Control device, program, and control method

Abstract

Provided is a control device comprising: a collection unit that collects, from some radio base stations among a plurality of radio base stations included in a base station group, communication data communicated between a radio base station and a user terminal; a model generation unit that, using the communication data collected by the collection unit, generates an artificial intelligence (AI) model to be used in communication between the radio base station and the user terminal; and a model provision unit that provides the AI model generated by the model generation unit to the plurality of radio base stations included in the base station group.

Description

Control device, program, and control method

[0001] The present invention relates to a control device, a program, and a control method.

[0002] Patent Document 1 states, "Training the ML model according to Embodiment A3 corresponds to the case where both the training of the ML models according to Embodiments A1 and A2 are performed (for example, in parallel, simultaneously, or at different times). That is, BS performs training of the ML model as shown in Figure 2 and obtains a trained ML model. On the other hand, UE performs training of the ML model as shown in Figure 4, obtains a trained ML model, and transmits information about the trained ML model to the BS." "BS may update (for example, fine-tune) the ML model it has trained based on the information about the ML model trained by the UE that it has received from the UE. Furthermore, BS may perform transfer learning of the ML model based on the information about the ML model trained by the UE that it has received from the UE. It is stated that BS may perform learning. Furthermore, BS may combine its own trained ML model with the ML model received from the UE (for example, it may use the output of the ML model received from the UE as input to its own trained ML model). In addition, BS may perform the above updates, transfer learning, etc., based on information from multiple trained ML models received from multiple communicating UEs. Patent Document 2 states, "A communication environment model is a model that takes environmental information, or environmental information and communication information, as input information and outputs communication information at the same time or in the future. Here, communication information refers to communication quality or control information that maximizes / minimizes parameters including communication quality. The communication environment model may be generated so that the output value of communication quality is highly accurate, or it may generate control information generated by reinforcement learning to maximize / minimize parameters including communication quality." "For example, when the communication environment model outputs received power at a future time, the received power at a future time is used as a parameter of the target information to form training data, and the communication environment model is generated so that the error between the predicted value of received power, which becomes the output of the communication environment model, and the actual received power is reduced."Alternatively, when the terminal device 102 is an autonomous mobile robot, in order to output the robot's X-coordinate velocity, Y-coordinate velocity, and rotation command so as to maximize received power, reinforcement learning is used to set a reward for maximizing received power, and a communication environment model is generated to output an X-coordinate velocity, a Y-coordinate velocity, and a rotation command that increase the reward. [Prior Art Documents] [Patent Documents] [Patent Document 1] International Publication No. WO 2022 / 208673 [Patent Document 2] International Publication No. WO 2022 / 079834.

[0003] In order to improve performance by applying an Artificial Intelligence (AI) model to a Radio Access Network (RAN), training data may be used. In general, efforts are being made to improve RAN performance by collecting Key Performance Indicators (KPIs) and the like related to RAN, and generating and updating an AI model using the collected KPIs and the like as training data. Application of such efforts to Open-RAN (O-RAN) is also being actively discussed.

[0004] For layers at Layer (L) 3 and above, a training format may be adopted in which KPIs and the like are collected in a central data center, a training data set is created in a batch, and an AI model is trained. However, in lower layers such as L1 and L2, the large data volume of the data set poses a problem. Furthermore, in the aforementioned training format, for lower layers such as L1 and L2, the transfer rate when transferring the data set to the central data center becomes a problem. On the other hand, in the training format where an AI model is trained at each distributed data center, computing resources of a Graphics Processing Unit (GPU) and a Central Processing Unit (CPU) for training the AI model must be deployed at each distributed data center. As a result, there is a risk that the return on investment in capital expenditure for communication infrastructure will decrease.

[0005] According to one embodiment of the present invention, a control device is provided. The control device may include a collection unit that collects communication data communicated between a wireless base station and a user terminal from some of the wireless base stations among a plurality of wireless base stations included in a base station group. The control device may include a model generation unit that generates an AI model for use in communication between the wireless base station and the user terminal using the communication data collected by the collection unit. The control device may include a model providing unit that provides the AI ​​model generated by the model generation unit to the plurality of wireless base stations included in the base station group.

[0006] In the control device, the collection unit may collect communication data transmitted between the malfunctioning wireless base station and the user terminal from among the plurality of wireless base stations to which the model providing unit has provided the AI ​​model, specifically from the wireless base station that has experienced a malfunction as a result of using the AI ​​model. In the control device, the model generation unit may update the AI ​​model using the communication data collected by the collection unit from the malfunctioning wireless base station. In the control device, the model providing unit may provide the AI ​​model updated by the model generation unit to the malfunctioning wireless base station.

[0007] In any of the control devices, the collection unit may collect the communication data from the wireless base station that has the highest volume of communication with the user terminal among the plurality of wireless base stations included in the base station group.

[0008] In any of the control devices, the collection unit may collect the communication data from the wireless base station with the lowest load among the plurality of wireless base stations included in the base station group.

[0009] Any of the control devices may include a group generation unit that groups the plurality of radio base stations based on the base station information of each of the plurality of radio base stations, and generates a plurality of base station groups, each containing a plurality of radio base stations.

[0010] In any of the control devices, the base station information may include information on the placement of the wireless base stations. In any of the control devices, the base station information may include information on the communication trends of the wireless base stations. In any of the control devices, the base station information may include the wireless base station density of the area including the wireless base station. In any of the control devices, the base station information may include information on the terminal trends of user terminals located within the wireless base station's service area.

[0011] Any of the control devices may include a group generation unit that groups the plurality of radio base stations based on the base station information of each of the plurality of radio base stations to generate a plurality of base station groups, each containing a plurality of radio base stations. In any of the control devices, the collection unit may collect communication data communicated between the malfunctioning radio base station and the user terminal from among the plurality of radio base stations to which the model provision unit has provided the AI ​​model, specifically from the radio base station that experienced a malfunction as a result of using the AI ​​model. In any of the control devices, the model generation unit may update the AI ​​model using the communication data collected by the collection unit from the malfunctioning radio base station. In any of the control devices, the model provision unit may provide the AI ​​model updated by the model generation unit to the malfunctioning radio base station. In any of the control devices, the group generation unit may move the radio base station to another base station group if, after the updated AI model has been provided to the malfunctioning radio base station by the model provision unit, the malfunctioning radio base station experiences another malfunction as a result of using the updated AI model.

[0012] In any of the control devices, the collection unit may collect the communication data from the radio base station that has the greatest overall correlation with the other radio base stations included in the base station group, among the plurality of radio base stations included in the base station group.

[0013] According to one embodiment of the present invention, a program is provided for causing a computer to function as one of the control devices.

[0014] According to one embodiment of the present invention, a control method is provided. The control method may include a collection step of collecting communication data communicated between a wireless base station and a user terminal from some of the wireless base stations among a plurality of wireless base stations included in a base station group. The control method may include a model generation step of generating an AI model for use in communication between the wireless base station and the user terminal using the communication data collected in the collection step. The control method may include a model provision step of providing the AI ​​model generated in the model generation step to the plurality of wireless base stations included in the base station group.

[0015] It should be noted that the above summary of the invention does not enumerate all the necessary features of the present invention. Furthermore, subcombinations of these features may also constitute an invention.

[0016] An example of the prior art is shown in general terms. An example of the prior art is shown in general terms. An example of the wireless communication system 10 is shown in general terms. An example of the wireless communication system 10 is shown in general terms. An example of the wireless communication system 10 is shown in general terms. An example of the functional configuration of the control device 100 is shown in general terms. An example of the processing flow by the control device 100 is shown in general terms. An example of the wireless communication system 10 is shown in general terms. An example of the hardware configuration of the computer 1200 that functions as the control device 100, information processing infrastructure 300, or management infrastructure 400 is shown in general terms.

[0017] The present invention will be described below through embodiments, but these embodiments are not intended to limit the scope of the claimed invention. Furthermore, not all combinations of features described in the embodiments are necessarily essential to the solution of the invention.

[0018] The control device 100 according to this embodiment generates base station groups by grouping multiple wireless base stations based on various similarities such as proximity of location, similarity of location attributes, and similarity of communication trends of user terminals providing wireless communication services. The control device 100 selects some of the wireless base stations included in the base station group as representatives of the base station group, uses the communication data of these wireless base stations as training data, and generates an AI model. The number of wireless base stations selected by the control device 100 may be adjusted as appropriate depending on the characteristics of the base station group.

[0019] The control device 100 applies the AI ​​model generated in this manner to all wireless base stations included in the base station group and evaluates the communication quality of each wireless base station. For wireless base stations included in the base station group where the application of the AI ​​model is thought to have degraded communication performance, the control device 100 acquires further communication data from that wireless base station to fine-tune the AI ​​model and provides the tuned AI model. In this way, the control device 100 reduces the amount of training data while providing a fine-tuned AI model to wireless base stations with problems, thereby achieving both efficient learning and improved communication quality.

[0020] The AI ​​model according to this embodiment may be any model that the control device 100 can use directly or indirectly to control the wireless communication system. For example, examples of AI models include a channel estimation model that estimates channel state information from a received signal, and a channel interpolation model that interpolates channel state information. Other examples of AI models include a transmission signal estimation model that estimates the transmission signal on the receiving side, a transmission signal restoration model that restores the transmission signal on the receiving side, a beamforming optimization model that selects or adapts to the optimal beam, a scheduling optimization model that optimizes resource allocation, and a power control model that optimizes transmission power, but these are just examples and are not limited to these.

[0021] Figure 1 schematically shows an example of the conventional technology. In the example shown in Figure 1, a learning device 800 is connected to a network 70 that includes multiple wireless base stations. Each of the multiple wireless base stations provides wireless communication services to user terminals 72 located within its coverage area. The learning device 800 collects communication data related to communication with the user terminals 72 from the multiple wireless base stations and generates an AI model using the collected communication data.

[0022] In the example shown in Figure 1, the learning device 800 collects communication data 91 from the wireless base station 201, communication data 92 from the wireless base station 202, communication data 93 from the wireless base station 203, and communication data 99 from the wireless base station 209. The learning device 800 uses the collected communication data to generate an AI model 80. The AI ​​model 80 is used for the wireless communication service of each wireless base station.

[0023] Here, if the data size of the communication data is large, the transmission load of the communication data increases, and the load on the network 70 increases. In particular, when the communication data is low-layer communication data such as L1 or L2, the data size of the communication data tends to be large, and the communication load on the network tends to become a problem. For this reason, with a learning device configured as shown in Figure 1, it may be difficult to collect low-layer communication data and generate an AI model for low-layer communication data.

[0024] Figure 2 schematically shows an example of the prior art. The example shown in Figure 2 will mainly be explained in terms of the differences from the example shown in Figure 1. In the example shown in Figure 2, multiple learning devices are arranged. In the example shown in Figure 2, one learning device is connected to each wireless base station. A learning device 801 is connected to wireless base station 201, and the learning device 801 generates an AI model 81 using communication data 91, and the AI ​​model 81 is provided to wireless base station 201. Similarly, a learning device 802 is connected to wireless base station 202, and the learning device 802 generates an AI model 82 using communication data 92, and the AI ​​model 82 is provided to wireless base station 202. Similarly, a learning device 803 is connected to wireless base station 203, and the learning device 803 generates an AI model 83 using communication data 93, and the AI ​​model 83 is provided to wireless base station 203. Similarly, a learning device 809 is connected to the wireless base station 209, the learning device 809 generates an AI model 89 using communication data 99, and the AI ​​model 89 is provided to the wireless base station 209.

[0025] In the example shown in Figure 2, communication data is transmitted directly from the wireless base station to the learning device, thus reducing the communication load on the entire network. However, this requires a large number of distributed server devices for the learning device, which is not practical from an investment efficiency standpoint.

[0026] Thus, when the data size of the communication data to be collected from each base station is large, if a centralized learning model is adopted to prioritize investment efficiency, the transmission load of the communication data becomes a problem. Conversely, if a distributed learning model is adopted to prioritize reducing the transmission load of the communication data, it leads to a decrease in investment efficiency.

[0027] Figure 3 schematically shows an example of a wireless communication system 10. In the example shown in Figure 3, the wireless communication system 10 comprises a control device 100, an information processing infrastructure 300, and a plurality of wireless base stations. In the example shown in Figure 3, the information processing infrastructure 300 and the plurality of wireless base stations from wireless base station 201 to wireless base station 209 are connected via a network 70.

[0028] The information processing infrastructure 300 may be a data center corresponding to a region. In the example shown in Figure 3, the information processing infrastructure 300 is located in region 30. The information processing infrastructure 300 does not have to be located in region 30. For example, the information processing infrastructure 300 may be located in a region adjacent to region 30. Multiple types of devices may be located in the information processing infrastructure 300. Each of the multiple wireless base stations may provide wireless communication services to user terminals 72 located within the coverage area. In the example shown in Figure 3, the control device 100 is located in the information processing infrastructure 300.

[0029] The control device 100 may be applied to AI-RAN. AI-RAN may include three types: "AI for RAN", "AI on RAN", and "AI and RAN". "AI for RAN" may be a technology that utilizes AI and machine learning techniques to improve the frequency utilization efficiency and performance of existing RANs, or to realize automation of base station operations and power saving. "AI for RAN" is expected to optimize processing at all layers, such as channel estimation and scheduling processing performed by each cell of the RAN, optimize cooperation between RAN cells, and improve the equipment utilization rate of base stations. "AI on RAN" may be a technology that utilizes the computing infrastructure of the base station to provide highly immediate services to users and devices around the base station with low latency. "AI on RAN" enables the deployment of AI and machine learning technology applications at the network edge via RAN, thereby facilitating the creation of new industries and solutions that leverage low latency and confidentiality. "AI and RAN" may be a technology for performing RAN processing and AI and machine learning technology processing that is not directly related to RAN on the same computing infrastructure. By integrating AI and RAN processing with "AI and RAN," improvements in infrastructure utilization efficiency can be expected. The control device 100 may be applied in particular to "AI for RAN."

[0030] The control device 100 may perform RAN control and AI processing. The RAN control performed by the control device 100 may be vRAN (Virtual RAN). The information processing infrastructure 300 may configure vRAN. In this case, multiple wireless base stations are located on the information processing infrastructure 300, and wireless communication services are provided to multiple user terminals 72 by the multiple wireless base stations.

[0031] The user terminal 72 may be any terminal as long as it is capable of using wireless communication services. Examples of user terminals include, but are not limited to, smartphones, tablet terminals, PCs (Personal Computers), mobile Wi-Fi (registered trademark), wearable terminals such as smartwatches and smart glasses, IoT (Internet of Things) terminals such as sensors, smart home appliances, and smart meters, in-vehicle equipment, robots, and game consoles. User terminals may include any terminal that falls under the category of IoE (Internet of Everything).

[0032] The AI ​​processing performed by the control device 100 may include RAN control AI processing, which is AI processing related to RAN control. The AI ​​processing performed by the control device 100 may also include non-RAN control AI processing, which is AI processing not related to RAN control.

[0033] An example of AI-based RAN control processing is the RIC (RAN Intelligent Controller). The RIC is a technology that uses AI to optimize RAN wireless resources and automate RAN operations. The RIC includes Non-RT (Real Time) RIC and Near-RT RIC. The Non-RT RIC is sometimes called a Centralized RIC. The Non-RT RIC is located within the SMO (Service Management and Orchestration) that manages and orchestrates the RAN. The Non-RT RIC generates and notifies policies related to RAN control and transmits information to the Near-RT RIC. For example, a Non-RT RIC generates a trained model for RAN control by performing machine learning using data collected from the RAN, and sends it to a Near-RT RIC. A Near-RT RIC is sometimes called a Distributed RIC. Compared to a Non-RT RIC, a Near-RT RIC is located closer to the RAN nodes (RU (Radio Unit), DU (Distributed Unit), CU (Central Unit)) and performs control of the RAN nodes and resources. Compared to a Non-RT RIC, a Near-RT RIC performs processing with higher real-time capabilities. For example, a Near-RT RIC performs inference processing related to RAN control using the trained model obtained from a Non-RT RIC. RAN control AI processing is not limited to RICs.

[0034] Non-RAN-controlled AI processing may correspond to so-called MEC (Multi-access Edge Computing) applications. Examples of non-RAN-controlled AI processing include, but are not limited to, monitoring AI execution processing that determines the situation within the imaging range of an input image, and response AI execution processing that outputs a response to an inquiry made by a user.

[0035] The control device 100 may group multiple radio base stations to generate base station groups. Based on the base station information of each of the multiple radio base stations, the control device 100 may group the multiple radio base stations and generate multiple base station groups, each containing multiple radio base stations. In the example shown in Figure 3, the control device 100 groups radio base station 201, radio base station 202, and radio base station 203 to generate base station group 21. In the example shown in Figure 3, the remaining radio base stations, including radio base station 209, are grouped to generate base station group 29.

[0036] Base station information may include at least one of the following: information on the location of wireless base stations, information on the communication trends of wireless base stations, the density of wireless base stations in the area including wireless base stations, and information on the terminal trends of user terminals located within the area of ​​wireless base stations.

[0037] The location information for wireless base stations may include any information relating to the location of the wireless base stations. For example, the location information may include the geographical coordinates of the wireless base stations. The location information may also include attributes of the region where the wireless base stations are located. For example, the location information may include attributes such as urban areas, residential areas, rural areas, mountainous areas, commercial districts, high-rise building areas, and areas around train stations.

[0038] The communication trend information for a wireless base station may include any information relating to the communication trends within the coverage area of ​​the wireless base station. For example, the communication trend information may include information showing the bias in communication volume by time of day, information showing the bias in communication volume by season, and information showing the regularity of traffic inflow and outflow. For example, the communication trend information may include information such as the difference in communication volume between daytime and nighttime, the difference in communication volume between summer and other periods, and the slope of the time change in communication volume.

[0039] The base station density of an area including a base station may be, for example, the number of base stations per unit area. The base station density may be a value calculated for each communication standard. The base station density may be a value calculated for each frequency band. The base station density does not have to be a calculated value; it may be a value that expresses the degree of density of base station placement in multiple stages. For example, the base station densities in descending order of density may be 5, 4, 3, 2, and 1.

[0040] Terminal trend information for user terminals located within the service area of ​​a wireless base station may include any information relating to the trends of user terminals located within the service area of ​​a wireless base station. For example, terminal trend information may include information relating to the attributes of user terminals. Terminal trend information may include information indicating biases in the types of user terminals, types of operating systems (OS), supported frequency bands, and supported cooperative communication technologies. Terminal trend information may also include information relating to the attributes of the user of the user terminal. Terminal trend information may include information indicating biases in the user's age, place of residence, employment or schooling status, and location of workplace or school.

[0041] The control device 100 can group multiple wireless base stations based on the base station information described above, thereby classifying wireless base stations with similar characteristics into the same base station group and wireless base stations with different characteristics into different base station groups. This makes it possible to increase the size of a single base station group as much as possible and reduce the load on data transmission. The control device 100 may assign multiple wireless base stations to predefined categories based on the base station information. The control device 100 may also group multiple wireless base stations by clustering based on the base station information.

[0042] The control device 100 collects communication data communicated between a radio base station and a user terminal from some of the plurality of radio base stations included in a base station group. The control device 100 uses the collected communication data to generate an AI model for use in communication between a radio base station and a user terminal. The control device 100 provides the generated AI model to the plurality of radio base stations included in the base station group.

[0043] In the example shown in FIG. 3, the control device 100 collects communication data 91 from the radio base station 201, which is one of the plurality of radio base stations included in the base station group 21. The control device 100 uses the collected communication data 91 to generate an AI model 81. The control device 100 provides the generated AI model 81 to the radio base station 201, the radio base station 202, and the radio base station 203 included in the base station group 21.

[0044] The control device 100 may collect communication data from a plurality of radio base stations among the plurality of radio base stations included in the base station group. The control device 100 may use the communication data collected from the plurality of radio base stations to generate an AI model for use in communication between a radio base station and a user terminal. The control device 100 may provide the generated AI model to the plurality of radio base stations included in the base station group.

[0045] For example, in the example shown in FIG. 3, the control device 100 collects communication data 91 and communication data 92 from the radio base station 201 and the radio base station 202 included in the base station group 21. The control device 100 uses the communication data 91 and the communication data 92 to generate an AI model. The control device 100 provides the generated AI model to the radio base station 201, the radio base station 202, and the radio base station 203 included in the base station group 21.

[0046] The control device 100 may collect communication data from a radio base station having the highest total relevance with a plurality of other radio base stations included in a base station group among the plurality of radio base stations included in the base station group. For example, the control device 100 calculates a statistic representing a relationship between each radio base station among the plurality of radio base stations included in the base station group and each of the other radio base stations, and collects communication data from the radio base station determined based on the statistic. For example, when the statistic is a statistic that takes a larger value when the relationship between radio base stations is closer and more similar, the control device 100 collects communication data from the radio base station having the largest calculation result such as the sum of the statistics. Conversely, when the statistic is a statistic that takes a smaller value when the relationship between radio base stations is closer and more similar, the control device 100 may collect communication data from the radio base station having the smallest calculation result such as the sum of the statistics.

[0047] For example, the statistic may be various correlation coefficients, covariance, coefficient of determination, or the like. For example, when time-series communication traffic of radio base stations is used as base station data, the control device 100 may calculate a cross-correlation function between time-series communication traffic data of each radio base station among the plurality of radio base stations included in the base station group and communication traffic data of each of the other radio base stations, and collect communication data of the radio base station having the largest sum of the cross-correlation functions. In this case, the control device 100 may perform Fourier transform on the time-series communication traffic data before calculating the correlation function.

[0048] The above examples of deriving total relevance are merely illustrative, and the method for deriving total relevance is not limited thereto. Any indicator may be used as total relevance as long as it can evaluate the closeness of the relationship with the remaining other base stations among the plurality of radio base stations included in the base station group.

[0049] The control device 100 may collect communication data from the wireless base station with the highest volume of communication with user terminals among the multiple wireless base stations included in the base station group. By collecting communication data from the wireless base station with the highest volume of communication within the base station group, the control device 100 can generate an AI model as training data that better reflects the characteristics of the base station group, for example, if the base station group 21 is characterized by high traffic and frequent congestion.

[0050] The control device 100 may collect communication data from the wireless base station with the lowest load among the multiple wireless base stations included in the base station group. By collecting communication data from the wireless base station with the lowest communication load in the base station group, the control device 100 can reduce the load on the network associated with the transmission of communication data.

[0051] In the example shown in Figure 3, the control device 100 provides the AI ​​model 81 by transmitting it to the wireless base stations 201, 202, and 203, but the method of providing the AI ​​model is not limited to this. Instead of the control device 100 transmitting the AI ​​model to the wireless base stations, the control device 100 may store the AI ​​model in an area accessible from the wireless base stations. For example, the control device 100 provides the AI ​​model to multiple wireless base stations by storing the AI ​​model in the information processing infrastructure 300.

[0052] In the example shown in Figure 3, the processing performed by the control device 100 on the base station group 29 may be the same as the processing performed on the base station group 21 described above. In the example shown in Figure 3, the control device 100 collects communication data 99 from the radio base station 209, which is one of the multiple radio base stations included in the base station group 29. The control device 100 generates an AI model 89 using the collected communication data 99. The control device 100 provides the generated AI model 89 to the multiple radio base stations included in the base station group 29.

[0053] In the example shown in Figure 3, wireless base station 201 is selected as a representative from the base station group 21 to collect communication data 91, and the AI ​​model 81 generated using the communication data 91 is provided not only to wireless base station 201, but also to wireless base stations 202 and 203. This reduces the communication load on the network 70 associated with the transmission of training data and improves training efficiency.

[0054] On the other hand, malfunctions may occur because AI model 81 is not suitable for communication between wireless base stations 202 and 203. Figure 4 describes the process after wireless base stations 201, 202, and 203 performed wireless communication using AI model 81 following the example shown in Figure 3. Although no malfunctions occurred in the communication between wireless base stations 201 and 202, a malfunction occurred in the communication between wireless base station 203.

[0055] The control device 100 may collect communication data from a wireless base station that has malfunctioned as a result of using the AI ​​model, among a plurality of wireless base stations that have been provided with the AI ​​model, specifically the communication data exchanged between the malfunctioning wireless base station and the user terminal. The control device 100 may update the AI ​​model using the communication data collected from the malfunctioning wireless base station. The control device 100 may provide the updated AI model to the malfunctioning wireless base station.

[0056] In the example shown in Figure 4, the control device 100 collects communication data 93 from the malfunctioning wireless base station 203 by utilizing the AI ​​model 81. The control device 100 updates the AI ​​model 81 using the collected communication data 93 to generate the AI ​​model 83. The control device 100 provides the AI ​​model 83 to the wireless base station 203. Since the AI ​​model 83 is an AI model that has been fine-tuned using the communication data 93 from the wireless base station 203, the malfunction that occurred in the communication of the wireless base station 203 can be improved by using the AI ​​model 83. As a result, overall, it is possible to reduce the communication load and improve learning efficiency while maintaining communication quality.

[0057] In the example shown in Figure 4, updating the AI ​​model 81 by the control device 100 may include generating a new AI model 83. For example, the control device 100 may use communication data 93 but generate AI model 83 without using AI model 81. The control device 100 may use communication data 93 but not other communication data to generate AI model 83. In this case, since AI model 83 can be generated as an AI model specialized for the wireless base station 203, the possibility of improving malfunctions that occur in the wireless base station 203 increases.

[0058] In the example shown in Figure 4, the control device 100 provides the radio base station 203 with the AI ​​model 83 generated by updating the AI ​​model 81 using communication data 93. However, even with the use of the AI ​​model 83, there may be cases where the malfunction in the radio base station 203 is not resolved. For example, if the characteristics of the radio base station 203 within the base station group 21 are not similar to the characteristics of the other radio base stations, updating the AI ​​model 81 using communication data 93 may not resolve the malfunction in the radio base station 203. Figure 5 describes the processing after the example shown in Figure 4, when the radio base station 203 performs wireless communication using the AI ​​model 83, resulting in another malfunction in the radio base station 203's communication.

[0059] The control device 100 may move a radio base station to another base station group if, after the updated AI model has been provided to the radio base station experiencing a malfunction, the radio base station experiences another malfunction as a result of using the updated AI model. In the example shown in Figure 5, the control device 100 moves the radio base station 203, which has experienced another malfunction, to base station group 29. The control device 100 may then provide the radio base station 203 with the AI ​​model 89. By reviewing and updating the base station groups in this way, the likelihood of improving the malfunction in the communication of the radio base station 203 increases, especially when the characteristics of the radio base stations included in base station group 29 are similar to those of the radio base station 203.

[0060] If a malfunction occurs again at a wireless base station, the control device 100 may move the malfunctioning wireless base station to another base station group based on the base station information of each of the multiple wireless base stations and the base station information of the wireless base station that has malfunctioned again. For example, the control device 100 may move the malfunctioning wireless base station to the base station group to which the wireless base station with the highest total correlation to the malfunctioning wireless base station belongs. The control device 100 may move the malfunctioning wireless base station to the base station group to which the proportion of wireless base stations whose total correlation to the malfunctioning wireless base station is above a predetermined threshold is the largest among the multiple wireless base stations. For each of the multiple base station groups, the control device 100 may compare a statistical amount calculated based on the base station information of the multiple wireless base stations included in the multiple base stations with a statistical amount calculated based on the base station information of the malfunctioning wireless base station, and move the malfunctioning wireless base station to the base station group with a similar statistical amount.

[0061] Figure 6 schematically shows an example of the functional configuration of the control device 100. In the example shown in Figure 6, the control device 100 includes a group generation unit 110, a collection unit 120, a storage unit 130, a model generation unit 140, and a model provision unit 150. It is not necessarily required that the control device 100 include all of these.

[0062] The group generation unit 110 may group multiple wireless base stations based on the base station information of each of the multiple wireless base stations and generate multiple base station groups, each containing multiple wireless base stations. The group generation unit 110 may group multiple wireless base stations such that all of the multiple wireless base stations belong to one of the multiple base station groups. The group generation unit 110 may group multiple wireless base stations such that some of the wireless base stations belong to one of the multiple base station groups, while the remaining some wireless base stations do not belong to any base station group. The group generation unit 110 may group multiple wireless base stations such that all of the multiple wireless base stations belong to one base station group.

[0063] The group generation unit 110 may move a malfunctioning radio base station to another base station group if, after the updated AI model has been provided to the malfunctioning radio base station, the malfunctioning radio base station experiences another malfunction as a result of using the updated AI model. By reviewing and updating base station groups in this way, the likelihood of improving the malfunction in the communication of the malfunctioning radio base station increases, especially when the characteristics of the malfunctioning radio base station are more similar to those of radio base stations in other base station groups than to those of other radio base stations in the base station group to which it currently belongs.

[0064] If a malfunction occurs again at a wireless base station, the group generation unit 110 may move the malfunctioning wireless base station to another base station group based on the base station information of each of the multiple wireless base stations and the base station information of the wireless base station that has malfunctioned again.

[0065] The data collection unit 120 collects communication data transmitted between a wireless base station and a user terminal from some of the multiple wireless base stations included in the base station group. The data collection unit 120 may collect communication data from one of the multiple wireless base stations included in the base station group. The data collection unit 120 may collect communication data from two or more wireless base stations included in the base station group.

[0066] The collection unit 120 may collect communication data from the radio base station that has the highest overall correlation with the other radio base stations included in the base station group, among the multiple radio base stations included in the base station group. The collection unit 120 may collect communication data from the radio base station determined by the control device 100 based on a statistical quantity representing the relationship between each radio base station and each of the other radio base stations included in the base station group. The overall correlation may be an index that evaluates the closeness of the relationship with the remaining base stations among the multiple radio base stations included in the base station group.

[0067] The data collection unit 120 may collect communication data from the wireless base station with the highest volume of communication with user terminals among the multiple wireless base stations included in the base station group. By collecting communication data from the wireless base station with the highest volume of communication within the base station group, the control device 100 can generate an AI model as training data that better reflects the characteristics of the base station group, for example, if the base station group 21 has characteristics such as high traffic and is prone to congestion.

[0068] The data collection unit 120 may collect communication data from the wireless base station with the lowest load among the multiple wireless base stations included in the base station group. By collecting communication data from the wireless base station with the lowest communication load in the base station group, the control device 100 can reduce the load on the network associated with the transmission of communication data.

[0069] The collection unit 120 may collect communication data transmitted between the malfunctioning wireless base station and the user terminal from among the multiple wireless base stations that provided the AI ​​model, specifically from the wireless base station that experienced a malfunction as a result of using the AI ​​model.

[0070] The memory unit 130 stores various types of information. The memory unit 130 may store communication data. The memory unit 130 may store base station information. The memory unit 130 may store AI models.

[0071] The model generation unit 140 generates various AI models. The model generation unit 140 uses the communication data collected by the collection unit 120 to generate an AI model for use in communication between the wireless base station and the user terminal. The learning method used by the model generation unit 140 to generate the AI ​​model is not particularly limited. For example, the model generation unit 140 may use various machine learning techniques used to solve optimization problems. For example, the model generation unit 140 generates an AI model that outputs values ​​for communication control parameters that maximize the communication quality between the wireless base station and the user terminal through optimization learning using communication data as training data.

[0072] The model generation unit 140 may update the AI ​​model using communication data collected from the malfunctioning wireless base station. The model generation unit 140 may update the AI ​​model using communication data collected from the malfunctioning wireless base station, but without using the AI ​​model provided to the malfunctioning wireless base station. The model generation unit 140 may update the AI ​​model using communication data collected from the malfunctioning wireless base station, but without using any other communication data. In this case, since the AI ​​model 83 can be generated as an AI model specific to the wireless base station 203, the likelihood of improving the malfunction in the wireless base station 203 increases.

[0073] The model provision unit 150 provides the AI ​​model generated by the model generation unit 140 to multiple wireless base stations included in the base station group. The model provision unit 150 may also provide the updated AI model from the model generation unit 140 to wireless base stations experiencing malfunctions. This allows for overall reduction of communication load and improvement of learning efficiency while maintaining communication quality.

[0074] Figure 7 schematically shows an example of the processing flow by the control device 100. In step 102 (steps may be abbreviated as S), the group generation unit 110 groups multiple wireless base stations to generate a base station group.

[0075] In S104, the collection unit 120 collects communication data from some of the wireless base stations in the base station group generated by the group generation unit 110 in S102.

[0076] In S106, the model generation unit 140 generates an AI model using the communication data collected by the collection unit 120 in S104.

[0077] In S108, the model provision unit 150 provides the AI ​​model generated by the model generation unit 140 in S106 to multiple wireless base stations of the base station group.

[0078] In S110, based on the results of using the AI ​​model provided by the model provision unit 150 in S108, it is determined whether there are any wireless base stations among the multiple wireless base stations included in the base station group that have experienced problems with communication with the user terminal. If the determination result is YES, the process ends; if the determination result is NO, the process proceeds to S112. This determination may be performed by the collection unit 120. The collection unit 120 may perform this determination by acquiring the results of communication by multiple wireless base stations. The collection unit 120 may also acquire the determination result from an external device that has acquired the results of communication by multiple wireless base stations.

[0079] In S112, the data collection unit 120 collects communication data from the malfunctioning wireless base station using the AI ​​model provided by the model provision unit 150 in S108.

[0080] In S114, the model generation unit 140 updates the AI ​​model using the communication data collected in S112 by the collection unit 120 from the wireless base station where a malfunction occurred.

[0081] In S116, the model provision unit 150 provides the AI ​​model updated by the model generation unit 140 in S114 to the wireless base station experiencing the malfunction.

[0082] Figure 8 schematically shows an example of a wireless communication system 10. In Figure 8, the differences from the example shown in Figure 3 will be mainly explained. In the example shown in Figure 8, the wireless communication system 10 comprises a management infrastructure 400 and a plurality of information processing infrastructures 300. The management infrastructure 400 may manage the plurality of information processing infrastructures 300. The management infrastructure 400 may be a data center that manages the plurality of information processing infrastructures 300. Multiple types of devices may be arranged on the management infrastructure 400. The management infrastructure 400 may be composed of multiple devices. The management infrastructure 400 may be implemented on a virtualization infrastructure with multiple devices. The management infrastructure 400 may be implemented by a single device. That is, the management infrastructure 400 may be a management device.

[0083] In the example shown in Figure 8, the configuration of each of the multiple information processing infrastructures 300 is the same as in the example shown in Figure 1. Note that in Figure 8, the control device 100, network 70, base station group, communication data, and AI model are omitted from the description.

[0084] In the example shown in Figure 8, the control device 100 may be located on the management infrastructure 400. In this case, the control device 100 may be further located on at least one of the multiple information processing infrastructures 300. Alternatively, the control device 100 may not be located on the management infrastructure 400, but on each of the multiple information processing infrastructures 300.

[0085] In the example shown in Figure 8, base station groups may be generated such that wireless base stations under different information processing infrastructures 300 belong to the same base station group. For example, wireless base stations located in different regions 30 may have similar attributes. In such cases, the control device 100 may collect communication data from some of the wireless base stations in a base station group that includes multiple wireless base stations with similar attributes, generate an AI model using the collected communication data, and provide the generated AI model to the multiple wireless base stations with similar attributes included in the base station group.

[0086] The management infrastructure 400 may be called the Core Brain, and the information processing infrastructure 300 may be called the Regional Brain. Note that Figure 8 illustrates a case where a single-layer information processing infrastructure 300 is located below the management infrastructure 400, but this is not the only example. The information processing infrastructure 300 may have multiple layers. For example, if two layers of information processing infrastructure 300 are located below the management infrastructure 400, the management infrastructure 400 may be called the Core Brain, the lower layer of information processing infrastructure 300 may be called the Regional Brain, and the further lower layer of information processing infrastructure 300 may be called the Sub-Regional Brain. In this case, the control device 100 may be located in the Regional Brain. In any of the above cases, the control device 100 may be connected to the information processing infrastructure 300 or management infrastructure 400 instead of being located on them.

[0087] Figure 9 schematically shows an example of the hardware configuration of a computer 1200 that functions as a control device 100, an information processing platform 300, or a management platform 400. A program installed on the computer 1200 can cause the computer 1200 to function as one or more "parts" of the apparatus according to this embodiment, or to cause the computer 1200 to execute operations associated with the apparatus according to this embodiment or such one or more "parts", and / or to cause the computer 1200 to execute a process or a stage of such process according to this embodiment. Such a program may be executed by the CPU 1212 to cause the computer 1200 to execute specific operations associated with some or all of the blocks in the flowcharts and block diagrams described herein.

[0088] The computer 1200 according to this embodiment includes a CPU 1212, a GPU 1213, a RAM 1214, and a graphics controller 1216, which are interconnected by a host controller 1210. The computer 1200 also includes input / output units such as a communication interface 1222, a storage device 1224, a DVD drive, and an IC card drive, which are connected to the host controller 1210 via an input / output controller 1220. The DVD drive may be a DVD-ROM drive and a DVD-RAM drive, etc. The storage device 1224 may be a hard disk drive and a solid-state drive, etc. The computer 1200 also includes legacy input / output units such as a ROM 1230 and a keyboard, which are connected to the input / output controller 1220 via an input / output chip 1240.

[0089] The CPU 1212 operates according to the programs stored in the ROM 1230 and RAM 1214, thereby controlling each unit. The graphics controller 1216 acquires the image data generated by the CPU 1212 and stores it in the frame buffer provided in the RAM 1214 or within itself, so that the image data is displayed on the display device 1218.

[0090] The communication interface 1222 communicates with other electronic devices via a network. The storage device 1224 stores programs and data used by the CPU 1212 in the computer 1200. The DVD drive reads programs or data from a DVD-ROM or the like and provides them to the storage device 1224. The IC card drive reads programs and data from an IC card and / or writes programs and data to an IC card.

[0091] The ROM 1230 stores boot programs and / or hardware-dependent programs of the computer 1200, which are executed by the computer 1200 when activated. The input / output chip 1240 may also connect various input / output units to the input / output controller 1220 via USB ports, parallel ports, serial ports, keyboard ports, mouse ports, etc.

[0092] The program is provided on a computer-readable storage medium such as a DVD-ROM or IC card. The program is read from the computer-readable storage medium and installed on a storage device 1224, RAM 1214, or ROM 1230, which are examples of computer-readable storage media, and executed by the CPU 1212. The information processing described within these programs is read by the computer 1200, resulting in coordination between the program and the various types of hardware resources described above. The apparatus or method may be configured to realize the operation or processing of information in accordance with the use of the computer 1200.

[0093] For example, when communication is performed between a computer 1200 and an external device, the CPU 1212 may execute a communication program loaded into the RAM 1214 and, based on the processing described in the communication program, instruct the communication interface 1222 to perform communication processing. Under the control of the CPU 1212, the communication interface 1222 reads transmission data stored in a transmission buffer area provided in a recording medium such as the RAM 1214, storage device 1224, DVD-ROM, or IC card, transmits the read transmission data to the network, or writes received data received from the network to a reception buffer area or the like provided on the recording medium.

[0094] Furthermore, the CPU 1212 may read all or necessary parts of a file or database stored on an external recording medium such as a storage device 1224, a DVD drive (DVD-ROM), or an IC card into the RAM 1214, and perform various types of processing on the data in the RAM 1214. The CPU 1212 may then write the processed data back to the external recording medium.

[0095] Various types of information, such as various types of programs, data, tables, and databases, may be stored on the recording medium and subjected to information processing. The CPU 1212 may perform various types of processing on the data read from the RAM 1214, including various types of operations, information processing, conditional judgments, conditional branching, unconditional branching, information retrieval / replacement, etc., as described throughout this disclosure and specified by the program instruction sequence, and write the results back to the RAM 1214. The CPU 1212 may also retrieve information in files, databases, etc., within the recording medium. For example, if a plurality of entries are stored in the recording medium, each having an attribute value of a first attribute associated with an attribute value of a second attribute, the CPU 1212 may search among the plurality of entries for an entry that matches the specified condition for the attribute value of the first attribute, read the attribute value of the second attribute stored in that entry, and thereby obtain the attribute value of the second attribute associated with the first attribute that satisfies a predetermined condition.

[0096] The program or software module described above may be stored on or near the computer 1200 in a computer-readable storage medium. Alternatively, a recording medium such as a hard disk or RAM provided within a server system connected to a dedicated communication network or the Internet can be used as a computer-readable storage medium, thereby providing the program to the computer 1200 via the network.

[0097] In this embodiment, blocks in the flowchart and block diagram may represent a stage in a process in which an operation is performed or a "part" of a device that has the role of performing an operation. A particular stage and "part" may be implemented by a dedicated circuit, a programmable circuit supplied with computer-readable instructions stored on a computer-readable storage medium, and / or a processor supplied with computer-readable instructions stored on a computer-readable storage medium. The dedicated circuit may include digital and / or analog hardware circuits, and may include integrated circuits (ICs) and / or discrete circuits. The programmable circuit may include reconfigurable hardware circuits, such as field-programmable gate arrays (FPGAs) and programmable logic arrays (PLAs), which include logical AND, logical OR, exclusive OR, negated AND, negated OR, and other logical operations, flip-flops, registers, and memory elements.

[0098] A computer-readable storage medium may include any tangible device capable of storing instructions to be executed by a suitable device, and as a result, a computer-readable storage medium having instructions stored therein will comprise a product that includes instructions that can be executed to create means for performing operations specified in a flowchart or block diagram. Examples of computer-readable storage media may include electronic storage media, magnetic storage media, optical storage media, electromagnetic storage media, semiconductor storage media, etc. More specific examples of computer-readable storage media may include floppy disks, diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), electrically erasable programmable read-only memory (EEPROM), static random access memory (SRAM), compact disk read-only memory (CD-ROM), digital multipurpose disk (DVD), Blu-ray® disk, memory stick, integrated circuit card, etc.

[0099] Computer-readable instructions may include assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk®, Java®, C++, and conventional procedural programming languages ​​such as the C programming language or similar programming languages.

[0100] Computer-readable instructions may be provided locally or via a wide area network (WAN) such as a local area network (LAN) or the internet to a processor or programmable circuit of a general-purpose computer, special-purpose computer, or other programmable data processing device, so that the processor or programmable circuit of the programmable data processing device, such as a computer, can execute the instructions to generate means for performing operations specified in a flowchart or block diagram. Here, the computer may be a PC (personal computer), tablet computer, smartphone, workstation, server computer, general-purpose computer, or special-purpose computer, and may also be a computer system in which multiple computers are connected. Such a computer system in which multiple computers are connected is also called a distributed computing system and is a computer in a broad sense. In a distributed computing system, multiple computers execute a program collectively by each computer executing a part of the program and passing data during program execution between computers as needed.

[0101] Examples of processors include computer processors, central processing units (CPUs), processing units, microprocessors, digital signal processors, controllers, and microcontrollers. A computer may have one or more processors. In a multiprocessor system with multiple processors, each processor executes a portion of the program, and the processors collectively execute the program by passing program execution data between them as needed. For example, in the execution of multitasks, each of the multiple processors may execute a portion of each task in small chunks by switching tasks at each time slice. In this case, which part of a program each processor executes changes dynamically. Which part of a program each of the multiple processors executes may also be statically determined by multiprocessor-aware programming.

[0102] The invention according to this embodiment enables AI model learning even for low-layer wireless communication processing. This improves the performance and efficiency of wireless communication services, making it possible to provide a higher-performance and more efficient communication infrastructure. Therefore, it can contribute to achieving at least one of the Sustainable Development Goals (SDGs): Goal 9, "Build resilient infrastructure, promote inclusive and sustainable industrialization and foster innovation," and Goal 11, "Make cities and human settlements inclusive, safe, resilient and sustainable."

[0103] Although the present invention has been described above using embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various modifications or improvements can be made to the above embodiments. It will be clear from the claims that such modified or improved forms may also be included in the technical scope of the present invention.

[0104] It should be noted that the execution order of operations, procedures, steps, and stages in the devices, systems, programs, and methods shown in the claims, specifications, and drawings is not explicitly stated as "before" or "prior to," and that these can be performed in any order unless the output of a previous operation is used in a later operation. Even if the operation flow in the claims, specifications, and drawings is described using phrases such as "first," and "next," for convenience, this does not mean that it is mandatory to perform the operations in that order.

[0105] 10 Wireless communication system, 21 Base station group, 29 Base station group, 30 Region, 70 Network, 72 User terminal, 80 AI model, 81 AI model, 82 AI model, 83 AI model, 89 AI model, 91 Communication data, 92 Communication data, 93 Communication data, 99 Communication data, 100 Control device, 110 Group generation unit, 120 Collection unit, 130 Storage unit, 140 Model generation unit, 150 Model provision unit, 201 Wireless base station, 202 Wireless base station, 203 Wireless base station, 209 Wireless base station, 300 Information processing infrastructure, 400 Management infrastructure, 800 Learning device, 801 Learning device, 802 Learning device, 803 Learning device, 809 Learning device, 1200 Computer, 1210 Host controller, 1212 CPU, 1213 GPU, 1214 RAM, 1216 Graphics controller, 1218 Display device, 1220 Input / Output controller, 1222 Communication interface, 1224 Storage device, 1230 ROM, 1240 Input / Output chip

Claims

1. A control device comprising: a collection unit that collects communication data communicated between a wireless base station and a user terminal from some of the wireless base stations included in a base station group; a model generation unit that generates an AI (Artificial Intelligence) model for use in communication between the wireless base station and the user terminal using the communication data collected by the collection unit; and a model provision unit that provides the AI ​​model generated by the model generation unit to the multiple wireless base stations included in the base station group.

2. The control device according to claim 1, wherein the collection unit collects communication data communicated between the malfunctioning wireless base station and the user terminal from among the plurality of wireless base stations to which the model providing unit has provided the AI ​​model, and from the wireless base station that has malfunctioned as a result of using the AI ​​model; the model generation unit updates the AI ​​model using the communication data collected by the collection unit from the malfunctioning wireless base station; and the model providing unit provides the AI ​​model updated by the model generation unit to the malfunctioning wireless base station.

3. The control device according to claim 1 or 2, wherein the collection unit collects the communication data from the wireless base station with the highest volume of communication with the user terminal among the plurality of wireless base stations included in the base station group.

4. The control device according to claim 1 or 2, wherein the collection unit collects the communication data from the wireless base station with the lowest load among the plurality of wireless base stations included in the base station group.

5. The control device according to any one of claims 1 to 4, further comprising a group generation unit that groups the plurality of wireless base stations based on the base station information of each of the plurality of wireless base stations to generate a plurality of base station groups, each of which includes a plurality of wireless base stations.

6. The control device according to claim 5, wherein the base station information includes at least one of the following: location information of the wireless base station, communication trend information of the wireless base station, wireless base station density of the area including the wireless base station, and terminal trend information of user terminals located at the wireless base station.

7. The control device according to claim 1, comprising: a group generation unit that groups the plurality of wireless base stations based on the base station information of each of the plurality of wireless base stations to generate a plurality of base station groups, each of which includes a plurality of wireless base stations; a collection unit that collects communication data communicated between the malfunctioning wireless base station and a user terminal from among the plurality of wireless base stations to which the model provision unit has provided the AI ​​model, and from the wireless base station that has malfunctioned due to the use of the AI ​​model; a model generation unit that updates the AI ​​model using the communication data collected by the collection unit from the malfunctioning wireless base station; a model provision unit that provides the AI ​​model updated by the model generation unit to the malfunctioning wireless base station; and a group generation unit that moves the wireless base station to another base station group if, after the updated AI model has been provided to the malfunctioning wireless base station, a malfunction occurs again in the malfunctioning wireless base station due to the use of the updated AI model.

8. The control device according to any one of claims 1 to 7, wherein the collection unit collects the communication data from the radio base station that has the greatest total correlation with the other radio base stations included in the base station group, among the plurality of radio base stations included in the base station group.

9. A program for causing a computer to function as a control device according to any one of claims 1 to 8.

10. A control method comprising: a collection step of collecting communication data communicated between a wireless base station and a user terminal from some of the multiple wireless base stations included in a base station group; a model generation step of generating an AI model for use in communication between the wireless base station and the user terminal using the communication data collected in the collection step; and a model provision step of providing the AI ​​model generated in the model generation step to the multiple wireless base stations included in the base station group.

Images