculation results of each communication quality prediction model based on the performance evaluation of each communication quality prediction model.

[0028] The evaluation unit 17 also has a function of evaluating the performance of each control model by comparing the calculation results of each control model with the effects of wireless communication R and control methods of the mobile communication terminal 4 actually performed without using each control model, and determining a method for selecting and combining the calculation results of each control model based on the performance evaluation of each control model.

[0029] The evaluation unit 17 also has a function of comparing the current performance evaluation of the communication quality prediction model / control model with past performance evaluations to determine whether or not there is a change in the performance of the model.

[0030] The integration unit 18 has a function of integrating the calculation results of each communication quality prediction model based on the selection method or synthesis method determined by the evaluation unit 17. The integration unit 18 also has a function of performing similar processing on the control model. As a result, the calculation results of each model are integrated based on the performance evaluation of each model.

[0031] FIG. 2 is a diagram showing the processing flow of the model generation and update method.

[0032] Step S101; First, the multiple first acquisition units 11 each acquire physical world information (dynamic objects, static objects), which is physical information of the wireless communication area AR, from multiple sensors 2 installed at different positions near the wireless communication area AR, and store each piece of physical world information in the memory unit 13.

[0033] Step S102; Next, the second acquisition unit 12 acquires communication quality information (throughput, received signal strength, etc.) of the wireless communication R conducted between the base station 3 in the wireless communication area AR and one or more mobile communication terminals 4, and stores the communication quality information in the memory unit 13.

[0034] Step S103; Next, the extraction unit 14 reads out each piece of physical world information and communication quality information from the storage unit 13, and extracts information necessary for generating and updating each communication quality prediction model. When generating and updating a control model, the extraction unit 14 extracts information necessary for generating and updating each control model.

[0035] Step S104; Finally, the plurality of generation / update units 15 use the extracted information to generate and update respective communication quality prediction models for predicting the communication quality of wireless communication R in the wireless communication area AR from each piece of physical world information.

[0036] For example, the first generating and updating unit 15 executes a process of associating physical world information of a wireless communication area AR acquired from sensor #1 at a certain timing with the communication quality of wireless communication R acquired from a base station 3 in the same wireless communication area AR at the same or similar timing. Then, the first generating and updating unit 15 repeats this process in chronological order, repeatedly learning the communication quality corresponding to various changing states of the physical world. As a result, the first generating and updating unit 15 generates and updates a communication quality prediction model that can predict future communication quality corresponding to sensor #1.

[0037] When generating and updating a control model, the plurality of generating and updating units 15 use the information extracted in step S103 to generate and update each control model for controlling the wireless communication R and the mobile communication terminal 4 in the wireless communication area AR from each piece of physical world information. As a result, the first generating and updating unit 15 generates and updates each control model that can output a control method for the wireless communication R and the mobile communication terminal 4 corresponding to the sensor #1. The control method is handover control, transmission power control, etc.

[0038] FIG. 3 is a diagram showing a processing flow of the communication quality prediction and communication control method.

[0039] Step S201; First, the plurality of first acquisition units 11 acquire, from the plurality of sensors 2, current physical world information, which is physical information of the wireless communication area AR.

[0040] Step S202; Next, the multiple processing units 16 input the current physical world information acquired from the multiple sensors 2 to the communication quality prediction models, respectively, and output the calculation results of the communication quality prediction models. For example, a first processing unit 16 inputs the current physical world information acquired from sensor #1 to the communication quality prediction model corresponding to that sensor #1, and acquires a predicted value of the future communication quality of wireless communication R corresponding to the current physical world information.

[0041] When a control model is used, the multiple processing units 16 input the current physical world information acquired from the multiple sensors 2 to each control model, and output the calculation results of each control model. For example, a first processing unit 16 inputs the current physical world information acquired from sensor #1 to the control model corresponding to that sensor #1, and acquires the wireless communication R and the control method for the mobile communication terminal 4 corresponding to the current physical world information.

[0042] Step S203; Next, the evaluation unit 17 evaluates the performance of each communication quality prediction model by comparing the calculation result of each communication quality prediction model (predicted value of future communication quality of wireless communication R) with the current communication quality of wireless communication R that is actually obtained without using each communication quality prediction model. After that, the evaluation unit 17 determines a method for selecting and combining the calculation results of each communication quality prediction model based on the performance evaluation of each communication quality prediction model.

[0043] For example, if the performance evaluation value of a certain communication quality prediction model is a low evaluation value, the evaluation unit 17 decides not to select the predicted value of that communication quality prediction model, and detects a failure or abnormality in the sensor 2 corresponding to that communication quality prediction model.

[0044] When a control model is used, the evaluation unit 17 evaluates the performance of each control model by comparing the calculation results (control methods of wireless communication R and mobile communication terminal 4) of each control model with the effects of wireless communication R and control methods of mobile communication terminal 4 that are actually performed without using each control model. After that, the evaluation unit 17 determines a method for selecting and combining the calculation results of each control model based on the performance evaluation of each control model.

[0045] Step S204; Finally, the integrating unit 18 integrates the calculation results of each communication quality prediction model based on the determined selection method or synthesis method. At this time, the integrating unit 18 may integrate the calculation results of each communication quality prediction model using a neural network, normalization, or other statistical method. When a control model is used, the integrating unit 18 also processes the control model in the same manner.

[0046] In step S203, the evaluation unit 17 can compare the current performance evaluation of the communication quality prediction model / control model with the past performance evaluation of the same communication quality prediction model / control model to determine whether the performance of the communication quality prediction model / control model has improved or deteriorated. For example, if the current performance evaluation value of a certain communication quality prediction model has deteriorated compared with the past performance evaluation value, the evaluation unit 17 determines not to select the communication quality prediction model or its predicted value, and detects a failure or abnormality, etc., of the sensor 2 corresponding to the communication quality prediction model.

[0047] Here, the effect of the integrating unit 18 will be explained. An aerial view of the indoor laboratory used is shown in Fig. 4. Five sensors 2a to 2e were installed at different positions in the room, and two base stations 3a and 3b were also installed. Although the total number of base stations 3 could be one, two were prepared to ensure that wireless communication was not disabled due to obstructions to the wireless communication path caused by pillars or the like.

[0048] While a mobile communication terminal 4 was moving through a corridor in the same room, physical world information (camera image data) was captured from different angles by five sensors 2a to 2e, and the communication quality of wireless communication between the mobile communication terminal 4 and two base stations 3a and 3b was predicted by the information processing device 1. Note that in Figure 4, the figure eight line indicates the movement route of the mobile communication terminal 4.

[0049] Figure 5 shows the error when predicting throughput one second later. The horizontal axis, "Pre Err Throughput," represents throughput. The vertical axis, "CDF," represents the error in the predicted value. The solid line represents the error in the predicted value when physical world information obtained from all five sensors 2a to 2e is input into a single communication quality prediction model. The dashed line represents the error in the predicted value when the predicted value of the communication quality prediction model that outputs the predicted value closest to the correct value is integrated from the results of the five communication quality prediction models corresponding to each of the sensors 2a to 2e. Figure 5 suggests that selecting the optimal sensor can produce better results than using all sensors.

[0050] According to this embodiment, a model is generated and updated for each sensor, and the calculation results of each model are integrated based on the performance evaluation of each model. This makes it possible to construct a small-scale model that can input each piece of physical world information collected from multiple sensors, and to obtain performance that is equal to or better than that of a single model that inputs all physical world information from all multiple sensors.

[0051] Furthermore, according to this embodiment, since multiple models are used, it is possible to reduce the calculation cost for communication quality prediction, etc., even when the wireless communication area is wide. Furthermore, since a model is generated and updated for each sensor, it is possible to generate and update a model targeting only each sensor, and it is possible to easily add a model as sensors are added, and it is possible to quantitatively evaluate whether a sensor is effectively contributing to communication quality prediction, etc.

[0052] [Second embodiment] FIG. 6 is a diagram showing a functional block configuration of an information processing device according to the second embodiment.

[0053] The information processing device 1 further includes a selection unit 19 in addition to the components of the first embodiment. The selection unit 19 has a function of selecting a communication quality prediction model to be used. The selection unit 19 also has a function of performing similar processing on a control model.

[0054] Next, the operation of the information processing device 1 will be described.

[0055] Between the above step S201 and step S202, the selection unit 19 selects one or more communication quality prediction models for performing communication quality prediction from among the plurality of communication quality prediction models. When a control model is used, the selection unit 19 selects one or more control models for outputting a control method from among the plurality of control models. In other words, the selection unit 19 selects one or more processing units 16 from among the plurality of processing units 16.

[0056] Possible methods for selecting a model include selecting a model corresponding to the sensor 2 that detected the mobile communication terminal 4 based on physical world information, or empirically selecting a model corresponding to the sensor 2 that has high accuracy in outputting predictions and control methods for the position of the mobile communication terminal 4.

[0057] The selection unit 19 can also update the selection method based on the model evaluation results performed by the evaluation unit 17. In this case, the evaluation unit 17 determines the selection method and synthesis method of the calculation results of each model described in the first embodiment, and generates input selection determination information indicating whether or not to input the physical world information from each first acquisition unit 11 to each processing unit 16, and transmits the input selection determination information to the selection unit 19. The selection unit 19 determines whether or not to input the physical world information from each first acquisition unit 11 to each processing unit 16, based on the input selection determination information.

[0058] Here, the effect of the selection unit 19 will be explained. Fig. 7 shows the number of selected sensors corresponding to the communication quality prediction model that output the predicted value closest to the correct value for each area obtained by dividing the indoor laboratory shown in Fig. 4 into a grid, expressed as a two-dimensional histogram for each sensor. Figs. 7(a) to (e) correspond to sensors 2a to 2e shown in Fig. 4.

[0059] For example, from Fig. 7(a), it can be seen that the histogram of the southwest area is high in the communication quality prediction model corresponding to sensor 2a. Similarly, it can be seen that the histograms of the areas close to the positions of the corresponding sensors are high in the communication quality prediction models corresponding to the other sensors 2b to 2e. Therefore, by the selector 19 selecting a communication quality prediction model corresponding to a sensor with high prediction accuracy according to the area, it is possible to further improve the prediction accuracy of communication quality.

[0060] Figure 8 shows the error when predicting throughput one second later when a communication quality prediction model is selected for each area. The solid and dashed lines are the same as those shown in Figure 5. The dashed and dotted line shows the error in the predicted value when a communication quality prediction model is selected for each area. It can be seen that the same level of accuracy can be obtained as when all sensors are used.

[0061] According to this embodiment, a model is selected in advance, which eliminates the need to perform calculations in all processing units 16, thereby reducing calculation costs. In the first embodiment, the evaluation unit 17 selects a predetermined calculation result from the calculation results in all processing units 16, but in the second embodiment, the processing units 16 to be used are selected in advance, which reduces calculation costs more than in the first embodiment.

[0062] [Third embodiment] FIG. 9 is a diagram showing a functional block configuration of an information processing device according to the third embodiment.

[0063] The information processing device 1 further includes a filter unit 20 in addition to the components of the first embodiment. The filter unit 20 has a function of extracting only information necessary for generating and updating each communication quality prediction model from the physical world information acquired by each of the plurality of first acquisition units 11 and the communication quality information acquired by the second acquisition unit 12, and storing the extracted information in the storage unit 13. The filter unit 20 also has a function of similarly processing the control model.

[0064] Next, the operation of the information processing device 1 will be described.

[0065] In the above step S101, the filter unit 20 extracts only the information required by the communication quality prediction model to predict the communication quality of wireless communication from each piece of physical world information acquired by the multiple first acquisition units 11, and stores only that information in the memory unit 13.

[0066] When using a control model, the filter unit 20 extracts only the information required by the control model to output a control method for wireless communication or a mobile communication terminal from each piece of physical world information acquired by the multiple first acquisition units 11, and stores only that information in the memory unit 13.

[0067] In the above step S102, the filter unit 20 extracts only the information required by the communication quality prediction model to predict the communication quality of wireless communication from the communication quality information acquired by the second acquisition unit 12, and stores only that information in the memory unit 13.

[0068] When a control model is used, the filter unit 20 extracts only the information required by the control model to output a control method for wireless communication or a mobile communication terminal from the communication quality information acquired by the second acquisition unit 12, and stores only that information in the memory unit 13.

[0069] According to this embodiment, only the information necessary for generating and updating a model is extracted from the physical world information and communication quality information and stored in the storage unit 13, thereby reducing the amount of information transmitted to the storage unit 13. If the storage unit 13 is located outside the information processing device 1, the processing load on the network, data server, and cloud can be reduced.

[0070] [others] The present invention is not limited to the above-described embodiments. Numerous modifications of the present invention are possible within the scope of the gist of the present invention. It is also possible to combine the first to third embodiments.

[0071] The information processing device 1 of the present embodiment described above can be realized, for example, by using a general-purpose computer system including a CPU 901, a memory 902, a storage 903, a communication device 904, an input device 905, and an output device 906, as shown in Fig. 10. The memory 902 and the storage 903 are storage devices. In the computer system, the CPU 901 executes a predetermined program loaded onto the memory 902, thereby realizing each function of the information processing device 1.

[0072] The information processing device 1 may be implemented by one computer. The information processing device 1 may be implemented by multiple computers. The information processing device 1 may be a virtual machine implemented on a computer. The program for the information processing device 1 may be stored in a computer-readable recording medium such as an HDD, SSD, USB memory, CD, or DVD. The program for the information processing device 1 may also be distributed via a communication network. [Explanation of symbols]

[0073] 1: Information processing device 11:First acquisition part 12:Second acquisition part 13: Storage part 14:Extraction part 15: Generation / update section 16: Processing section 17: Evaluation section 18: Integration Department 19: Selection section 20: Filter section 901:CPU 902: Memory 903:Storage 904:Communication equipment 905: Input device 906: Output device

Claims

1. In an information processing method performed by an information processing device, a first step of acquiring physical world information, which is physical information of a wireless communication area, from a plurality of information collection means installed at different locations; a second step of acquiring communication quality information of wireless communication from wireless communication devices in the wireless communication area; a third step of generating or updating, using the respective physical world information and the communication quality information, respective models for predicting communication quality of wireless communication in the wireless communication area from the respective physical world information, or respective models for controlling wireless communication or wireless communication devices in the wireless communication area from the respective physical world information; a fourth step of inputting each piece of current physical world information acquired from each of the plurality of information collecting means into each of the models and outputting a calculation result of each model; a fifth step of evaluating the performance of each model by comparing the calculation results of each model with the communication quality of the current wireless communication or the effect of control of the wireless communication or the wireless communication device that is actually obtained without using each model, and determining a method for selecting or combining the calculation results of each model based on the performance evaluation of each model; a sixth step of integrating the calculation results of each of the models based on the selection method or the synthesis method; An information processing method that performs the above.

2. Between the third step and the fourth step, The information processing method according to claim 1, further comprising the step of selecting a model to be used.

3. Between the second step and the third step, The information processing method according to claim 1 , further comprising the step of extracting only information necessary for generating or updating each of the models from the respective physical world information and the communication quality information.

4. In the fifth step, The information processing method according to claim 1 , wherein the performance evaluation of the model is compared with a past performance evaluation of the model to determine whether or not there has been a change in the performance of the model.

5. a plurality of first acquisition units that acquire physical world information, which is physical information of a wireless communication area, from a plurality of information collection means installed at different positions; a second acquisition unit that acquires communication quality information of wireless communication from a wireless communication device in the wireless communication area; a plurality of generation / update units that use the physical world information and the communication quality information to generate or update models for predicting communication quality of wireless communication in the wireless communication area from each piece of physical world information, or models for controlling wireless communication or wireless communication devices in the wireless communication area from each piece of physical world information; a plurality of processing units that input the current physical world information acquired from the plurality of information collecting means to the respective models and output the calculation results of the respective models; an evaluation unit that evaluates the performance of each of the models by comparing a calculation result of each of the models with a communication quality of a current wireless communication or an effect of control of a wireless communication or a wireless communication device that is actually obtained without using each of the models, and determines a method for selecting or combining the calculation results of each of the models based on the performance evaluation of each of the models; an integration unit that integrates the calculation results of each of the models based on the selection method or the synthesis method; An information processing device comprising:

6. An information processing system including a plurality of information collection means, a wireless communication device, and an information processing device, The information processing device includes: a plurality of first acquisition units that acquire physical world information, which is physical information of a wireless communication area, from a plurality of information collection means installed at different positions; a second acquisition unit that acquires communication quality information of wireless communication from a wireless communication device in the wireless communication area; a plurality of generation / update units that use the physical world information and the communication quality information to generate or update models for predicting communication quality of wireless communication in the wireless communication area from each piece of physical world information, or models for controlling wireless communication or wireless communication devices in the wireless communication area from each piece of physical world information; a plurality of processing units that input the current physical world information acquired from the plurality of information collecting means to the respective models and output the calculation results of the respective models; an evaluation unit that evaluates the performance of each of the models by comparing a calculation result of each of the models with a communication quality of a current wireless communication or an effect of control of a wireless communication or a wireless communication device that is actually obtained without using each of the models, and determines a method for selecting or combining the calculation results of each of the models based on the performance evaluation of each of the models; an integration unit that integrates the calculation results of each of the models based on the selection method or the synthesis method; An information processing system comprising:

Images