CONTROL SYSTEM, CONTROL METHOD, AND CONTROL PROGRAM

Abstract

A first control system (10) comprises: a collection unit (11) for collecting identification data that is used for identification by an identification model that identifies a control pertaining to a wireless network, such as a wireless access network; and a transmission unit (12) for transmitting the identification data collected by the collection unit (11) to a second control system (20) that carries out training by means of a training model that learns a control pertaining to a wireless network, such as a wireless access network, said identification data being transmitted as training data used in the training model.

Description

Control system, control device, control method, and non-transitory computer-readable medium

[0001] The present disclosure relates to a control system, a control device, a control method, a control program, and a non-transitory computer-readable medium.

[0002] In recent years, the introduction of 5G (5th Generation) has been promoted as a wireless communication technology that achieves large capacity, low latency, and multiple connectivity. In next-generation wireless communication systems including 5G, the openness of the Radio Access Network (RAN) is being promoted to accommodate increasingly sophisticated and complex systems, and the Open Radio Access Network (O-RAN) Alliance is discussing the openness and intelligentization of the RAN.

[0003] Related Patent Literature 1 and Non-Patent Literature 1 describe that Non-RT (Real Time) RIC and Near-RT RIC are provided as RAN Intelligent Controllers (RICs) that utilize AI / ML (Artificial Intelligence / Machine Learning) to intelligently control the RAN. The Near-RT RIC is placed near E2 nodes including O-DUs (O-RAN Distributed Units) and O-CUs (O-RAN Central Units) and controls the RAN in quasi-real time. The Non-RT RIC is placed at a location away from the E2 nodes and controls the RAN in non-real time.

[0004] International Publication No. 2022 / 070363

[0005] O-RAN ALLIANCE, O-RAN Working Group 2,"AI / ML workflow description and requirements", Technical Report, O-RAN.WG2.AIML-v01.03, 2021.07.20

[0006] According to Patent Document 1 and Non-Patent Document 1, an inference model that infers RAN control and a learning model that performs learning to build the inference model can be placed in either a Near-RT RIC or a Non-RT RIC, or can be distributed. This makes it possible to perform learning using the learning model using operational data while executing control using the inference model. However, since it is necessary to collect learning data to perform learning using the learning model, this may impose a load on the transmission path, device, etc. that transmits the data.

[0007] In view of such problems, one of the objects of the present disclosure is to provide a control system, a control device, a control method, and a non-transitory computer-readable medium that can reduce the load.

[0008] The control system of the present disclosure comprises a collection means for collecting identification data to be used in an identification model that identifies control related to a wireless network, and a transmission means for transmitting the identification data to another control system that learns using the learning model as learning data to be used in a learning model that learns control related to the wireless network.

[0009] The control system according to the present disclosure comprises a receiving means for receiving, as learning data, identification data collected by another control system in order to identify control relating to a wireless network using a specific model, from the other control system, and a learning means for using the learning data to learn control relating to the wireless network using a learning model.

[0010] The control device according to the present disclosure includes a collection means for collecting identification data used by an identification model that identifies control related to a wireless network, and a transmission means for transmitting the identification data to another control system that learns using the learning model as learning data to be used by a learning model that learns control related to the wireless network.

[0011] The control device according to the present disclosure comprises a receiving means for receiving, as learning data, identification data collected by another control system to identify control relating to a wireless network using a specific model from the other control system, and a learning means for using the learning data to learn control relating to the wireless network using a learning model.

[0012] The control method disclosed herein involves a specification model that specifies control related to a wireless network collecting specification data to be used for the specification, and transmitting the specification data to another control system that performs learning using the learning model as learning data to be used for a learning model that learns control related to the wireless network.

[0013] The control method disclosed herein receives, from another control system, identification data collected by the other control system to identify control related to a wireless network using a specific model, as learning data, and uses the learning data to learn control related to the wireless network using a learning model.

[0014] The non-transitory computer-readable medium of the present disclosure is a non-transitory computer-readable medium storing a control program for causing a computer to execute a process in which an identification model that identifies controls related to a wireless network collects identification data to be used for the identification, and transmits the identification data to another control system that performs learning using the learning model as learning data to be used for a learning model that learns controls related to the wireless network.

[0015] The non-transitory computer-readable medium of the present disclosure is a non-transitory computer-readable medium having stored thereon a control program for causing a computer to execute a process of receiving, as learning data, identification data collected by another control system to identify control related to a wireless network using a specific model, from the other control system, and using the learning data to learn control related to the wireless network using a learning model.

[0016] According to the present disclosure, it is possible to provide a control system, a control device, a control method, a control program, and a non-transitory computer-readable medium that are capable of reducing the load.

[0017] 1 is a configuration diagram showing an overview of a first control system according to an embodiment. 2 is a configuration diagram showing an overview of a second control system according to an embodiment. 3 is a configuration diagram showing an overview of a first control device according to an embodiment. 4 is a configuration diagram showing an overview of a second control device according to an embodiment. 5 is a flowchart showing an overview of a first control method according to an embodiment. 6 is a flowchart showing an overview of a second control method according to an embodiment. 7 is a configuration diagram showing an example of a configuration of a RAN system according to embodiment 1. 8 is a diagram for explaining a data flow in a comparative example. 9 is a diagram for explaining a data flow in embodiment 1. 10 is a configuration diagram showing an example of a configuration of a Near-RT RIC according to embodiment 1. 11 is a configuration diagram showing an example of a configuration of a Non-RT RIC according to embodiment 1. 12 is a flowchart showing an overview of a learning process in the RAN system according to embodiment 1. 13 is a flowchart showing an operation example of the learning process in the RAN system according to embodiment 1. 14 is a sequence diagram showing an operation example of the learning process in the RAN system according to embodiment 1. 15 is a configuration diagram showing an example of a configuration of a Near-RT RIC according to embodiment 2. 16 is a configuration diagram showing an example of a configuration of a Near-RT RIC according to embodiment 3. 17 is a configuration diagram showing an example of a configuration of a Near-RT RIC according to embodiment 4. Fig. 10 is a configuration diagram showing a configuration example of a RAN system according to embodiment 5. Fig. 11 is a configuration diagram showing an overview of hardware of a computer according to an embodiment.

[0018] Hereinafter, embodiments will be described with reference to the drawings. In the drawings, the same elements are denoted by the same reference numerals, and redundant description will be omitted as necessary.

[0019] For example, if an inference model is placed in a Near-RT RIC and a learning model is placed in a Non-RT RIC, one possible method is for the Near-RT RIC to collect the necessary data from the E2 node for inference, and for the Non-RT RIC to collect the necessary data from the E2 node for learning. The inventors of the present disclosure considered such a collection method and found a problem in that it could place a load on the E2 node and the transmission path over which the data is transmitted. Therefore, in the embodiment, it is possible to reduce the load on the E2 node and the transmission path over which the data is transmitted.

[0020] (Outline of the embodiment) First, an outline of the embodiment will be described. Fig. 1 shows a schematic configuration of a first control system 10 according to the embodiment, and Fig. 2 shows a schematic configuration of a second control system 20 according to the embodiment. For example, the first control system 10 and the second control system 20 constitute a system that controls a wireless network such as a RAN. For example, the first control system 10 includes a Near-RT RIC, and the second control system 20 includes a Non-RT RIC, but is not limited to these.

[0021] As shown in FIG. 1 , the first control system 10 includes a collection unit 11 and a transmission unit 12. The collection unit 11 collects identification data used for identification by an identification model that identifies control related to the wireless network. The collection unit 11 collects identification data from a RAN including an O-DU and an O-CU. The control related to the wireless network is, for example, control of the operation of the RAN, such as control of a wireless resource allocation scheduler, beams, handover, etc., which is possible by configuring the O-DU and O-CU. The identification model can also be considered an inference model that infers control related to the wireless network, and the identification data can be considered inference data used for inference. For example, the identification model identifies control of the RAN according to data such as wireless quality collected from the RAN. The identification model is, for example, included in the first control system 10, but may also be located outside the first control system 10.

[0022] The transmitter 12 transmits the identification data collected by the collector 11 to the second control system 20, which performs learning using a learning model, as learning data to be used in a learning model that learns control related to a wireless network. For example, the learning model is a model for constructing an identification model. The learning data transmitted by the transmitter 12 may include identification result data, which is the result of identification by the identification model. The identification result data may be, for example, analysis information obtained by analyzing the identification data or control information for controlling the RAN. Furthermore, the transmitter 12 may transmit the learning data depending on the communication status between the first control system 10 and the second control system 20. For example, the transmitter 12 may transmit the learning data depending on the availability of an interface connecting to the second control system 20.

[0023] 2, the second control system 20 includes a receiving unit 21 and a learning unit 22. The receiving unit 21 receives learning data transmitted from the transmitting unit 12 of the first control system 10. That is, the receiving unit 21 receives, as learning data, identification data collected by the first control system 10 to identify control related to a wireless network using a specific model.

[0024] The learning unit 22 uses the learning data received by the receiving unit 21 to learn control related to the wireless network using a learning model. That is, it learns control of the RAN according to data such as wireless quality. The learning model is included in the second control system 20, for example, but may also be located outside the second control system 20. The learning unit 22 applies the learned learning model, which has been trained using the learning data, to a specific model of the first control system 10.

[0025] The first control system 10 and the second control system 20 may each be configured by a single device or multiple devices. FIG. 3 shows an example configuration of a first control device 30 according to an embodiment, and FIG. 4 shows an example configuration of a second control device 40 according to an embodiment. As shown in FIG. 3, the first control device 30 may include the collection unit 11 and the transmission unit 12 shown in FIG. 1. This example is not limited to this example, and the collection unit 11 and the transmission unit 12 may be implemented in separate devices. As shown in FIG. 4, the second control device 40 may include the reception unit 21 and the learning unit 22 shown in FIG. 2. This example is not limited to this example, and the reception unit 21 and the learning unit 22 may be implemented in separate devices. As with the first control system 10 and the second control system 20, for example, the first control device 30 may be a Near-RT RIC, and the second control device 40 may be a Non-RT RIC.

[0026] Furthermore, part or all of the first control system 10 and the second control system 20 may be placed at an edge or in the cloud using virtualization technology or the like. They may be placed in a specific location, or may be distributed across multiple locations. The edge is a location or infrastructure on the base station side that includes the O-DU and O-CU. The cloud is a location or infrastructure on the core network side that is distant from the base station. For example, the collection unit 11 and the transmission unit 12 may be placed at the edge, and the reception unit 21 and the learning unit 22 may be placed in the cloud. Furthermore, the collection unit 11, the transmission unit 12, the reception unit 21, and the learning unit 22 may each be distributed.

[0027] Fig. 5 shows a first control method according to an embodiment, and Fig. 6 shows a second control method according to an embodiment. For example, the first control method is executed by the first control system 10 in Fig. 1 or the first control device 30 in Fig. 3. The second control method is executed by the second control system 20 in Fig. 2 or the second control device 40 in Fig. 4.

[0028] 5, the collection unit 11 collects identification data used by an identification model that identifies control related to a wireless network (S11). Next, the transmission unit 12 transmits the collected identification data to the second control system 20 that performs learning using the learning model, as learning data to be used by the learning model that learns control related to the wireless network (S12).

[0029] 6 , the receiver 21 receives the learning data transmitted from the transmitter 12 of the first control system 10, i.e., the learning data that is collected as identification data to be used by the identification model (S21). Next, the learner 22 uses the received learning data to learn control related to the wireless network using the learning model (S22). Furthermore, the learner 22 applies the learned learning model to the identification model of the first control system 10.

[0030] Thus, in this embodiment, a first control system such as a Near-RT RIC transmits the identification data collected for the identification model to a second control system such as a Non-RT RIC as training data to be used in the training model. The second control system then performs training using the training data received from the first control system. This allows the first control system to transfer data collected from a wireless network to the second control system, and the second control system can then train the training model using the transferred data, thereby reducing the load on the nodes providing the wireless network data and the transmission paths that transmit the data.

[0031] (Embodiment 1) Next, a description will be given of embodiment 1. In this embodiment, an example will be described in which data used in inference by a Near-RT RIC is transferred to a Non-RT RIC as learning data.

[0032] 7 shows an example of the configuration of a RAN system 1 according to this embodiment. As shown in FIG. 7, the RAN system 1 includes a Near-RT RIC 100, a Non-RT RIC 200, an O-DU 300, and an O-CU 400.

[0033] The Non-RT RIC 200 and the Near-RT RIC 100, and the Non-RT RIC 200 and the E2 node including the O-DU 300 and the O-CU 400 are communicatively connected via the O1 interface. The O1 interface is an interface for sending and receiving data and messages required mainly for operation and management. Note that the interface is a connection interface defined by a communication protocol for sending and receiving data and messages, and includes logical transmission paths and networks, as well as physical transmission paths and networks.

[0034] The Non-RT RIC 200 and the Near-RT RIC 100 are communicatively connected via an A1 interface. The Near-RT RIC 100 and the E2 node including the O-DU 300 and the O-CU 400 are connected via an E2 interface. The A1 interface and the E2 interface are interfaces primarily for transmitting and receiving data and messages required for control. The O-DU 300 and the O-CU 400 are communicatively connected via an F1 interface.

[0035] The O-DU 300 and the O-CU 400 are nodes that constitute the RAN and are also referred to as E2 nodes. The RAN is a wireless network accessed by UEs (User Equipment) and is connected to a core network such as the 5G Core network (5GC) or the Evolved Packet Core (EPC). The RAN may include an O-RU (O-RAN Remote Unit) that constitutes an antenna. The UE is a terminal device that connects to the RAN and performs wireless communication, and may be a mobile phone, smartphone, tablet terminal, IoT (Internet of Things) terminal, etc. The UE may also be an application device such as a robot, drone, or self-driving car that implements terminal functions.

[0036] The O-DU 300 and the O-CU 400 provide base station functions. The base station is, for example, a gNB (next generation Node B) or an eNB (evolved Node B), but is not limited to these. The O-DU 300 and the O-CU 400 are examples of nodes that provide base station functions, and may be other network nodes.

[0037] The O-DU 300 is a logical node that provides the radio signal control function and Layer 2 control function of the base station. The O-DU 300 accommodates the O-RU, controls the antenna radio signal (beam) in the accommodated O-RU, and performs protocol processing such as MAC (Media Access Control) and RLC (Radio Link Control) required between the O-RU and the O-CU 400.

[0038] The O-CU 400 is a logical node that provides the radio resource control function of the base station and data processing functions higher than Layer 2. The O-CU 400 accommodates the O-DU 300 and performs data transmission and reception via the accommodated O-DU 300, QoS (Quality of Service) control, cell / UE management, handover control, and protocol processing such as PDCP (Packet Data Convergence Protocol), SDAP (Service Data Adaptation Protocol), and RRC (Radio Resource Control) required between the O-DU 300 and the core network.

[0039] The RAN system 1 may include any number of O-DUs 300 and O-CUs 400, one or more, as the E2 node. That is, the RAN system 1 may include multiple base stations. The number of O-DUs 300 and O-CUs 400 does not necessarily have to be the same. The O-DUs 300 and O-CUs 400 may be located in different locations or in the same location. Furthermore, the O-DUs 300 and O-CUs 400 may be implemented by different virtual machines running on an edge virtualization platform, or by the same virtual machine. The O-DUs 300 and O-CUs 400 may be virtualized distributed units (vDUs) and virtualized central units (vCUs), and may constitute virtual base stations. The O-DUs 300 and O-CUs 400 may also be physical DUs and CUs. The E2 node may also be a base station device including the functions of the O-DUs 300 and O-CUs 400.

[0040] The Near-RT RIC 100 is a logic function that controls and optimizes the RAN in near real time. The Near-RT RIC 100 controls the RAN in a short control period, for example, greater than or equal to 10 ms and less than 1 s. The Near-RT RIC 100 collects and analyzes RAN data from E2 nodes including either or both of the O-DU 300 and the O-CU 400 via the E2 interface, and controls the E2 nodes according to the RAN data. For example, the Near-RT RIC 100 performs control according to the RAN data in accordance with a control policy acquired from the Non-RT RIC 200 via the A1 interface. The control policy is a policy related to RAN control, such as an A1 policy. The A1 policy is guidance for RAN optimization defined via the A1 interface. The RAN data is data collected from either or both of the O-DU 300 and the O-CU 400. The RAN data is radio-related data related to the radio of the RAN, and includes radio quality data and location information for each UE, and may include the number of active UEs for each base station (cell). For example, radio quality data may be acquired from the O-DU 300, or handover information may be acquired from the O-CU 400. The Near-RT RIC 100 is located in the same location as either or both of the O-DU 300 and the O-CU 400, or in a location close to either or both of the O-DU 300 and the O-CU 400. For example, the Near-RT RIC 100 may be implemented in the same edge virtual machine as either or both of the O-DU 300 and the O-CU 400.

[0041] Some functions of the Near-RT RIC 100 are realized by xApps (Near-RT RIC Applications). The xApps include applications that analyze RAN data, control the RAN, and the like. For example, the xApps include an inference model (also called an inference engine), which is a trained model, and use the inference model to analyze inference data including RAN data and control the RAN. The Near-RT RIC 100 may include multiple xApps as necessary.

[0042] The Non-RT RIC 200 is a logic function that controls and optimizes the RAN in non-real time. The Non-RT RIC 200 controls the RAN with a long control period of, for example, 1 second or more. The Non-RT RIC 200 manages control policies, manages the operation of the E2 node including the O-DU 300 and O-CU 400 and the Near-RT RIC 100, learns (trains) learning models, and updates inference models. For example, the Non-RT RIC 200 generates control policies and notifies the Near-RT RIC 100 of the generated control policies via the A1 interface. The Non-RT RIC 200 also manages and sets the configuration information (Configuration) of the E2 node based on data acquired from the E2 node and the Near-RT RIC 100 via the O1 interface. The Non-RT RIC 200 is arranged in a Service Management and Orchestration (SMO) that manages and orchestrates the RAN. The SMO is arranged in a location away from the O-DU 300, the O-CU 400, and the Near-RT RIC 100, for example, in a cloud. The Non-RT RIC 200 may include the functionality of the SMO.

[0043] Some functions of the Non-RT RIC 200 are realized by rApp (Non-RT RIC Application). The rApp includes applications that generate control policies and manage the inference model of the Near-RT RIC 100. For example, the rApp includes a learning model (also called a learning engine) and generates a learning model that learns RAN control using learning data acquired from the E2 node and the Near-RT RIC 100 via the O1 interface, and applies the generated trained learning model to the xApp of the Near-RT RIC 100. Note that applying the trained learning model to the inference model is also referred to as deploying. Deploying refers to placing and deploying the model in the application execution environment and making the model executable. The Non-RT RIC 200 may include multiple rApps as necessary.

[0044] Fig. 8 shows the flow of data for inference and learning in a comparative example, and Fig. 9 shows the flow of data for inference and learning in this embodiment. Note that the arrows shown in Fig. 8 and Fig. 9 indicate the direction of data collection for convenience of explanation, and are not intended to limit the data transmission and reception associated with such data collection.

[0045] As shown in Figure 8, in the comparative example, when the learning model (rApp) of the Non-RT RIC 200 performs learning, the Non-RT RIC 200 collects RAN data (D1, C1) as learning data from the O-DU 300 and O-CU 400 via the O1 interface. Also, the Near-RT RIC 100 collects RAN data (D2, C2) as inference data from the O-DU 300 and O-CU 400 via the E2 interface. The Near-RT RIC 100 uses the collected RAN data to perform inference using the inference model (xApp), and controls the O-DU 300 and O-CU 400 based on the inference results. The Near-RT RIC 100 may collect RAN data from either the DU 300 or the O-CU 400 and control either the DU 300 or the O-CU 400. The Near-RT RIC 100 transmits the inferred inference result data (N1) as learning data to the Non-RT RIC 200 via the O1 interface. The learning model of the Non-RT RIC 200 performs learning using the RAN data (D1, C1) collected from the O-DU 300 and the O-CU 400 and the inference result data (N1) collected from the Near-RT RIC 100 via the O1 interface as learning data. The Non-RT RIC 200 may collect RAN data from either the DU 300 or the O-CU 400, similar to the Near-RT RIC 100. In this example, the RAN data (D1, C1) and the RAN data (D2, C2) are the same data.

[0046] In the comparative example, since either or both of the O-DU300 and the O-CU400 transmit RAN data via two interfaces, the O1 interface and the E2 interface, the communication load on either or both of the O-DU300 and the O-CU400 is doubled compared to when transmitting via a single interface. As a result, in the comparative example, a communication load solely for the purpose of acquiring and collecting learning data occurs on either or both of the O-DU300 and the O-CU400, and therefore it is necessary to increase the communication resources of either or both of the O-DU300 and the O-CU400 for occasional learning.

[0047] In addition, in the comparative example, RAN data from either or both of the O-DU 300 and the O-CU 400 and inference result data from the Near-RT RIC 100 are transmitted via the O1 interface, so the load on the O1 interface is also heavy. In particular, if the transmission buffers of the O-DU 300 and the O-CU 400 are small, the transmission timing cannot be adjusted, so the communication load on the O1 interface is likely to increase. As a result, in the comparative example, the communication load is concentrated on the O1 interface when acquiring and collecting learning data, so it is necessary to increase the communication resources of the O1 interface for occasional learning.

[0048] Therefore, in this embodiment, when learning data used by the learning model installed in Non-RT RIC 200 is acquired and collected from either or both of O-DU 300 and O-CU 400, the communication load on either or both of O-DU 300 and O-CU 400 can be reduced, and concentration of the communication load on the O1 interface can be prevented.

[0049] Specifically, in this embodiment, as shown in FIG. 9, when the learning model (rAPP) of the Non-RT RIC 200 performs learning, the Near-RT RIC 100 collects RAN data (D2, C2) as inference data via the E2 interface, and the inference model (xApp) performs inference using the collected RAN data. As in the comparative example, the Near-RT RIC 100 may collect RAN data from either the DU 300 or the O-CU 400 and control either the DU 300 or the O-CU 400. In this embodiment, the Near-RT RIC 100 accumulates data required for learning from the data used in inference as a log in a database (DB). For example, the collected RAN data (D2, C2) and the inferred inference result data (N1) are accumulated.

[0050] The Near-RT RIC 100 transfers the accumulated RAN data (D2, C2) and inference result data (N1) as learning data to the Non-RT RIC 200 via the O1 interface. For example, the Near-RT RIC 100 effectively utilizes timing when the O1 interface is free to transmit to the Non-RT RIC 200. The learning model of the Non-RT RIC 200 performs learning using the RAN data (D2, C2) and inference result data (N1) collected from the Near-RT RIC 100 via the O1 interface as learning data.

[0051] As a result, in this embodiment, compared to the comparative example, either or both of the O-DU 300 and the O-CU 400 transmit RAN data only via the E2 interface, which reduces the load on either or both of the O-DU 300 and the O-CU 400. Furthermore, since data is transmitted only from the Near-RT RIC 100 via the O1 interface, the load on the O1 interface can also be reduced.

[0052] It should be noted that the comparative example of Fig. 8 may be implemented in combination with the present embodiment of Fig. 9. For example, some RAN data may be transferred from either or both of the O-DU 300 and the O-CU 400 to the Non-RT RIC 200 via the O1 interface as in Fig. 8, and other RAN data may be transferred from the Near-RT RIC 100 to the Non-RT RIC 200 via the O1 interface as in Fig. 9.

[0053] 10 shows an example of the configuration of a Near-RT RIC 100 according to this embodiment. As shown in FIG. 10, the Near-RT RIC 100 includes an E2 communication unit 101, an O1 communication unit 102, an A1 communication unit 103, a data collection unit 111, a data analysis unit 112, a control content identification unit 113, a model storage unit 120, a learning data accumulation control unit 131, a learning database 132, and a learning data transmission unit 133. Note that this configuration is merely an example, and other configurations may be used as long as the operation according to this embodiment, which will be described later, is possible. Furthermore, a configuration for realizing functions required for the Near-RT RIC 100 may also be included.

[0054] The E2 communication unit 101 is a communication unit that communicates with the O-DU 300 and the O-CU 400 via the E2 interface. For example, the E2 communication unit 101 transmits and receives various data including RAN data and control messages between the O-DU 300 and the O-CU 400 in accordance with a communication method defined as the E2 interface.

[0055] The O1 communication unit 102 is a communication unit that communicates with the Non-RT RIC 200 via the O1 interface. For example, the O1 communication unit 102 transmits and receives various data including learning data and control messages to and from the Non-RT RIC 200 in accordance with the communication method defined as the O1 interface.

[0056] The A1 communication unit 103 is a communication unit that communicates with the Non-RT RIC 200 via the A1 interface. For example, the A1 communication unit 103 transmits and receives control messages including various data and control policies to and from the Non-RT RIC 200 in accordance with the communication method defined as the A1 interface.

[0057] The data collection unit 111 collects RAN data from either or both of the O-DU 300 and the O-CU 400 via the E2 interface via the E2 communication unit 101. For example, the data collection unit 111 and the E2 communication unit 101 correspond to the collection unit 11 in FIG. 1 . The data collection unit 111 periodically collects RAN data as inference data to be used for inference by the inference model of the data analysis unit 112 and the control content identification unit 113. The data collection unit 111 may instruct either or both of the O-DU 300 and the O-CU 400 on the data to be collected and the collection period. The data collection unit 111 outputs the collected RAN data to the data analysis unit 112 and the learning data accumulation control unit 131.

[0058] The model storage unit 120 stores an inference model 121 that the data analysis unit 112 and the control content identification unit 113 use for analysis processing and control processing. The inference model 121 is a trained model that infers control of an E2 node including either or both of the O-DU300 and the O-CU400 according to RAN data. The inference model 121 is, for example, a model that can analyze and predict time-series data. The inference model 121 may be a CNN (Convolutional Neural Network), an RNN (Recurrent Neural Network), an LSTM (Long-Short Term Model), or other neural networks. The inference model 121 is not limited to a neural network and may be other machine learning models. The model storage unit 120 may store multiple inference models 121. For example, the model storage unit 120 stores an inference model 121a for data analysis used by the data analysis unit 112 and an inference model 121b for control content identification used by the control content identification unit 113. The inference model 121a is an analysis model for analyzing data, and the inference model 121b is an identification model for identifying control content.

[0059] The data analysis unit 112 analyzes the RAN data collected by the data collection unit 111, i.e., data for inference. The function of the data analysis unit 112 may be realized by executing an xApp for data analysis processing, i.e., for wireless analysis processing. The data analysis unit 112 analyzes the RAN data using an inference model 121a for data analysis stored in the model storage unit 120. The data analysis unit 112 inputs the collected RAN data to the inference model 121a and analyzes the RAN data. The data analysis unit 112 outputs the analysis results of the RAN data, i.e., the inference results, as wireless analysis information. For example, if the RAN data is wireless quality data, the inference model 121a predicts future wireless quality from the input wireless quality data and outputs the predicted wireless quality results as wireless analysis information. The data analysis unit 112 outputs the wireless analysis information to the control content identification unit 113 and the learning data accumulation control unit 131.

[0060] The control content identification unit 113 identifies the control content of the E2 node, including either or both of the O-DU 300 and the O-CU 400, based on the analysis results of the RAN data analyzed by the data analysis unit 112, and performs control over the E2 node. The function of the control content identification unit 113 may be realized by executing an xApp for control content identification, i.e., for wireless control processing. The control content identification unit 113 may also control the E2 node in accordance with a control policy notified via the A1 interface via the A1 communication unit 103. The control content identification unit 113 identifies the control content (control information) of the E2 node using an inference model 121b for control content identification stored in the model storage unit 120. The control content identification unit 113 inputs wireless analysis information, which is the analysis result of the data analysis unit 112, into the inference model 121b and identifies the control content according to the wireless analysis information. For example, multiple control contents may be inferred and the control content to be used for control may be identified in accordance with the control policy. The control content specification unit 113 outputs the result of specifying the control content, i.e., the inference result, as wireless control information. For example, if the data analysis unit 112 predicts future wireless quality, it specifies the wireless intensity and modulation method to be set in the E2 node according to the predicted wireless quality, and outputs control information to be set in the corresponding E2 node as wireless control information. The control content specification unit 113 transmits wireless control information indicating the specified control content to either or both of the O-DU 300 and the O-CU 400 via the E2 communication unit 101 and the E2 interface. Furthermore, the control content specification unit 113 outputs the wireless control information to the learning data accumulation control unit 131.

[0061] The learning database 132 is a learning data storage unit that stores learning data. The learning data storage control unit 131 controls the learning data stored in the learning database 132. Specifically, the learning data storage control unit 131 acquires and selects data necessary for learning, and stores the acquired and selected data in the learning database 132. The learning data storage control unit 131 also performs integration processing to format the accumulated data for transmission. It can also be said that the learning data storage control unit 131 includes a learning data acquisition unit, a learning data selection unit, and a learning data integration unit.

[0062] For example, the learning data accumulation control unit 131 starts accumulation control of learning data in response to an instruction from the Non-RT RIC 200. The learning data accumulation control unit 131 acquires RAN data, which is inference data collected by the data collection unit 111, radio analysis information generated by the data analysis unit 112, and radio control information generated by the control content identification unit 113, and selects, i.e., extracts, learning data from the collected and acquired data. The learning data accumulation control unit 131 may collect and acquire all of the RAN data, radio analysis information, and radio control information, or may collect and acquire any of the data, or arbitrarily selected data. The radio analysis information and radio control information are also inference result data inferred by the inference model. The selected learning data is data used for learning by the learning model of the Non-RT RIC 200, and may be instructed by the Non-RT RIC 200 or may be predetermined. For example, the learning data to be selected may be determined for each learning model or application. The learning data accumulation control unit 131 may narrow down the learning data based on predetermined conditions. For example, if the RAN data includes location information of the UE, the learning data may be narrowed down based on the location information. For example, RAN data included in a specific area from which wireless quality or the like is desired to be acquired may be extracted. Furthermore, RAN data included in other areas may be extracted, excluding areas for which data has already been acquired.

[0063] It is also possible to store all collected and acquired RAN data, radio analysis information, and radio control information, and then select the learning data to be transmitted from the stored data when transmitting the data to the Non-RT RIC 200.

[0064] The learning data accumulation control unit 131 accumulates the selected learning data in the learning database 132. For example, the learning data may be accumulated in a predetermined format, or in a format instructed by the Non-RT RIC 200. The learning data accumulation control unit 131 may accumulate RAN data collected from a plurality of O-DUs 300 and a plurality of O-CUs 400. The learning data accumulation control unit 131 may also accumulate a plurality of pieces of radio analysis information and a plurality of pieces of radio control information analyzed and controlled for a plurality of O-DUs 300 and a plurality of O-CUs 400.

[0065] After starting to store training data, the training data storage control unit 131 terminates the storage in a predetermined case. For example, the storage of training data may be terminated at a timing or a predetermined timing instructed by the Non-RT RIC 200, when a specified period or a predetermined period has elapsed, or when a specified amount or a predetermined amount of data has been stored. When the storage is terminated, the training data storage control unit 131 integrates the training data stored in the training database 132 into a file for transmission. The integration of training data includes integrating RAN data of multiple O-DUs 300 and multiple O-CUs 400, and integrating RAN data of multiple O-DUs 300 and multiple O-CUs 400 with multiple pieces of radio analysis information and multiple pieces of radio control information. The integrated file may be referred to as a training data file or integrated data.

[0066] When integrating the training data, the training data accumulation control unit 131 may process or reshape the data as necessary. For example, the training data accumulation control unit 131 may compress the training data to reduce the data volume. Compression may be performed on the data before integration, or on the data after integration. For example, the training data accumulation control unit 131 may reduce the data by dimensionality reduction. Dimensionality reduction reduces multidimensional data to information with fewer dimensions without affecting the learning of the training model. For example, the feature quantities of the training data input to the training model during training may be reduced to a degree that does not reduce the accuracy of data analysis (radio analysis) or control content identification (radio control). Furthermore, deduplication may be performed to eliminate duplicate data, such as data from the same area, or the data may be compressed using a predetermined encoding method. Instead of compression, the integrated data may be divided into multiple pieces of data. For example, multiple training data files may be transmitted.

[0067] The learning data transmission unit 133 transmits the learning data integrated by the learning data accumulation control unit 131, i.e., the learning data file, to the non-RT RIC 200 via the O1 communication unit 102 and the O1 interface. For example, the learning data transmission unit 133 and the O1 communication unit 102 correspond to the transmission unit 12 in FIG. 1 . The learning data transmission unit 133 may control the transmission timing and transfer rate of the learning data file. The learning data transmission unit 133 may control the transmission timing and transfer rate in response to instructions from the non-RT RIC 200, or may control the transmission timing and transfer rate in response to the communication status of the O1 interface. For example, the learning data transmission unit 133 may transmit the learning data file when there is available transmission bandwidth in the O1 interface, and may delay the transmission timing of the learning data file when there is no available transmission bandwidth. The learning data file may be transmitted when the available transmission bandwidth is greater than a predetermined threshold, or when the available transmission bandwidth is greater than the bandwidth required for the file size to be transmitted. The transfer rate may also be controlled in accordance with the available transmission bandwidth. The available transmission bandwidth may be a bandwidth acquired by the O1 communication unit 102 by monitoring the traffic volume of the O1 interface, or may be a bandwidth predicted from the monitoring results.

[0068] Furthermore, when transmitting multiple pieces of training data, i.e., multiple training data files, the training data transmission unit 133 may control the transmission order of the training data files. For example, the training data may be transmitted in an order based on the priority of the training data included in the training data file. The priority may be set based on the order of data used by the training model for training, whether the data is required for training the training model, or whether the data is optional, or may be specified by the non-RT RIC 200. The transmission order of the training data files may be determined based on the highest priority of the training data included in the training data file. Furthermore, the training data may be integrated for each priority of the training data, and the transmission order of the training data files may be determined based on the priority of the integrated training data file. Furthermore, when training data is integrated and transmitted for each training model, the training data files may be transmitted in an order based on the priority of the training model (application). For example, the transmission order of the training data files may be determined based on the importance of the application or the urgency of training the training model. Furthermore, if there is traffic other than training data being transmitted via the O1 interface due to other applications, the transmission timing and transmission order may be adjusted between the traffic. For example, transmission scheduling may be performed taking into consideration desired delivery times among traffic for multiple uses.

[0069] 11 shows an example of the configuration of the Non-RT RIC 200 according to this embodiment. As shown in FIG. 11, the Non-RT RIC 200 includes an O1 communication unit 201, an A1 communication unit 202, a system management unit 211, a learning unit 212, and a model storage unit 220. Note that this configuration is just an example, and other configurations may be used as long as the operation according to this embodiment, which will be described later, is possible. Furthermore, the Non-RT RIC 200 may include a configuration for realizing functions required for the Non-RT RIC 200.

[0070] The O1 communication unit 201 is a communication unit that communicates with the Near-RT RIC 100 via the O1 interface. For example, the O1 communication unit 201 transmits and receives various data including learning data and control messages between the Near-RT 100 in accordance with the communication method defined as the O1 interface. The O1 communication unit 201 can also transmit and receive necessary data and control messages between the O-DU 300 and the O-CU 400 via the O1 interface.

[0071] The A1 communication unit 202 is a communication unit that communicates with the Near-RT RIC 100 via the A1 interface. For example, the A1 communication unit 202 transmits and receives control messages including various data and control policies to and from the Near-RT RIC 100 in accordance with the communication method defined as the A1 interface.

[0072] The system management unit 211 manages the settings and operation of the RAN system including the O-DU 300, O-CU 400, and Near-RT RIC 100. The functions of the system management unit 211 may be realized by executing an rApp for system management processing. For example, the system management unit 211 is a policy generation unit that generates a control policy. The system management unit 211 may generate the control policy based on instructions input from an operator or an external device, or may generate the control policy based on data acquired from any or all of the O-DU 300, O-CU 400, and Near-RT RIC 100. The system management unit 211 notifies the Near-RT RIC 100 of the generated control policy via the A1 communication unit 202 and the A1 interface.

[0073] The model storage unit 220 stores a learning model 221 for constructing the inference model 121 of the Near-RT RIC 100. The learning model 221 is the same model as the inference model 121, and is, for example, a model that performs learning so as to be able to analyze and predict time-series data. As with the Near-RT RIC 100, the model storage unit 220 can store multiple learning models 221, and stores, for example, a learning model 221a for data analysis and a learning model 221b for specifying control content.

[0074] The learning unit 212 performs machine learning using training data, i.e., training data files, received from the Near-RT RIC 100 via the O1 communication unit 201 and the O1 interface. For example, the learning unit 212 corresponds to the learning unit 22 in FIG. 2. The function of the learning unit 212 may be realized by executing an rApp for the training process. The learning unit 212 can also be said to include a receiving unit that receives training data files from the Near-RT RIC 100 via the O1 interface. For example, the O1 communication unit 201 and the learning unit 212 also correspond to the receiving unit 21 in FIG. 2. If the received training data files are compressed or split, the learning unit 212 performs decompression or joining processes as necessary. The learning unit 212 performs machine learning such as deep learning to generate a trained learning model 221. The learning unit 212 inputs a learning data file into a learning model 221a for data analysis in the model storage unit 220 and a learning model 221b for control content identification in the model storage unit 220, and trains each model. The received learning data file may be subjected to the necessary data processing for inputting the learning models 221a and 221b. Furthermore, similar to the Near-RT RIC 100, a learning unit for data analysis and a learning unit for control content identification may be provided. The learning data in the received learning data file includes RAN data from either or both of the O-DU 300 and the O-CU 400, and inference result data (radio analysis information and radio control information) from the Near-RT RIC 100. By using these data, analysis and control according to the RAN data can be learned. In other words, accurate training can be achieved by using the RAN data before control input into the inference model, the inference result data output from the inference model, and the RAN data after control based on the inference result as learning data. The learning data may be any of RAN data, wireless analysis information, and wireless control information. The learning unit 212 stores the learned learning models 221a and 221b in the model storage unit 220, and further transmits the learned learning models 221a and 221b to the Near-RT RIC 100 and applies them to the inference models 121a and 121b.

[0075] 12 shows an overview of the learning process in the RAN system 1 according to this embodiment. As shown in FIG. 12, first, a learning model is prepared and placed in the RAN system 1 (S101). For example, an external learning device generates a learned model trained (learned) by simulation, and the generated model is stored as an inference model (xApp) in the model storage unit 120 of the Near-RT RIC 100. The same model is also stored as a learning model (rApp) in the model storage unit 220 of the Non-RT RT 200. Instead of simulation, a model equipped with a predetermined algorithm may be placed in the Near-RT RIC 100 as an inference model, and the same model may be placed in the Non-RT RIC 200 as a learning model.

[0076] Next, the RAN system 1 performs initial learning when the system is installed (S102). When the system is installed, the Near-RT RIC 100's inference model (xApp) is operated or trial-run in a real environment to collect learning data. Furthermore, the learning data is transferred from the Near-RT RIC 100 to the Non-RT RIC 200, and the learning model (rApp) is trained using the transferred learning data. For example, one hour's worth of learning data may be collected and used for training. The trained learning model learned through the initial learning is applied to the Near-RT RIC 100 as an inference model. Note that, because the initial learning has little impact on the system load, as in the comparative example, the Non-RT RIC 200 may collect RAN data from either or both the O-DU 300 and the O-CU 400 via the O1 interface.

[0077] Next, the RAN system 1 determines whether relearning is necessary (S103), and if relearning is necessary, performs the second or subsequent relearning (S104). For example, the Non-RT RIC 200 determines that relearning is necessary and performs relearning when an instruction is input from an operator or an external device, when the on-site environment including the UE changes, at regular intervals, or when the accuracy of the inference model decreases. The environmental change may be detected from a change in wireless quality, or a signal indicating an environmental change such as a layout change may be input. The accuracy of the inference model may be determined from the inference results of the inference model and RAN data, etc. In relearning, the Near-RT RIC 100 uses the inference model (xApp) that was previously trained and applied, and collects training data during actual operation while operating the inference model in a real environment. Furthermore, the Near-RT RIC 100 transfers the training data to the Non-RT RIC 200, and the training model (rApp) is trained using the transferred training data. For example, one hour's worth of learning data may be collected and used for training. The learned model obtained by relearning is applied as an inference model to the Near-RT RIC 100. Thereafter, relearning is repeatedly performed in steps S103 and S104.

[0078] Fig. 13 shows an example of the operation of the learning process in the RAN system 1 according to this embodiment. For example, Fig. 13 shows an example of the learning process in the re-learning (S104) shown in Fig. 12, but it may also be applied to the learning process in the initial learning (S102).

[0079] 13, the Near-RT RIC 100 starts storing learning data (S201). For example, the learning unit 212 instructs the Near-RT RIC 100 to start learning. In response to the instruction from the Non-RT RIC 200, the learning data storage control unit 131 starts storing learning data in the learning database 132.

[0080] Next, the Near-RT RCI 100 determines whether the accumulation of the learning data has been completed (S202), and if the accumulation has been completed, integrates the learning data (S203). For example, when the period instructed by the Non-RT RIC 200 has ended, the learning data accumulation control unit 131 integrates the learning data accumulated in the learning database 132 into a learning data file.

[0081] Next, the Near-RT RIC 100 transfers the learning data file to the Non-RT RIC 200 (S204). For example, the learning data transmission unit 133 transfers the integrated learning data file to the Non-RT RIC 200 at the timing instructed by the Non-RT RIC 200 or when the O1 interface is free.

[0082] Next, the Non-RT RIC 200 performs a learning process for the learning model using the received learning data file (S205). When the learning unit 212 receives the learning data file from the Near-RT RIC 100, it trains the learning models 221a and 221b using the learning data included in the received learning data file. When the learning unit 212 finishes learning, it stores the learned learning models 221a and 221b in the model storage unit 220.

[0083] Next, the Non-RT RIC 200 applies the trained learning model to the Near-RT RIC 100 (S206). The learning unit 212 transmits the trained learning models 221a and 221b stored in the model storage unit 220 to the Near-RT RIC 100. The data analysis unit 112 and the control content identification unit 113 update the inference models 121a and 121b in the model storage unit 120 using the received learning models 221a and 221b. The Near-RT RIC 100 performs data analysis and control using the updated inference models 121a and 121b.

[0084] Fig. 14 shows a sequence diagram of the learning process in the RAN system 1 according to this embodiment. For example, Fig. 14 shows the operations from accumulation to transfer of the learning data (S201 to S204) shown in Fig. 13.

[0085] As shown in FIG. 14, the Near-RT RIC 100 collects RAN data from either or both of the O-DU 300 and the O-CU 400 (S301). The data collection unit 111 repeatedly collects RAN data from either or both of the O-DU 300 and the O-CU 400 via the E2 interface. For example, the data collection unit 111 instructs either or both of the O-DU 300 and the O-CU 400 on the data and period to be collected, and either or both of the O-DU 300 and the O-CU 400 may transmit data according to the instruction, or either or both of the O-DU 300 and the O-CU 400 may transmit data according to predetermined data and period. In addition, the data analysis unit 112 analyzes the collected RAN data to generate radio analysis information, and the control content identification unit 113 infers radio control information according to the radio analysis information. The control content specification unit 113 transmits radio control information to either or both of the O-DU 300 and the O-CU 400 via the E2 interface, and controls the operation of either or both of the O-DU 300 and the O-CU 400. In parallel with the repeated collection and control of RAN data by S301, the processes of S302 to S307 are performed. That is, the Near-RT RIC 100 accumulates and transfers learning data while collecting inference data and performing control using an inference model.

[0086] First, the Non-RT RIC 200 transmits a learning instruction message to the Near-RT RIC 100 (S302). The learning unit 212 transmits the learning instruction message to the Near-RT RIC 100 via the O1 interface in response to instructions from the operator or changes in the environment. The learning instruction message is a message that instructs the Near-RT RIC 100 to start collecting (accumulating) learning data for learning a learning model. For example, the learning instruction message includes a target learning model and a learning data collection period. The target learning model indicates identification information that identifies the learning model (learning engine) to be learned, and the learning data collection period indicates the time to start and end collection of learning data (training data). The learning instruction message is not limited to the target learning model, but may also include information that identifies the learning data to be collected. The learning data collection period may specify the length of time for collecting learning data. The learning instruction message is not limited to the learning data collection period, but may also include a cycle for collecting learning data. The target learning model and learning data collection period may be set in advance or by the operator. For example, when a change in the environment is detected, the target learning model and the learning data collection period may be set in accordance with the change in the environment. When the wireless quality changes, the learning model of the application corresponding to the location of the changed wireless quality may be determined as the target learning model. The learning instruction message may be transmitted via the O1 interface or the A1 interface. In the Near-RT RIC 100, the learning data transmission unit 133 receives the learning instruction message via the O1 or A1 interface, and the learning data accumulation control unit 131 acquires the received learning instruction message.

[0087] Next, when the learning data collection period specified in the received learning instruction message begins, the Near-RT RIC 100 begins accumulating learning data (S303). The learning data accumulation control unit 131 selects the learning data necessary for learning the target learning model specified in the learning instruction message. For example, the learning model and learning data are associated with each other in advance, and the learning data corresponding to the target learning model specified in the learning instruction message is identified. Furthermore, when the learning instruction message specifies the learning data to be collected, the specified learning data is identified as the data to be accumulated. The learning data accumulation control unit 131 selects (extracts) the identified learning data from any or all of the RAN data collected by the data collection unit 111, the radio analysis information generated by the data analysis unit 112, and the radio control information generated by the control content identification unit 113, and accumulates the selected learning data in the learning database 132. Necessary data may be accumulated each time the data collection unit 111 collects RAN data, each time the data analysis unit 112 generates radio analysis information, and each time the control content identification unit 113 generates radio control information. If a learning data collection cycle is specified in the learning instruction message, data may be accumulated according to the specified cycle. Note that the Near-RT RIC 100 may voluntarily start accumulating learning data without receiving a learning instruction message from the Non-RT RIC 200. In this case, predetermined learning data may be accumulated for a predetermined period.

[0088] Next, when the training data collection period specified in the training instruction message ends, the Near-RT RIC 100 integrates the training data (S304). The training data accumulation control unit 131 formats the training data accumulated in the training database 132 and integrates it into a single training data file. For example, the training model and the format of the training data file may be associated with each other in advance, and the training data may be formatted and integrated into the training data file in accordance with a format corresponding to the target training model specified in the training instruction message. A format may be specified in the training instruction message, and the training data may be integrated according to the specified format. Furthermore, the integrated training data file may be compressed as needed. For example, the training instruction message may specify a compression method such as dimensionality reduction, and the data may be compressed using the specified compression method, or the data may be compressed using a compression method corresponding to the target training model specified in the training instruction message.

[0089] Next, the Near-RT RIC 100 transmits a collection end message to the Non-RT RIC 200 (S305). Once the learning data has been integrated, the learning data transmission unit 133 transmits a collection end message to the Non-RT RIC 200 via the O1 interface. The collection end message is a message that notifies that all learning data has been collected, i.e., that the collection (accumulation) of the learning data has been completed. The collection end message may include a target learning model, similar to the learning instruction message. The collection end message may be transmitted via either the O1 interface or the A1 interface. In the Non-RT RIC 200, the learning unit 212 receives the collection end message via the O1 or A1 interface.

[0090] Next, the Non-RT RIC 200 transmits a transfer instruction message to the Near-RT RIC 100 (S306). Upon receiving the collection end message, the learning unit 212 transmits a transfer instruction message to the Near-RT RIC 100 via the O1 interface. The transfer instruction message is a message that instructs the Near-RT RIC 100 to transfer the collected training data. For example, the transfer instruction message includes timing and bit rate. The timing indicates the transmission timing of the training data and may be the transmission time, etc. The bit rate indicates the transmission bit rate (bps) of the training data. Like the training instruction message, the transfer instruction message may also include a target training model. Note that the transfer instruction message may be transmitted via either the O1 interface or the A1 interface. In the Near-RT RIC 100, the training data transmission unit 133 receives the transfer instruction message via the O1 or A1 interface.

[0091] Next, the Near-RT RIC 100 transmits the training data file to the Non-RT RIC 200 (S307). At the timing specified in the received transfer instruction message, the training data transmission unit 133 transmits the integrated training data file to the Non-RT RIC 200 via the O1 interface at the bit rate specified in the transfer instruction message. For example, the training data file may be transmitted if the transmission bandwidth of the O1 interface is available after the specified timing. The training data file may be transmitted via either the O1 interface or the A1 interface. For example, if the transmission bandwidth of the O1 interface is not available, the training data file may be transmitted via the A1 interface. The Near-RT RIC 100 may transmit the training data file autonomously without receiving a transfer instruction from the Non-RT RIC 200. The training data file may be transmitted when the accumulation of training data is completed and the training data file is generated, or at a predetermined timing.

[0092] As described above, in this embodiment, when the Near-RT RIC receives a learning instruction, for example, it accumulates data required for learning from the data used in the inference model inference as a log, in parallel with inference by an inference model including radio analysis and radio control. Furthermore, after collecting the log, the Near-RT RIC transmits the accumulated data to the Non-RT RIC via the O1 interface, for example, effectively utilizing timing when the O1 interface is free, and the Non-RT RIC performs learning of the learning model. This eliminates the communication load solely for the purpose of acquiring and collecting learning data, thereby reducing the communication load on either or both the O-DU and O-CU and reducing the required communication resources. Furthermore, for example, by effectively utilizing timing when the O1 interface is free, the Non-RT RIC can collect learning data, thereby distributing the communication load on the O1 interface and reducing the communication resources required for the O1 interface.

[0093] (Embodiment 2) Next, embodiment 2 will be described. In this embodiment, an example of learning using Near-RT RIC will be described. Note that this embodiment can be implemented in combination with embodiment 1, and may be implemented by using the configuration of embodiment 1 as appropriate.

[0094] 15 shows an example of the configuration of the Near-RT RIC 100 according to this embodiment. For example, as shown in FIG. 15, the Near-RT RIC 100 according to this embodiment includes a learning unit 141 in addition to the configuration of the first embodiment. The other configurations are the same as those of the first embodiment.

[0095] The learning unit 141 performs learning using inference data and applies the learned learning model to the inference model. The function of the learning unit 141 may be realized by executing an xApp for learning processing. Similar to the learning unit 212 of the Non-RT RIC 200, the learning unit 141 trains the learning model using any or all of the RAN data collected by the data collection unit 111, the radio analysis information analyzed by the data analysis unit 112, and the radio control information identified by the control content identification unit 113. The learning unit 141 may acquire necessary data from the data collection unit 111, the data analysis unit 112, and the control content identification unit 113, or may acquire necessary data from the learning database 132. The learning unit 141 performs learning on some of the inference models 121 stored in the model storage unit 120. Learning may be performed for either the inference model 121a or the inference model 121b. For example, when learning can be limited to a specific cell, as in wireless analysis such as time-series analysis of SINR (Signal-to-Interference plus Noise power Ratio), learning can be performed with relatively light processing, and therefore learning is performed in the learning unit 141 of the Near-RT RIC 100. For example, the learning unit 141 generates a learning model that learns wireless analysis using the SINR of the specific cell, and updates the inference model 121a in the model storage unit 120 with the generated learned learning model.

[0096] In this way, learning can be performed in the Near-RT RIC, and the learned learning model can be applied to the inference model inside the Near-RT RIC. This allows inference and learning to be performed inside the Near-RT RIC, enabling high-speed learning and reducing the load on the O1 interface and the Non-RT RIC.

[0097] (Third Embodiment) Next, a third embodiment will be described. In this embodiment, an example will be described in which the data collection frequency and the time granularity of the data to be transferred are adjusted in accordance with changes in the data to be collected. Note that this embodiment can be implemented in combination with either the first or second embodiment, and may be implemented by using the configuration of either the first or second embodiment as appropriate.

[0098] 16 shows an example of the configuration of the Near-RT RIC 100 according to this embodiment. For example, as shown in FIG. 16, the Near-RT RIC 100 according to this embodiment includes a data change detection unit 142 in addition to the configuration of the first embodiment. The other configurations are the same as those of the first embodiment.

[0099] The data change detection unit 142 detects changes in the RAN data acquired from either or both of the O-DU 300 and the O-CU 400. For example, it detects the magnitude of changes in the radio quality data, or the UE's moving speed from the UE's location information. The function of the data change detection unit 142 may be included in the data analysis unit 112 or another xApp. For example, the data analysis unit 112 may output radio analysis information including the detection results of the data change.

[0100] In this embodiment, the data collection unit 111 adjusts the data collection frequency according to the detection result of the data change detection unit 142. For example, the data collection unit 111 adjusts the collection frequency from either or both of the O-DU 300 and the O-CU 400 according to the intensity of the time-series changes in the radio quality data and the moving speed of the UE. As the changes in the radio quality data increase, the data collection frequency may be increased, i.e., the data collection interval may be shortened, and as the changes in the radio quality data decrease, the data collection frequency may be decreased, i.e., the data collection interval may be lengthened. Furthermore, as the moving speed of the UE increases, the data collection frequency may be increased, i.e., the data collection interval may be shortened, and as the moving speed of the UE decreases, the data collection frequency may be decreased, i.e., the data collection interval may be lengthened.

[0101] Furthermore, the learning data accumulation control unit 131 may adjust the time granularity of the data accumulated in the learning database 132, i.e., the time granularity of the data to be transferred as learning data, according to the detection result of the data change detection unit 142. For example, the learning data accumulation control unit 131 adjusts the time granularity of the data to be accumulated and transferred according to the intensity of time-series changes in the radio quality data and the moving speed of the UE. For example, the time granularity of the data can be made coarse by reducing the number of data items per unit time. As the change in the radio quality data decreases, the time granularity of the data may be made coarser, i.e., the interval between the time-series data may be lengthened. As the moving speed of the UE decreases, the time granularity of the data may be made coarser, i.e., the interval between the time-series data may be lengthened.

[0102] In this way, when collecting data using the Near-RT RIC, the data collection frequency and the time granularity of the data to be transferred may be adjusted depending on changes in the collected data. By adjusting the data collection frequency, it is possible to reduce the load on either or both of the O-DU and O-CU, as well as the load on the E2 interface. Furthermore, by adjusting the time granularity of the data to be transferred, it is possible to further reduce the load on the O1 interface.

[0103] (Fourth Embodiment) Next, a fourth embodiment will be described. In this embodiment, an example will be described in which data to be stored and transferred is selected in accordance with changes in the environment. Note that this embodiment can be implemented in combination with any of the first to third embodiments, and may also be implemented by using any of the configurations of the first to third embodiments as appropriate.

[0104] 17 shows an example of the configuration of the Near-RT RIC 100 according to this embodiment. For example, as shown in Fig. 17, the Near-RT RIC 100 according to this embodiment includes an environmental change detection unit 143 in addition to the configuration of the first embodiment. The other configurations are the same as those of the first embodiment.

[0105] The environmental change detection unit 143 detects environmental changes based on RAN data acquired from either or both of the O-DU 300 and the O-CU 400. The environmental change is a change in the UE's radio environment. For example, a change in the radio environment that makes it necessary to train a learning model, such as a change in the radio environment caused by a layout change, is detected. The environmental change detection unit 143 detects changes in radio quality, and determines that the radio environment has changed if the change in radio quality is greater than a predetermined value. The unit may also identify the location where the radio quality has changed significantly. The function of the environmental change detection unit 143 may be included in the data analysis unit 112 or another xApp. For example, the data analysis unit 112 may output radio analysis information including the detection results of the environmental change.

[0106] In this embodiment, the learning data accumulation control unit 131 selects data necessary for learning, i.e., data to be accumulated, in accordance with environmental changes detected by the environmental change detection unit 143. For example, in locations where wireless quality has changed significantly, it is highly likely that the wireless environment has changed due to changes in the placement or layout of obstructions, making it necessary to train the learning model. For this reason, the learning data is narrowed down to data on locations (areas) where wireless quality has changed significantly. The learning data accumulation control unit 131 narrows down data on identified locations from the collected RAN data and accumulates the narrowed down learning data in the learning database 132, thereby reducing unnecessary data.

[0107] In this way, when collecting and storing data in Near-RT RIC, the data to be stored may be selected according to changes in the environment. This reduces the amount of data for training, further reducing the load on the O1 interface.

[0108] (Fifth Embodiment) Next, a fifth embodiment will be described. In this embodiment, an example of acquiring data from an external application server will be described. Note that this embodiment can be implemented in combination with any of the first to fourth embodiments, and may be implemented by appropriately using any of the configurations of the first to fourth embodiments.

[0109] Fig. 18 shows an example of the configuration of a RAN system 1 according to this embodiment. As shown in Fig. 18, the RAN system 1 according to this embodiment includes an external application server 500 in addition to the configuration of embodiment 1. Other configurations are the same as, for example, embodiment 1. The Near-RT RIC 100 and the external application server 500 are connected to each other so that they can communicate with each other via any interface. They may also be connected by an interface that allows a general application server to provide data. Note that the Non-RT RIC 200 and the external application server 500 may also be connected to each other so that they can communicate with each other.

[0110] The external application server 500 is a server external to the RAN including at least the O-DU 300 and the O-CU 400. The external application server 500 can also be said to be a server external to a system including the O-DU 300, the O-CU 400, the Non-RT RIC 200, and the Near-RT RIC 100. The external application server 500 is a data providing device that provides application data. For example, the external application server 500 may be a web server or a social networking service (SNS) server, a management server for an automated guided vehicle (AGV) system or an autonomous mobile robot (AMR) system, a management server for an automobile platooning system or an autonomous driving system, a management server for an automated construction system, or a server that executes an application that provides a predetermined service in cooperation with the RAN. The external application server 500 may be, for example, a server on the Internet as long as it can provide application data to the Near-RT RIC 100. The external application server 500 may be a physical server or a virtual server on the cloud.

[0111] The application data is data related to an application generated by the external application server 500, and is external data that cannot be collected within the RAN. The application data may include information related to the application status or information related to the application's communication requirements. The information related to the application status is, for example, the location, movement speed, and movement trajectory of each UE. The information related to the application's communication requirements is, for example, communication quality requirements of each UE. The application data may also include various other data that can be used for inference of the inference model of the Near-RT RIC 100. For example, the application data may include weather information, traffic information, map information, and the like.

[0112] In this embodiment, the Near-RT RIC 100 collects and uses for inference application data from the external application server 500. That is, the Near-RT RIC 100 uses RAN data collected from either or both of the O-DU 300 and the O-CU 400 and application data collected from the external application server 500 as inference data, and an inference model performs inference and controls either or both of the O-DU 300 and the O-CU 400.

[0113] Further, the Near-RT RIC 100 accumulates application data used for inference and transfers it as learning data to the Non-RT RIC 200. That is, data selected from the RAN data and application data, which are inference data, and the radio analysis information and radio control information, which are inference result data, is accumulated in the learning database 132 as learning data, and the accumulated learning data is transferred to the Non-RT RIC 200. The Near-RT RIC 100 acquires the transferred learning data and performs learning of the learning model using, for example, the RAN data, application data, radio analysis information, and radio control information included in the learning data.

[0114] In this way, the Near-RT RIC may acquire application data from an external application server and use it for inference together with RAN data from the E2 node. This allows for wireless network control suitable for various situations and applications. Furthermore, the application data used for inference in the Near-RT RIC is transferred to the Non-RT RIC as training data along with RAN data, etc. This allows for training data that is consistent with the RAN data, etc., and application data to be created and transmitted to the Non-RT RIC 200. Furthermore, since there is no need to collect application data in duplicate as inference data and training data, the communication load on the external application server and on the communication lines between the external application server and the system including the O-DU 300, O-CU 400, Non-RT RIC 200, and Near-RT RIC 100 can be reduced.

[0115] The present disclosure is not limited to the above-described embodiment, and can be modified as appropriate within the scope of the present disclosure.

[0116] Each component in the above-described embodiments may be configured with hardware, software, or both, and may be configured with a single piece of hardware or software, or may be configured with multiple pieces of hardware or software. Each device, including the Non-RT RIC and Near-RT RIC, and each function (processing) may be implemented by a computer 50 having a network interface 51, a processor 52 such as a CPU (Central Processing Unit), and a memory 53 serving as a storage device, as shown in FIG. 19 . The network interface 51 may include a network interface card (NIC) for communicating with devices including network nodes. For example, a program for performing the method (control method) in the embodiment may be stored in the memory 53, and each function may be implemented by executing the program stored in the memory 53 by the processor 52.

[0117] These programs include instructions (or software code) that, when loaded into a computer, cause the computer to perform one or more functions described in the embodiments. The programs may be stored on a non-transitory computer-readable medium or a tangible storage medium. By way of example and not limitation, computer-readable media or tangible storage media include random-access memory (RAM), read-only memory (ROM), flash memory, solid-state drive (SSD) or other memory technology, CD-ROM, digital versatile disc (DVD), Blu-ray disc or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage device. The programs may also be transmitted on a transitory computer-readable medium or communication medium. By way of example and not limitation, transitory computer-readable media or communication media include electrical, optical, acoustic, or other forms of propagated signals.

[0118] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above-described embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure.

[0119] Some or all of the above embodiments can be described as, but are not limited to, the following supplementary notes. (Supplementary Note 1) A control system comprising: a collection means for collecting identification data used by an identification model that identifies control related to a wireless network; and a transmission means for transmitting the identification data to another control system that performs learning using the learning model as learning data to be used by a learning model that learns control related to the wireless network. (Supplementary Note 2) The control system according to Supplementary Note 1, wherein the identification model is constructed by a learning model that has performed the learning. (Supplementary Note 3) The control system according to Supplementary Note 1 or 2, wherein the learning data includes identification result data that is a result of identification by the identification model using the identification data. (Supplementary Note 4) The control system according to any one of Supplements 1 to 3, further comprising: a selection means for selecting, from the collected identification data, learning data to be used by the learning model for learning, and the transmission means for transmitting the selected learning data. (Supplementary Note 5) The control system according to Supplementary Note 4, wherein the selection means selects the learning data based on a position of a user terminal corresponding to the identification data. (Supplementary Note 6) The selection means selects the learning data based on an environmental change corresponding to the identification data. The control system according to Supplementary Note 4 or 5. (Supplementary Note 7) The control system according to any one of Supplements 1 to 6, further comprising: storage means for storing the learning data to be transmitted; and the transmission means for transmitting integrated data that integrates the accumulated learning data. (Supplementary Note 8) The control system according to Supplementary Note 7, wherein the storage means stores learning data including identification data collected from a plurality of nodes included in the wireless network. (Supplementary Note 9) The control system according to any one of Supplements 1 to 8, wherein the transmission means transmits the learning data depending on a communication status between the control system and the other control system. (Supplementary Note 10) The control system according to any one of Supplementary Notes 1 to 9, wherein the transmission means transmits the learning data in response to other traffic transmitted from the control system to the other control system.(Supplementary Note 11) The control system according to any one of Supplements 1 to 10, wherein, when transmitting a plurality of pieces of training data, the transmission means transmits the training data according to a priority of each piece of training data. (Supplementary Note 12) The control system according to any one of Supplements 1 to 11, wherein the transmission means transmits compressed data obtained by dimensionally compressing the training data. (Supplementary Note 13) The control system according to any one of Supplements 1 to 12, wherein the collection means adjusts a frequency of collecting the identification data according to a change in the collected identification data. (Supplementary Note 14) The control system according to any one of Supplements 1 to 13, wherein the transmission means adjusts a time granularity of the training data to be transmitted according to a change in the collected identification data. (Supplementary Note 15) The control system according to any one of Supplements 1 to 14, further comprising: learning means for generating a learning model trained using the identification data and updating the identification model with the generated learning model. (Supplementary Note 16) The control system according to any one of Supplements 1 to 15, wherein the control system and the other control system include a RAN Intelligent Controller (RIC) that controls a RAN (Radio Access Network). (Supplementary Note 17) The control system according to Supplementary Note 16, wherein the control system includes a Near-RT (Real Time) RIC, and the other control system includes a Non-RT RIC. (Supplementary Note 18) A control system comprising: a receiving means that receives, from the other control system, identification data collected by the other control system in order to identify control related to the wireless network using an identification model, as learning data; and a learning means that uses the learning data to learn control related to the wireless network using a learning model. (Supplementary Note 19) The control system according to Supplementary Note 18, wherein the learning means applies the learned learning model to the identification model.(Supplementary Note 20) A control device comprising: a collection means for an identification model that identifies a control related to a wireless network to collect identification data to be used for the identification; and a transmission means for transmitting the identification data to another control system that performs learning using the learning model as learning data to be used for a learning model that learns control related to the wireless network. (Supplementary Note 21) A control device comprising: a reception means for receiving, from the other control system, identification data collected by the other control system in order for the other control system to identify a control related to the wireless network using the identification model, as learning data; and a learning means for learning control related to the wireless network using the learning model using the learning data. (Supplementary Note 22) A control method, in which an identification model that identifies a control related to a wireless network collects identification data to be used for the identification, and transmits the identification data to another control system that performs learning using the learning model as learning data to be used for the learning model that learns control related to the wireless network. (Supplementary Note 23) A control method, comprising: receiving, from another control system, identification data collected by the other control system for identifying controls related to a wireless network using an identification model, as learning data; and using the learning data to learn controls related to the wireless network using a learning model. (Supplementary Note 24) A non-transitory computer-readable medium storing a control program for causing a computer to execute a process, wherein a identification model that identifies controls related to a wireless network collects identification data to be used for the identification, and transmits the identification data to another control system that performs learning using the learning model as learning data to be used for a learning model that learns controls related to the wireless network. (Supplementary Note 25) A non-transitory computer-readable medium storing a control program for causing a computer to execute a process, wherein another control system receives, from the other control system, identification data collected by the other control system for identifying controls related to a wireless network using an identification model, as learning data; and using the learning data to learn controls related to the wireless network using a learning model.

[0120] 1 RAN system 10 First control system 11 Collection unit 12 Transmission unit 20 Second control system 21 Reception unit 22 Learning unit 30 First control device 40 Second control device 50 Computer 51 Network interface 52 Processor 53 Memory 100 Near-RT RIC 101 E2 communication unit 102 O1 communication unit 103 A1 communication unit 111 Data collection unit 112 Data analysis unit 113 Control content identification unit 120 Model storage unit 121, 121a, 121b Inference model 131 Learning data accumulation control unit 132 Learning database 133 Learning data transmission unit 141 Learning unit 142 Data change detection unit 143 Environmental change detection unit 200 Non-RT RIC 201 O1 communication unit 202 A1 communication unit 211 System management unit 212 Learning unit 220 Model storage unit 221, 221a, 221b Learning model 300 O-DU 400 O-CU 500 External application server

Claims

1. a collection means for collecting identification data used by an identification model for identifying a control related to a wireless network; a transmission means for transmitting the identification data to another control system that performs learning using the learning model as learning data to be used in a learning model that learns control related to the wireless network; A control system comprising:

2. The specific model is constructed by a learning model that has undergone the learning. The control system of claim 1 .

3. The learning data includes identification result data that is a result of the identification model being identified using the identification data. A control system according to claim 1 or 2.

4. A selection means is provided for selecting learning data to be used for learning by the learning model from the collected identification data; The transmission means transmits the selected learning data. A control system according to claim 1 or 2.

5. The selection means selects the learning data based on a position of the user terminal corresponding to the identification data.

5. The control system of claim 4.

6. A receiving means for receiving, as learning data, identification data collected by the other control system in order to identify control related to the wireless network using a specific model from the other control system; a learning means for learning control regarding the wireless network by a learning model using the learning data; A control system comprising:

7. A characterization model for determining control over a wireless network collects characterization data for use in said characterization; transmitting the identification data to another control system that performs learning using the learning model as learning data to be used in a learning model that learns control related to the wireless network; Control methods.

8. receiving, as learning data, from another control system, identification data collected by the other control system in order to identify control related to the wireless network using a specific model; using the training data to learn a control for the wireless network using a training model; Control methods.

9. A characterization model for determining control over a wireless network collects characterization data for use in said characterization; transmitting the identification data to another control system that performs learning using the learning model as learning data to be used in a learning model that learns control related to the wireless network; A control program that causes a computer to execute processing.

10. receiving, as learning data, from another control system, identification data collected by the other control system in order to identify control related to the wireless network using a specific model; using the training data to learn a control for the wireless network using a training model; A control program that causes a computer to execute processing.