WO2023119623A1 - Management device, communication system, control method, and non-transitory computer-readable medium - Google Patents

Management device, communication system, control method, and non-transitory computer-readable medium Download PDF

Info

Publication number
WO2023119623A1
WO2023119623A1 PCT/JP2021/048186 JP2021048186W WO2023119623A1 WO 2023119623 A1 WO2023119623 A1 WO 2023119623A1 JP 2021048186 W JP2021048186 W JP 2021048186W WO 2023119623 A1 WO2023119623 A1 WO 2023119623A1
Authority
WO
WIPO (PCT)
Prior art keywords
engine
analysis
analysis result
management device
data
Prior art date
Application number
PCT/JP2021/048186
Other languages
French (fr)
Japanese (ja)
Inventor
正弘 渡邉
忠信 齋藤
祥平 市瀬
修平 高野
憲二 辻
英明 黒木
裕司 土屋
Original Assignee
日本電気株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 日本電気株式会社 filed Critical 日本電気株式会社
Priority to PCT/JP2021/048186 priority Critical patent/WO2023119623A1/en
Publication of WO2023119623A1 publication Critical patent/WO2023119623A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/445—Program loading or initiating
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning

Definitions

  • the present disclosure relates to management devices, communication systems, control methods, and non-transitory computer-readable media.
  • Cited Document 1 discloses a system in which a cloud-side server collects information such as temperature, vibration, speed, etc. of learning target devices installed in factories, etc., and uses AI to automatically control the learning target devices. It is
  • the AI engine is configured as software, and the software version is updated according to algorithm updates and the like.
  • the AI engine currently in operation may be changed to another vendor's AI engine. If an AI engine with a software version different from the AI engine running in the commercial environment, or an AI engine from a different vendor, is applied to the system in the commercial environment, analysis processing using data different from that in the test environment will be performed. Become. As a result, after verification in the test environment, the AI engine applied to the commercial environment may output analysis results with lower accuracy than expected, and the system in the commercial environment may not operate normally.
  • One of the objects of the present disclosure is a management device, a communication system, and a control method that can apply an AI engine to a system in a commercial environment so that the system in the commercial environment that uses data different from the test environment can be operated normally. and to provide a non-transitory computer-readable medium.
  • a management device includes a control unit that operates in parallel a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine.
  • a selection unit that selects either the first AI engine or the second AI engine based on a selection instruction signal containing identification information of the first AI engine or the second AI engine;
  • the analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis result of the second AI engine.
  • a communication unit that outputs to the controlled device that uses the.
  • a communication system includes: a first AI engine; a selection unit that selects either the first AI engine or the second AI engine based on a selection instruction signal containing identification information of the first AI engine or the second AI engine;
  • a management device comprising: a communication unit that transmits the analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine; and autonomous control using the analysis result. and a controlled device that executes
  • a control method operates in parallel with a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine, one of the first AI engine and the second AI engine is selected based on a selection instruction signal including identification information of one AI engine or the second AI engine;
  • a controlled device that uses the analysis result of the selected AI engine from the analysis result or the analysis result of the second AI engine, and the analysis result of the first AI engine or the analysis result of the second AI engine.
  • a program operates in parallel with a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine, selecting either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the AI engine or the second AI engine, and analyzing the first AI engine
  • the analysis result of the AI engine selected from the result or the analysis result of the second AI engine is sent to the controlled device that uses the analysis result of the first AI engine or the analysis result of the second AI engine. Make the computer do the output.
  • a management device a communication system, a control method, and a non-temporary system that can apply an AI engine to a system in a commercial environment so as to normally operate the system in the commercial environment using data different from the test environment
  • a computer readable medium can be provided.
  • FIG. 1 is a configuration diagram of a management device according to Embodiment 1;
  • FIG. 4 is a diagram showing the flow of control processing executed by the management device according to the first embodiment;
  • FIG. 1 is a configuration diagram of a communication system according to a second embodiment;
  • FIG. 2 is a configuration diagram of a management device according to a second embodiment;
  • FIG. 10 is a diagram showing a data flow in the management device according to the second embodiment;
  • FIG. FIG. 10 is a diagram showing the flow of analysis result transmission processing according to the second embodiment;
  • 4 is a configuration diagram of a management device according to each embodiment;
  • the management device 10 may be a computer device operated by a processor executing a program stored in memory.
  • the management device 10 may be, for example, a server device.
  • the management device 10 may perform operations or processes similar to those of a platform on which a plurality of computer devices operate cooperatively.
  • the management device 10 has a control unit 11, a selection unit 12, and a communication unit 13.
  • the control unit 11, the selection unit 12, and the communication unit 13 may be software or modules whose processing is executed by a processor executing a program stored in memory.
  • the control unit 11, the selection unit 12, and the communication unit 13 may be hardware such as circuits or chips.
  • the control unit 11 operates in parallel the first AI engine and the second AI engine that executes analysis processing for a common purpose with the first AI engine.
  • the analysis process may be a process of generating data indicating traffic transitions in which the passage of time is associated with the amount of traffic.
  • the analysis process may be, for example, a process for autonomously or automatically optimizing resources in a base station of a mobile communication system. Specifically, the analysis processing reduces the number of base stations that are operated during periods of low traffic and increases the number of base stations that are operated during periods of heavy traffic. It may be processing.
  • the analysis process may be, more specifically, a process of generating data indicating traffic transitions for each antenna of the base station.
  • the analysis processing may be analysis processing of location information of the terminal device for achieving optimization of beamforming in a base station having multiple antennas. More specifically, the analysis process may be a process of generating data indicating movement tendencies of terminal devices.
  • the analysis process may also include an optimum parameter value determination process or an optimum control content determination process, which is executed based on the analysis result.
  • the analysis processing may be to collect multiple data and statistically process the collected data. Also, the analysis processing may be performed using not only data of the same type but also data in which different types of data are combined.
  • the type of data may be, for example, data indicating changes in the amount of data transmitted and received for each time period, climate data, information regarding the date and time of an event such as a concert, and the like.
  • the AI engine may be software or an application with different algorithms depending on the service provided or the analysis content.
  • the software version of the AI engine may be updated by adding new functions, deleting existing functions, or updating existing functions.
  • the first AI engine and the second AI engine that execute analysis processing for a common purpose may be, for example, AI engines with different software versions.
  • the first AI engine and the second AI engine may be AI engines produced by different vendors.
  • Parallel operation may mean, for example, that the first AI engine and the second AI engine execute analysis processing at substantially the same timing.
  • parallel operation may be a partial overlap between the operating time of the first AI engine and the operating time of the second AI engine.
  • the parallel operation may be performing the analysis processing of the second AI engine after the analysis processing of the first AI engine is performed.
  • parallel operation may mean that a plurality of analysis results are generated as a result of executing analysis processing in each AI engine.
  • Parallel operation may be rephrased as parallel operation.
  • the control unit 11 may, for example, control the timing of operating the first AI engine and the second AI engine, and may control the duration of analysis processing in each AI engine. Further, the control unit 11 may output data used for analysis processing to each of the first AI engine and the second AI engine.
  • the selection unit 12 selects either the first AI engine or the second AI engine based on the selection instruction signal containing the identification information of the first AI engine or the second AI engine.
  • the identification information is information that uniquely identifies the AI engine, and may be, for example, information indicating the name given to the AI engine. Also, the identification information may include information indicating the software version. For the identification information, for example, the administrator of the management device 10 may manage the combination of each AI engine and the identification information, and the processor or the like in the management device 10 may manage the identification information given to each AI engine. It may be generated autonomously.
  • the selection instruction signal includes identification information of the AI engine, and the selection unit 12 selects either the first AI engine or the second AI engine based on the identification information included in the selection instruction signal.
  • the selection instruction signal may include, for example, identification information of either the first AI engine or the second AI engine input by a manager or the like who manages the management device 10 .
  • the selection instruction signal may be generated when the administrator inputs the identification information.
  • the selection instruction signal may be an input signal input by the administrator through the input interface.
  • the input interface may be, for example, a keyboard, a touch panel, or the like, or may be a voice input interface such as a microphone.
  • the communication unit 13 sends the analysis result of the selected AI engine from the analysis result of the first AI engine or the analysis result of the second AI engine to the analysis result of the first AI engine or the analysis result of the second AI engine. Output the analysis results to the control target device that uses them.
  • a device to be controlled is a device that can be controlled using software.
  • it may be a base station that executes autonomous control such as resource optimization, or it may be another device that constitutes a mobile communication system. good.
  • the controlled device that receives the analysis result may perform autonomous control such as resource optimization based on the analysis result.
  • the communication unit 13 may discard or discard the analysis results output from the unselected AI engines without outputting them to the control target device.
  • the analysis results output from the AI engines that have not been selected may be stored in the memory within the management device 10 .
  • the analysis results stored in the management device 10 may be managed, for example, as history information when verification is executed.
  • the selection unit 12 selects either a first AI engine or a second AI engine that executes analysis processing for a common purpose with the first AI engine and operates in parallel with the first AI engine. (S11).
  • the selection unit 12 selects either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine.
  • the communication unit 13 outputs the analysis result of the AI engine selected by the selection unit 12 out of the analysis result of the first AI engine or the analysis result of the second AI engine to the device to be controlled (S12).
  • the controlled device executes autonomous control and the like using the analysis result of the first AI engine or the analysis result of the second AI engine.
  • the management device 10 selects at least one analysis result from among the analysis results of a plurality of AI engines that execute analysis processing for a common purpose, and outputs the selected analysis result to the controlled device. do.
  • the management device 10 can operate two AI engines with different software versions in parallel in a commercial environment. can. For example, when the administrator or the like determines that the analysis result of the second AI engine is closer to the desired result than the analysis result of the first AI engine, the management device 10 performs the analysis of the second AI engine. The results can be applied to autonomous control or the like in the controlled device. On the other hand, when the administrator or the like determines that the analysis result of the first AI engine is closer to the desired result than the analysis result of the second AI engine, the management device 10 performs the analysis of the first AI engine. The results can be applied to autonomous control or the like in the controlled device.
  • the management device 10 operates AI engines of different versions in parallel in a commercial environment, and uses the AI engine of the old software version until the analysis result of the AI engine of the new software version does not satisfy the desired result. keep applying.
  • the first AI engine and the second AI engine are not limited to different software versions, and may be AI engines generated by different vendors, for example.
  • the communication system of FIG. 3 has a management device 20 and a radio device 30.
  • the wireless device 30 is a device included in the controlled device.
  • the management device 20 manages the wireless device 30 .
  • the management device 20 managing the wireless device 30 may include, for example, the management device 20 controlling the operation of the wireless device 30 and determining the values of various parameters set in the wireless device 30 .
  • management device 20 may generate information necessary for optimizing functions executed by wireless device 30 and transmit the generated information to wireless device 30 .
  • the management device 20 and the wireless device 30 may communicate via an IP network, for example.
  • the wireless device 30 actually provides wireless services to mobile communication devices, and the communication system configured by the management device 20 and the wireless device 30 can be said to be a system operating in a commercial environment.
  • the management device 20 may be, for example, an RIC (RAN Intelligent Controller) that configures SMO (Service Management and Orchestration) defined in the O-RAN (Open Radio Access Network) Alliance.
  • the RIC may include Non-Real Time RIC and Near-Real Time RIC with different control cycles.
  • the RIC manages the AI engine and performs data processing.
  • the RIC provides data to the AI engine and controls the wireless device based on the analysis results of the AI engine.
  • the AI engine may be an application that uses a different algorithm for each service content or analysis process.
  • the application may be rAPP (Non-Real Time RIC Application) defined in the O-RAN Alliance.
  • the AI engine may be rephrased as an application, and rAPP may be used as a specific example of the application.
  • Multi-vendor rAPPs are managed on the Non-Real Time RIC platform.
  • Non-Real Time RIC for example, utilizes AI and ML (Machine Learning) to perform autonomous control processing, optimization processing, or parameter value determination processing for wireless devices that make up the RAN. Execute.
  • AI and ML Machine Learning
  • a wireless device may be, for example, a base station.
  • the base station may be an eNB (evolved Node B) that supports LTE (Long Term Evolution), and supports the so-called 5G (Fifth Generation) wireless communication system specified in 3GPP (3rd Generation Partnership Project). It may be a base station.
  • the wireless device may be an RU (Radio Unit), a DU (Distributed Unit), or a CU (Central Unit) defined by the O-RAN Alliance.
  • RU, DU, and CU may be referred to as O(O-RAN)-RU, O-DU, and O-CU, respectively.
  • Signals are transmitted and received between the O-RU, O-DU, and O-CU and the Non-Real Time RIC and Near-Real Time RIC according to the interface specified by the O-RAN Alliance.
  • the management device 20 is configured by adding a generation unit 21 , a conversion unit 22 , and a display unit 23 to the management device 10 .
  • functions different from those of the management apparatus 10 and detailed descriptions of various functions described in the management apparatus 10 will be mainly described.
  • the management device 20 manages multiple AI engines 25. Each of the multiple AI engines 25 may be managed in a container. The AI engine 25 managed in the container can operate independently for each container.
  • a container may be, for example, a memory area provided within the management device 20 .
  • a plurality of AI engines 25 are AI engines that execute analysis processing for a common purpose.
  • Each of the plurality of AI engines 25 may be, for example, an AI engine produced by a different vendor, or an AI engine with a different software version.
  • the management device 20 is provided with a new AI engine that executes analysis processing for a common purpose with the existing AI engine, and when the new AI engine has desired analysis accuracy, to manage or control the wireless device 30 .
  • the new AI engine does not have the desired analysis accuracy, it manages or controls the wireless device 30 based on the analysis results of the existing AI engine.
  • the generation unit 21 generates analysis target data used for analysis processing in the AI engine 25.
  • the generation unit 21 generates analysis target data using monitoring data collected from wireless devices and various devices on a communication network.
  • Monitoring data may include, for example, sensor data when various devices on the communication network are sensor devices.
  • the monitoring data may be data generated by various devices, and may be referred to as management data in various devices.
  • Management data may include, for example, processor utilization in the device, data throughput, data transfer rate, and the like.
  • Management data may also be called maintenance data.
  • the communication network may be a mobile communication network connecting the radio access device and the core network device, etc., or may be a maintenance network to which the maintenance device is connected.
  • the generation unit 21 may generate analysis target data by statistically processing the monitoring data.
  • Statistical processing may be, for example, calculation of values such as an average value, median value, minimum value, or maximum value of a plurality of pieces of monitoring data, or generation of data combining a plurality of pieces of monitoring data.
  • Analysis target data may be referred to as processed data obtained by processing monitoring data.
  • the selection unit 12 may select at least one device from the wireless device that generates the monitoring data and various devices on the communication network. For example, the selection unit 12 may generate an acquisition instruction signal including identification information of the selected device and output the acquisition signal to the generation unit 21 .
  • the generation unit 21 may collect monitoring data from a device specified by identification information included in the acquisition instruction signal. The identification information of the device included in the acquisition instruction signal may be input to the selection unit 12 via the input interface by an administrator who operates the management device 20 .
  • the selection unit 12 selects two or more AI engines 25 that the control unit 11 operates in parallel from among the plurality of AI engines 25 .
  • the control unit 11 outputs the analysis target data generated by the generation unit 21 to the AI engine 25 selected by the selection unit 12 .
  • the selection unit 12 may generate an output destination instruction signal including identification information of the AI engine 25 that is the output destination of the analysis target data, and output the output destination instruction signal to the control unit 11 .
  • the control unit 11 may output the analysis target data to the AI engine 25 specified by the identification information included in the output destination instruction signal.
  • the AI engine identification information included in the output destination instruction signal may be input to the selection unit 12 via the input interface by a manager who operates the management device 20 .
  • the conversion unit 22 converts the analysis target data generated by the generation unit 21 into analysis target data in a form that can be analyzed by the AI engine 25 selected by the selection unit 12 .
  • a plurality of AI engines 25 with different software versions may differ in the data format or data structure of data used for analysis processing.
  • a plurality of AI engines 25 generated by different vendors may have different data formats or data structures of data used for analysis processing.
  • the conversion unit 22 converts the analysis target data generated by the generation unit 21 into a data format, data structure, etc. that can be analyzed by the AI engine 25 selected by the selection unit 12 .
  • the conversion unit 22 converts the analysis target data generated by the generation unit 21 into a data format, data structure, etc. that can be read by the AI engine 25 selected by the selection unit 12 .
  • the control unit 11 outputs the analysis target data converted by the conversion unit 22 to the AI engine 25 selected by the selection unit 12 .
  • the control unit 11 outputs the analysis results of the multiple AI engines 25 operating in parallel to the display unit 23 .
  • the display unit 23 may be, for example, a display device.
  • the display unit 23 may be a display device used integrally with the management device 20, or may be a display device connected to the management device 20 via a cable or network.
  • the display unit 23 displays the received analysis results. Specifically, the display unit 23 displays analysis results for each AI engine 25 .
  • the administrator of the management device 20 confirms the analysis result displayed on the display unit 23 and selects the AI engine 25 to be used for controlling the wireless device 30 . For example, the administrator inputs the selection result to the management device 20 via the input interface of the management device 20 . The administrator selects, for example, the AI engine 25 that outputs an analysis result close to the desired analysis result.
  • the selection unit 12 receives the selection result input by the administrator as a selection instruction signal.
  • the selection instruction signal includes identification information of the AI engine 25 selected by the administrator.
  • the communication unit 13 transmits the analysis result of the AI engine 25 selected by the selection unit 12 to the wireless device 30 .
  • the wireless device 30 uses the analysis results received from the management device 20, the wireless device 30 performs, for example, resource optimization, automatic control, and the like.
  • FIG. 5 shows that data handled in the management device 20 is transmitted to the radio device 30, which is the device to be controlled, through each process of data acquisition, data processing, AI/ML, and action execution. .
  • the generation unit 21 collects monitoring data from wireless devices and various devices on the communication network.
  • the generation unit 21 may collect monitoring data from wireless devices selected by the selection unit 12 or from various devices on the communication network.
  • the generation unit 21 generates analysis target data using collected monitoring data.
  • FIG. 5 shows that the number of AI engines 25 executing analysis processing is two, the number of AI engines 25 executing analysis processing is not limited to two. Also, for example, the two AI engines 25 that perform analysis processing may have different software versions.
  • FIG. 5 shows that the AI engine 25 of software version 1 (V1) and the AI engine 25 of software version 2 (V2) execute analysis processing.
  • software version 2 may be a newer version than software version 1.
  • a test may be conducted in a commercial environment to confirm the accuracy of the AI engine 25 analysis process for the newly provided software version 2.
  • the selection unit 12 may select the AI engine 25 that outputs analysis target data. That is, the selection unit 12 may select the AI engine 25 that executes analysis processing.
  • the communication unit 13 transmits to the wireless device 30 one of the analysis results of the two AI engines 25 using different software versions.
  • the communication unit 13 transmits the analysis result of the AI engine 25 selected by the selection unit 12 from among the two AI engines 25 to the wireless device 30 .
  • the selection unit 12 receives a selection result input by the administrator of the management device 20 and selects one of the two AI engines 25 .
  • the selection unit 12 first determines a data flow from data acquisition to action execution, and components of the management device 20 such as the generation unit 21 may execute processing according to the data flow determined by the selection unit 12. good.
  • the administrator confirms the analysis results of the two AI engines 25 displayed on the display unit 23, and whether or not the analysis results of the newly provided AI engine 25 operated by software version 2 show the desired results. Check whether If the newly provided analysis result in the AI engine 25 is not the desired result, the administrator decides to use the analysis result in the AI engine 25 operating with software version 1 that has already been operating in the commercial environment. may In this case, as shown in FIG. 5, the communication unit 13 transmits the analysis result of the AI engine 25 operating with the software version 1 to the wireless device 30 . The administrator may decide to utilize the analysis results in the AI engine 25 running with software version 2 if the newly provided analysis results in the AI engine 25 are the desired results.
  • the administrator of the management device 20 can change the data flow of FIG. 5 as appropriate. For example, the administrator may select monitoring data necessary to generate data to be analyzed, select an AI engine for executing processes related to AI/ML, and further select analysis results.
  • the selection unit 12 executes selection processing according to selection information input by the administrator via the input interface of the management device 20 .
  • the display unit 23 may display the data flow shown in FIG. 5 and the administrator may change the data flow as appropriate by selecting the data flow arrows displayed on the display unit 23 .
  • the generation unit 21 collects monitoring data from wireless devices selected by the selection unit 12 or various devices on the communication network (S21).
  • the monitoring data may include sensor data, data indicating the state of the wireless device, and the like.
  • the generation unit 21 generates analysis target data using the collected monitoring data (S22).
  • the selection unit 12 selects the AI engine 25 as the output destination of the analysis data (S23).
  • the conversion unit 22 converts the analysis target data so that it can be used by the AI engine 25 selected by the selection unit 12 (S24).
  • the conversion unit 22 may convert the data format or data structure of the analysis target data.
  • a data format may, for example, be referred to as a data format.
  • the data structure may be defined by, for example, an array of data.
  • the conversion unit 22 may manage each AI engine 25 and the data format or data structure in each AI engine 25 in association with each other in advance.
  • the control unit 11 may display analysis results, which are results of analysis processing in the plurality of AI engines 25 , on the display unit 23 .
  • the selection unit 12 selects the AI engine 25 (S26).
  • the selection unit 12 selects the AI engine 25 based on the selection result of the AI engine 25 input by the administrator or the like.
  • the communication unit 13 transmits the analysis result of the AI engine 25 selected by the selection unit 12 to the wireless device 30 (S27).
  • the management device 20 operates the AI engine 25 in parallel. Furthermore, the management device 20 selects an AI engine 25 selected by the administrator who has confirmed the analysis results of the plurality of AI engines 25, and transmits the analysis results generated by the selected AI engine 25 to the wireless device 30. In this way, a plurality of AI engines 25 are simultaneously operated in the commercial environment, and when the new AI engine 25 outputs desired analysis results, the AI engine 25 operating in the commercial environment is changed to the new AI engine 25. be able to.
  • the new AI engine 25 may be a new software version of the AI engine or an AI engine provided by a vendor different from the vendor that produced the currently running AI engine 25 .
  • the conversion unit 22 converts the data format or data structure of the data to be analyzed so that each AI engine 25 can use it.
  • a plurality of AI engines 25 can use common analysis target data to execute analysis processing.
  • FIG. 7 is a block diagram showing a configuration example of the management device 10 and the management device 20 (hereinafter referred to as the management device 10 and the like).
  • the management device 10 and the like include a network interface 1201 , a processor 1202 and a memory 1203 .
  • Network interface 1201 may be used to communicate with network nodes.
  • Network interface 1201 may include, for example, an IEEE 802.3 series compliant network interface card (NIC). IEEE stands for Institute of Electrical and Electronics Engineers.
  • NIC network interface card
  • the processor 1202 reads and executes software (computer program) from the memory 1203 to perform the processing of the management device 10 and the like described using the flowcharts in the above embodiments.
  • Processor 1202 may be, for example, a microprocessor, MPU, or CPU.
  • Processor 1202 may include multiple processors.
  • the memory 1203 is composed of a combination of volatile memory and non-volatile memory.
  • Memory 1203 may include storage remotely located from processor 1202 .
  • the processor 1202 may access the memory 1203 via an I/O (Input/Output) interface (not shown).
  • I/O Input/Output
  • memory 1203 is used to store software modules.
  • the processor 1202 reads and executes these software modules from the memory 1203, thereby performing the processing of the management apparatus 10 and the like described in the above embodiments.
  • each of the processors included in the management device 10 and the like in the above-described embodiments includes one or more programs containing instructions for causing a computer to execute the algorithm described with reference to the drawings. to run.
  • the program includes instructions (or software code) that, when read into a computer, cause the computer to perform one or more of the functions described in the embodiments.
  • the program may be stored in a non-transitory computer-readable medium or tangible storage medium.
  • computer readable media or tangible storage media may include random-access memory (RAM), read-only memory (ROM), flash memory, solid-state drives (SSD) or other memory technology, CDs - ROM, digital versatile disc (DVD), Blu-ray disc or other optical disc storage, magnetic cassette, magnetic tape, magnetic disc storage or other magnetic storage device.
  • the program may be transmitted on a transitory computer-readable medium or communication medium.
  • transitory computer readable media or communication media include electrical, optical, acoustic, or other forms of propagated signals.
  • (Appendix 1) a control unit that operates in parallel a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine; a selection unit that selects either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine; The analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis of the second AI engine. and a communication unit that outputs the result to a controlled device that uses the result.
  • (Appendix 2) The management device according to appendix 1, wherein the first AI engine and the second AI engine have different software versions.
  • (Appendix 3) The control unit outputting the analysis result of the first AI engine and the analysis result of the second AI engine to a display device;
  • the selection unit Either the first AI engine or the second AI engine is selected based on the selection instruction signal input by the user who visually recognizes the analysis result of the first AI engine and the analysis result of the second AI engine.
  • the management device according to appendix 1 or 2 which selects: (Appendix 4) 4.
  • the management according to any one of Appendices 4 to 6, further comprising a conversion unit that converts the analysis target data into data in a format that can be used by each of the first AI engine and the second AI engine.
  • Device (Appendix 8) A control unit that operates in parallel with a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine, and the first AI engine or the second AI a selection unit that selects either the first AI engine or the second AI engine based on a selection instruction signal containing engine identification information; a management device comprising a communication unit that transmits analysis results of a selected AI engine among the analysis results of the AI engines; and a control target device that executes autonomous control using the analysis result.
  • a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine are operated in parallel, selecting either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine;
  • the analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis of the second AI engine.
  • a control method that outputs results to controlled devices that use them.
  • a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine are operated in parallel, selecting either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine;
  • the analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis of the second AI engine.
  • a non-transitory computer-readable medium that stores a program that causes a computer to output results to a controlled device that utilizes the results.

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

The purpose of the present invention is to provide a management device (10) capable of applying AI engines to a commercial environment system, which uses data different from that of a test environment, so as to enable normal operation of the commercial environment system. The management device (10) comprises: a control unit (11) that operates in parallel a first AI engine and a second AI engine that executes analysis processing having a shared purpose with the first AI engine; a selection unit (12) that selects either one of the first AI engine and the second AI engine on the basis of a selection command signal including first AI engine identification information or second AI engine identification information; and a communication unit (13) that, from among the result of analysis by the first AI engine and the result of analysis by the second AI engine, outputs the result of analysis carried out by the selected AI engine to a controlled device that uses the result of analysis by the first AI engine or the result of analysis by the second AI engine.

Description

管理装置、通信システム、制御方法、及び非一時的なコンピュータ可読媒体Management device, communication system, control method, and non-transitory computer readable medium
　本開示は管理装置、通信システム、制御方法、及び非一時的なコンピュータ可読媒体に関する。 The present disclosure relates to management devices, communication systems, control methods, and non-transitory computer-readable media.
　近年、移動通信ネットワークにおける無線アクセスネットワーク（Radio Access Network, RAN）において、AI（Artificial Intelligence）を利用した高度な制御を実現することが検討されている。高度な制御として、例えば、RANにおけるリソース最適化を実現するためにAIを利用することが検討されている。 In recent years, the realization of advanced control using AI (Artificial Intelligence) has been considered in radio access networks (RAN) in mobile communication networks. As advanced control, for example, the use of AI to achieve resource optimization in RAN is under study.
　引用文献１には、クラウド側のサーバが、工場等に設置された学習対象機器の温度、振動、スピード等の情報を収集し、AIを利用して、学習対象機器を自動制御するシステムが開示されている。 Cited Document 1 discloses a system in which a cloud-side server collects information such as temperature, vibration, speed, etc. of learning target devices installed in factories, etc., and uses AI to automatically control the learning target devices. It is
特開２０１９－１３９７３４号公報JP 2019-139734 A
　ここで、自動制御内容が多岐にわたる場合、制御内容ごとに異なるアルゴリズムが適用される。その結果、制御内容ごとに異なるAIエンジンが用いられる。また、AIエンジンは、ソフトウェアとして構成されており、アルゴリズムの更新等に応じて、ソフトウェアバージョンが更新される。もしくは、現在稼働中のAIエンジンを、他のベンダのAIエンジンへ変更することもある。商用環境において稼働しているAIエンジンと異なるソフトウェアバージョンのAIエンジン、もしくは、異なるベンダのAIエンジンを、商用環境のシステムへ適用した場合、試験環境とは異なるデータを用いた分析処理を行うことになる。その結果、試験環境において検証後に、商用環境に適用されたAIエンジンは、想定よりも低い精度の分析結果を出力する場合があり、商用環境のシステムが正常に動作しなくなる可能性があるという問題がある。 Here, when the content of automatic control is diverse, different algorithms are applied for each content of control. As a result, a different AI engine is used for each control content. Also, the AI engine is configured as software, and the software version is updated according to algorithm updates and the like. Alternatively, the AI engine currently in operation may be changed to another vendor's AI engine. If an AI engine with a software version different from the AI engine running in the commercial environment, or an AI engine from a different vendor, is applied to the system in the commercial environment, analysis processing using data different from that in the test environment will be performed. Become. As a result, after verification in the test environment, the AI engine applied to the commercial environment may output analysis results with lower accuracy than expected, and the system in the commercial environment may not operate normally. There is
　本開示の目的の一つは、試験環境とは異なるデータを用いる商用環境のシステムを正常に稼働させるように、AIエンジンを商用環境のシステムへ適用することができる管理装置、通信システム、制御方法、及び非一時的なコンピュータ可読媒体を提供することにある。 One of the objects of the present disclosure is a management device, a communication system, and a control method that can apply an AI engine to a system in a commercial environment so that the system in the commercial environment that uses data different from the test environment can be operated normally. and to provide a non-transitory computer-readable medium.
　本開示の第１の態様にかかる管理装置は、第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働する制御部と、前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択する選択部と、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果を利用する制御対象装置へ出力する通信部と、を備える。 A management device according to a first aspect of the present disclosure includes a control unit that operates in parallel a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine. a selection unit that selects either the first AI engine or the second AI engine based on a selection instruction signal containing identification information of the first AI engine or the second AI engine; The analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis result of the second AI engine. and a communication unit that outputs to the controlled device that uses the.
　本開示の第２の態様にかかる通信システムは、第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働する制御部と、前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択する選択部と、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を送信する通信部と、を備える管理装置と、前記分析結果を利用した自律制御を実行する制御対象装置と、を備える。 A communication system according to a second aspect of the present disclosure includes: a first AI engine; a selection unit that selects either the first AI engine or the second AI engine based on a selection instruction signal containing identification information of the first AI engine or the second AI engine; A management device comprising: a communication unit that transmits the analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine; and autonomous control using the analysis result. and a controlled device that executes
　本開示の第３の態様にかかる制御方法は、第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働し、前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択し、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果を利用する制御対象装置へ出力する。 A control method according to a third aspect of the present disclosure operates in parallel with a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine, one of the first AI engine and the second AI engine is selected based on a selection instruction signal including identification information of one AI engine or the second AI engine; A controlled device that uses the analysis result of the selected AI engine from the analysis result or the analysis result of the second AI engine, and the analysis result of the first AI engine or the analysis result of the second AI engine. Output to
　本開示の第４の態様にかかるプログラムは、第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働し、前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択し、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果を利用する制御対象装置へ出力することをコンピュータに実行させる。 A program according to a fourth aspect of the present disclosure operates in parallel with a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine, selecting either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the AI engine or the second AI engine, and analyzing the first AI engine The analysis result of the AI engine selected from the result or the analysis result of the second AI engine is sent to the controlled device that uses the analysis result of the first AI engine or the analysis result of the second AI engine. Make the computer do the output.
　本開示により、試験環境とは異なるデータを用いる商用環境のシステムを正常に稼働させるように、AIエンジンを商用環境のシステムへ適用することができる管理装置、通信システム、制御方法、及び非一時的なコンピュータ可読媒体を提供することができる。 According to the present disclosure, a management device, a communication system, a control method, and a non-temporary system that can apply an AI engine to a system in a commercial environment so as to normally operate the system in the commercial environment using data different from the test environment A computer readable medium can be provided.
実施の形態１にかかる管理装置の構成図である。1 is a configuration diagram of a management device according to Embodiment 1; FIG. 実施の形態１にかかる管理装置において実行される制御処理の流れを示す図である。4 is a diagram showing the flow of control processing executed by the management device according to the first embodiment; FIG. 実施の形態２にかかる通信システムの構成図である。1 is a configuration diagram of a communication system according to a second embodiment; FIG. 実施の形態２にかかる管理装置の構成図である。2 is a configuration diagram of a management device according to a second embodiment; FIG. 実施の形態２にかかる管理装置におけるデータフローを示す図である。FIG. 10 is a diagram showing a data flow in the management device according to the second embodiment; FIG. 実施の形態２にかかる分析結果の送信処理の流れを示す図である。FIG. 10 is a diagram showing the flow of analysis result transmission processing according to the second embodiment; それぞれの実施の形態にかかる管理装置の構成図である。4 is a configuration diagram of a management device according to each embodiment; FIG.
　（実施の形態１）
　以下、図面を参照して本発明の実施の形態について説明する。図１を用いて実施の形態１にかかる管理装置１０の構成例について説明する。管理装置１０は、プロセッサがメモリに格納されたプログラムを実行することによって動作するコンピュータ装置であってもよい。管理装置１０は、例えば、サーバ装置であってもよい。もしくは、管理装置１０は、複数のコンピュータ装置が協働して動作するプラットフォームと同様の動作もしくは処理を実行してもよい。 (Embodiment 1)
BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described with reference to the drawings. A configuration example of the management apparatus 10 according to the first embodiment will be described with reference to FIG. The management device 10 may be a computer device operated by a processor executing a program stored in memory. The management device 10 may be, for example, a server device. Alternatively, the management device 10 may perform operations or processes similar to those of a platform on which a plurality of computer devices operate cooperatively.
　管理装置１０は、制御部１１、選択部１２、及び通信部１３を有している。制御部１１、選択部１２、及び通信部１３は、プロセッサがメモリに格納されたプログラムを実行することによって処理が実行されるソフトウェアもしくはモジュールであってもよい。または、制御部１１、選択部１２、及び通信部１３は、回路もしくはチップ等のハードウェアであってもよい。 The management device 10 has a control unit 11, a selection unit 12, and a communication unit 13. The control unit 11, the selection unit 12, and the communication unit 13 may be software or modules whose processing is executed by a processor executing a program stored in memory. Alternatively, the control unit 11, the selection unit 12, and the communication unit 13 may be hardware such as circuits or chips.
　制御部１１は、第１のAIエンジンと、第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働する。分析処理は、時間の経過とトラヒック量とを関連付けたトラヒックの推移を示すデータを生成する処理であってもよい。分析処理とは、例えば、移動通信システムの基地局においてリソースの最適化を自律的もしくは自動的に実行するための処理であってもよい。具体的には、分析処理は、トラヒックの少ない時間帯に動作させる基地局の数を減らし、トラヒックの多い時間帯に動作させる基地局の数を増加させる、との制御を実現するためのトラヒック分析処理であってもよい。　 The control unit 11 operates in parallel the first AI engine and the second AI engine that executes analysis processing for a common purpose with the first AI engine. The analysis process may be a process of generating data indicating traffic transitions in which the passage of time is associated with the amount of traffic. The analysis process may be, for example, a process for autonomously or automatically optimizing resources in a base station of a mobile communication system. Specifically, the analysis processing reduces the number of base stations that are operated during periods of low traffic and increases the number of base stations that are operated during periods of heavy traffic. It may be processing.　
　もしくは、分析処理は、より具体的には、基地局のアンテナ単位に、トラヒックの推移を示すデータを生成する処理であってもよい。 Alternatively, the analysis process may be, more specifically, a process of generating data indicating traffic transitions for each antenna of the base station.
　もしくは、分析処理は、複数のアンテナを有する基地局における、ビームフォーミングの最適化を実現するための、端末装置の位置情報の分析処理であってもよい。より具体的には、分析処理は、端末装置の移動傾向を示すデータを生成する処理であってもよい。 Alternatively, the analysis processing may be analysis processing of location information of the terminal device for achieving optimization of beamforming in a base station having multiple antennas. More specifically, the analysis process may be a process of generating data indicating movement tendencies of terminal devices.
　また、分析処理は、分析結果に基づいて実行される、最適なパラメータ値の決定処理もしくは最適な制御内容の決定処理を含んでもよい。 The analysis process may also include an optimum parameter value determination process or an optimum control content determination process, which is executed based on the analysis result.
　また、分析処理は、複数のデータを収集し、収集したデータを統計的に処理することであってもよい。また、分析処理は、同一種類のデータのみではなく、異なる種類のデータが組み合わされたデータを用いて実行されてもよい。データの種類は、例えば、時間帯ごとの送受信されるデータ量の推移を示すデータ、気候データ、コンサート等のイベント日時に関する情報等であってもよい。 Also, the analysis processing may be to collect multiple data and statistically process the collected data. Also, the analysis processing may be performed using not only data of the same type but also data in which different types of data are combined. The type of data may be, for example, data indicating changes in the amount of data transmitted and received for each time period, climate data, information regarding the date and time of an event such as a concert, and the like.
　AIエンジンは、提供するサービスもしくは分析内容によって異なるアルゴリズムを有するソフトウェアもしくはアプリケーションであってもよい。AIエンジンは、新たな機能の追加もしくは既存機能の削除、もしくは、既存機能の更新等によって、ソフトウェアバージョンが更新されてもよい。 The AI engine may be software or an application with different algorithms depending on the service provided or the analysis content. The software version of the AI engine may be updated by adding new functions, deleting existing functions, or updating existing functions.
　共通の目的の分析処理を実行する第１のAIエンジン及び第２のAIエンジンは、例えば、ソフトウェアバージョンの異なるAIエンジンであってもよい。もしくは、第１のAIエンジン及び第２のAIエンジンは、それぞれが異なるベンダによって生成されたAIエンジンであってもよい。 The first AI engine and the second AI engine that execute analysis processing for a common purpose may be, for example, AI engines with different software versions. Alternatively, the first AI engine and the second AI engine may be AI engines produced by different vendors.
　並行稼働とは、例えば、第１のAIエンジンと第２のAIエンジンとが、実質的に同一のタイミングに分析処理を実行することであってもよい。例えば、並行稼働とは、第１のAIエンジンの稼働時刻と、第２のAIエンジンの稼働時刻との一部が重複することであってもよい。もしくは、並行稼働とは、第１のAIエンジンの分析処理が行われた後に、第２のAIエンジンの分析処理を行うことであってもよい。つまり、並行稼働とは、それぞれのAIエンジンにおける分析処理が実行された結果、複数の分析結果が生成されることであってもよい。並行稼働は、並行運用と言い換えられてもよい。 "Parallel operation" may mean, for example, that the first AI engine and the second AI engine execute analysis processing at substantially the same timing. For example, parallel operation may be a partial overlap between the operating time of the first AI engine and the operating time of the second AI engine. Alternatively, the parallel operation may be performing the analysis processing of the second AI engine after the analysis processing of the first AI engine is performed. In other words, parallel operation may mean that a plurality of analysis results are generated as a result of executing analysis processing in each AI engine. Parallel operation may be rephrased as parallel operation.
　制御部１１は、例えば、第１のAIエンジンと第２のAIエンジンとを稼働するタイミングを制御してもよく、それぞれのAIエンジンにおける分析処理の継続時間を制御してもよい。さらに、制御部１１は、第１のAIエンジン及び第２のAIエンジンのそれぞれに、分析処理に用いられるデータを出力してもよい。 The control unit 11 may, for example, control the timing of operating the first AI engine and the second AI engine, and may control the duration of analysis processing in each AI engine. Further, the control unit 11 may output data used for analysis processing to each of the first AI engine and the second AI engine.
　選択部１２は、第１のAIエンジンもしくは第２のAIエンジンの識別情報を含む選択指示信号に基づいて、第１のAIエンジン及び第２のAIエンジンのいずれかを選択する。 The selection unit 12 selects either the first AI engine or the second AI engine based on the selection instruction signal containing the identification information of the first AI engine or the second AI engine.
　識別情報は、AIエンジンを一意に識別する情報であり、例えば、AIエンジンに付与された名称を示す情報であってもよい。また、識別情報にはソフトウェアバージョンを示す情報が含まれてもよい。識別情報は、例えば、管理装置１０の管理者が、それぞれのAIエンジンと識別情報との組み合わせを管理してもよく、管理装置１０内のプロセッサ等が、それぞれのAIエンジンに付与する識別情報を自律的に生成してもよい。 The identification information is information that uniquely identifies the AI engine, and may be, for example, information indicating the name given to the AI engine. Also, the identification information may include information indicating the software version. For the identification information, for example, the administrator of the management device 10 may manage the combination of each AI engine and the identification information, and the processor or the like in the management device 10 may manage the identification information given to each AI engine. It may be generated autonomously.
　選択指示信号は、AIエンジンの識別情報を含み、選択部１２は、選択指示信号に含まれる識別情報に基づいて、第１のAIエンジン及び第２のAIエンジンのいずれかを選択する。選択指示信号は、例えば、管理装置１０を管理する管理者等が入力した第１のAIエンジン及び第２のAIエンジンのいずれかの識別情報を含んでもよい。つまり、選択指示信号は、管理者によって識別情報が入力されたことを契機として生成されてもよい。言い換えると、選択指示信号は、管理者によって入力インタフェースを介して入力される入力信号であってもよい。入力インタフェースは、例えば、キーボード、タッチパネル等であってもよく、マイク等の音声入力インタフェースであってもよい。 The selection instruction signal includes identification information of the AI engine, and the selection unit 12 selects either the first AI engine or the second AI engine based on the identification information included in the selection instruction signal. The selection instruction signal may include, for example, identification information of either the first AI engine or the second AI engine input by a manager or the like who manages the management device 10 . In other words, the selection instruction signal may be generated when the administrator inputs the identification information. In other words, the selection instruction signal may be an input signal input by the administrator through the input interface. The input interface may be, for example, a keyboard, a touch panel, or the like, or may be a voice input interface such as a microphone.
　通信部１３は、第１のAIエンジンの分析結果もしくは第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を、第１のAIエンジンの分析結果もしくは第２のAIエンジンの分析結果を利用する制御対象装置へ出力する。 The communication unit 13 sends the analysis result of the selected AI engine from the analysis result of the first AI engine or the analysis result of the second AI engine to the analysis result of the first AI engine or the analysis result of the second AI engine. Output the analysis results to the control target device that uses them.
　制御対象装置は、ソフトウェアを用いて制御可能な装置であり、例えば、リソースの最適化等の自律制御を実行する基地局であってもよく、移動通信システムを構成する他の装置であってもよい。分析結果を受け取った制御対象装置は、分析結果に基づいたリソースの最適化等の自律制御を行ってもよい。 A device to be controlled is a device that can be controlled using software. For example, it may be a base station that executes autonomous control such as resource optimization, or it may be another device that constitutes a mobile communication system. good. The controlled device that receives the analysis result may perform autonomous control such as resource optimization based on the analysis result.
　通信部１３は、例えば、選択されなかったAIエンジンから出力された分析結果を、制御対象装置へ出力することなく、廃棄もしくは破棄してもよい。もしくは、選択されなかったAIエンジンから出力された分析結果は、管理装置１０内のメモリに、格納されてもよい。管理装置１０内に格納された分析結果は、例えば、検証を実行した際の履歴情報として管理されてもよい。 For example, the communication unit 13 may discard or discard the analysis results output from the unselected AI engines without outputting them to the control target device. Alternatively, the analysis results output from the AI engines that have not been selected may be stored in the memory within the management device 10 . The analysis results stored in the management device 10 may be managed, for example, as history information when verification is executed.
　続いて、図２を用いて実施の形態１にかかる管理装置において実行される制御処理の流れについて説明する。はじめに、選択部１２は、第１のAIエンジンと、第１のAIエンジンと共通の目的の分析処理を実行し、第１のAIエンジンと並行稼働する第２のAIエンジンとのいずれかを選択する（Ｓ１１）。選択部１２は、第１のAIエンジンもしくは第２のAIエンジンの識別情報を含む選択指示信号に基づいて、第１のAIエンジン及び第２のAIエンジンのいずれかを選択する。 Next, the flow of control processing executed by the management device according to the first embodiment will be described using FIG. First, the selection unit 12 selects either a first AI engine or a second AI engine that executes analysis processing for a common purpose with the first AI engine and operates in parallel with the first AI engine. (S11). The selection unit 12 selects either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine.
　次に、通信部１３は、第１のAIエンジンの分析結果もしくは第２のAIエンジンの分析結果の内、選択部１２において選択されたAIエンジンの分析結果を制御対象装置へ出力する（Ｓ１２）。制御対象装置は、第１のAIエンジンの分析結果もしくは第２のAIエンジンの分析結果を利用して、自律制御等を実行する。 Next, the communication unit 13 outputs the analysis result of the AI engine selected by the selection unit 12 out of the analysis result of the first AI engine or the analysis result of the second AI engine to the device to be controlled (S12). . The controlled device executes autonomous control and the like using the analysis result of the first AI engine or the analysis result of the second AI engine.
　以上説明したように、管理装置１０は、共通の目的の分析処理を実行する複数のAIエンジンにおける分析結果の中から、少なくとも１つの分析結果を選択し、選択した分析結果を制御対象装置へ出力する。 As described above, the management device 10 selects at least one analysis result from among the analysis results of a plurality of AI engines that execute analysis processing for a common purpose, and outputs the selected analysis result to the controlled device. do.
　例えば、第２のAIエンジンが、既存の第１のAIエンジンのソフトウェアバージョンを更新したAIエンジンである場合、管理装置１０は、ソフトウェアバージョンの異なる二つのAIエンジンを商用環境において並行稼働することができる。管理装置１０は、例えば、管理者等が、第２のAIエンジンの分析結果のほうが、第１のAIエンジンの分析結果よりも所望の結果に近いと判断した場合、第２のAIエンジンの分析結果を、制御対象装置における自律制御等に適用することができる。一方、管理装置１０は、管理者等が、第１のAIエンジンの分析結果のほうが、第２のAIエンジンの分析結果よりも所望の結果に近いと判断した場合、第１のAIエンジンの分析結果を、制御対象装置における自律制御等に適用することができる。このようにして、管理装置１０は、商用環境において異なるバージョンのAIエンジンを並行稼働し、新たなソフトウェアバージョンのAIエンジンの分析結果が所望の結果を満たさないうちは、古いソフトウェアバージョンのAIエンジンを適用し続ける。なお、第１のAIエンジンと第２のAIエンジンは、異なるソフトウェアバージョンの場合だけではなく、例えば、異なるベンダによって生成されたAIエンジンであってもよい。 For example, if the second AI engine is an AI engine obtained by updating the software version of the existing first AI engine, the management device 10 can operate two AI engines with different software versions in parallel in a commercial environment. can. For example, when the administrator or the like determines that the analysis result of the second AI engine is closer to the desired result than the analysis result of the first AI engine, the management device 10 performs the analysis of the second AI engine. The results can be applied to autonomous control or the like in the controlled device. On the other hand, when the administrator or the like determines that the analysis result of the first AI engine is closer to the desired result than the analysis result of the second AI engine, the management device 10 performs the analysis of the first AI engine. The results can be applied to autonomous control or the like in the controlled device. In this way, the management device 10 operates AI engines of different versions in parallel in a commercial environment, and uses the AI engine of the old software version until the analysis result of the AI engine of the new software version does not satisfy the desired result. keep applying. Note that the first AI engine and the second AI engine are not limited to different software versions, and may be AI engines generated by different vendors, for example.
　（実施の形態２）
　続いて、図３を用いて実施の形態２にかかる通信システムの構成例について説明する。図３の通信システムは、管理装置２０及び無線装置３０を有している。無線装置３０は、制御対象装置に含まれる装置である。図３の通信システムにおいては、管理装置２０が無線装置３０を管理する。管理装置２０が無線装置３０を管理するとは、例えば、管理装置２０が、無線装置３０の動作を制御すること、無線装置３０において設定される各種パラメータの値を決定することを含んでもよい。もしくは、管理装置２０は、無線装置３０が実行する機能を最適化させるために必要な情報を生成し、生成した情報を無線装置３０へ送信してもよい。管理装置２０と無線装置３０とは、例えば、IPネットワークを介して通信を行ってもよい。 (Embodiment 2)
Next, a configuration example of the communication system according to the second embodiment will be described with reference to FIG. The communication system of FIG. 3 has a management device 20 and a radio device 30. The wireless device 30 is a device included in the controlled device. In the communication system of FIG. 3 , the management device 20 manages the wireless device 30 . The management device 20 managing the wireless device 30 may include, for example, the management device 20 controlling the operation of the wireless device 30 and determining the values of various parameters set in the wireless device 30 . Alternatively, management device 20 may generate information necessary for optimizing functions executed by wireless device 30 and transmit the generated information to wireless device 30 . The management device 20 and the wireless device 30 may communicate via an IP network, for example.
　また、無線装置３０は、実際に移動通信装置に対して無線サービスを提供しており、管理装置２０及び無線装置３０が構成する通信システムは、商用環境において動作しているシステムといえる。 Also, the wireless device 30 actually provides wireless services to mobile communication devices, and the communication system configured by the management device 20 and the wireless device 30 can be said to be a system operating in a commercial environment.
　管理装置２０は、例えば、O-RAN（Open Radio Access Network）アライアンスにおいて規定されているSMO（Service Management and Orchestration ）を構成するRIC（RAN Intelligent Controller）であってもよい。RICは、具体的には、制御周期の異なるNon-Real Time RIC及びNear-Real Time RICを含んでもよい。 The management device 20 may be, for example, an RIC (RAN Intelligent Controller) that configures SMO (Service Management and Orchestration) defined in the O-RAN (Open Radio Access Network) Alliance. Specifically, the RIC may include Non-Real Time RIC and Near-Real Time RIC with different control cycles.
　RICは、AIエンジンの管理及びデータ処理を実行する。RICは、AIエンジンに対して、データを提供し、AIエンジンにおける分析結果に基づいて、無線装置を制御する。AIエンジンは、サービス内容や、分析処理毎に異なるアルゴリズムが用いられるアプリケーションであってもよい。アプリケーションは、O-RANアライアンスにおいて規定されているrAPP（Non-Real Time RIC Application）であってもよい。つまり、AIエンジンは、アプリケーションと言い換えられてもよく、アプリケーションの具体例として、rAPPが用いられてもよい。Non-Real Time RICのプラットフォームにおいて、マルチベンダのrAPPが管理される。 RIC manages the AI engine and performs data processing. The RIC provides data to the AI engine and controls the wireless device based on the analysis results of the AI engine. The AI engine may be an application that uses a different algorithm for each service content or analysis process. The application may be rAPP (Non-Real Time RIC Application) defined in the O-RAN Alliance. In other words, the AI engine may be rephrased as an application, and rAPP may be used as a specific example of the application. Multi-vendor rAPPs are managed on the Non-Real Time RIC platform.
　Non-Real Time RICは、例えば、AI及びML（Machine Learning）を活用して、RANを構成する無線装置の自律制御処理、最適化処理、もしくは無線装置において設定されるパラメータ値の決定処理等を実行する。 Non-Real Time RIC, for example, utilizes AI and ML (Machine Learning) to perform autonomous control processing, optimization processing, or parameter value determination processing for wireless devices that make up the RAN. Execute.
　無線装置は、例えば、基地局であってもよい。基地局は、LTE（Long Term Evolution）をサポートするeNB（evolved Node B）であってもよく、3GPP（3rd Generation Partnership Project）において規定されているいわゆる5G（Fifth Generation）の無線通信方式をサポートする基地局であってもよい。もしくは、無線装置は、O-RANアライアンスにおいて規定されている、RU（Radio Unit）、DU（Distributed Unit）、もしくはCU（Central Unit）であってもよい。RU、DU、CUは、それぞれ、O（O-RAN）-RU、O-DU、O-CU、と称されてもよい。 A wireless device may be, for example, a base station. The base station may be an eNB (evolved Node B) that supports LTE (Long Term Evolution), and supports the so-called 5G (Fifth Generation) wireless communication system specified in 3GPP (3rd Generation Partnership Project). It may be a base station. Alternatively, the wireless device may be an RU (Radio Unit), a DU (Distributed Unit), or a CU (Central Unit) defined by the O-RAN Alliance. RU, DU, and CU may be referred to as O(O-RAN)-RU, O-DU, and O-CU, respectively.
　O-RU、O-DU、及びO-CUと、Non-Real Time RIC及びNear-Real Time RICとの間においては、O-RANアライアンスにおいて規定されているインタフェースに従って信号が送受信される。 Signals are transmitted and received between the O-RU, O-DU, and O-CU and the Non-Real Time RIC and Near-Real Time RIC according to the interface specified by the O-RAN Alliance.
　続いて、図４を用いて実施の形態２にかかる管理装置２０の構成例について説明する。管理装置２０は、管理装置１０に生成部２１、変換部２２、及び表示部２３が追加された構成である。以下においては、管理装置１０と異なる機能及び管理装置１０において説明した各種機能の詳細な説明等を主に説明する。 Next, a configuration example of the management device 20 according to the second embodiment will be described using FIG. The management device 20 is configured by adding a generation unit 21 , a conversion unit 22 , and a display unit 23 to the management device 10 . In the following, functions different from those of the management apparatus 10 and detailed descriptions of various functions described in the management apparatus 10 will be mainly described.
　管理装置２０は、複数のAIエンジン２５を管理する。複数のAIエンジン２５のそれぞれは、コンテナにおいて管理されてもよい。コンテナにおいて管理されているAIエンジン２５は、コンテナごとに独立して動作することが可能である。コンテナは、例えば、管理装置２０内に設けられたメモリの領域であってもよい。 The management device 20 manages multiple AI engines 25. Each of the multiple AI engines 25 may be managed in a container. The AI engine 25 managed in the container can operate independently for each container. A container may be, for example, a memory area provided within the management device 20 .
　複数のAIエンジン２５は、共通の目的の分析処理を実行するAIエンジンである。複数のAIエンジン２５のそれぞれは、例えば、異なるベンダによって生成されAIエンジンであってもよく、もしくは、ソフトウェアバージョンの異なるAIエンジンであってもよい。 A plurality of AI engines 25 are AI engines that execute analysis processing for a common purpose. Each of the plurality of AI engines 25 may be, for example, an AI engine produced by a different vendor, or an AI engine with a different software version.
　管理装置２０は、既存のAIエンジンと共通の目的の分析処理を実行する新たなAIエンジンが提供され、新たなAIエンジンが所望の分析精度を有する場合に、新たなAIエンジンにおける分析結果に基づいて、無線装置３０を管理もしくは制御する。言い換えると、新たなAIエンジンが所望の分析精度を有さない場合、既存のAIエンジンによる分析結果に基づいて、無線装置３０を管理もしくは制御する。 The management device 20 is provided with a new AI engine that executes analysis processing for a common purpose with the existing AI engine, and when the new AI engine has desired analysis accuracy, to manage or control the wireless device 30 . In other words, if the new AI engine does not have the desired analysis accuracy, it manages or controls the wireless device 30 based on the analysis results of the existing AI engine.
　生成部２１は、AIエンジン２５における分析処理に用いられる分析対象データを生成する。例えば、生成部２１は、無線装置、さらに、通信ネットワーク上の各種装置等から収集した監視データを用いて分析対象データを生成する。監視データとは、例えば、通信ネットワーク上の各種装置がセンサ装置である場合に、センサデータを含んでもよい。さらに、監視データは、各種装置が生成するデータであってもよく、各種装置における管理データと称されてもよい。管理データは、例えば、装置におけるプロセッサの使用率、データ処理量、データ転送速度、等を含んでもよい。管理データは、保守データと言い換えられてもよい。通信ネットワークは、無線アクセス装置及びコアネットワーク装置等を接続する移動通信ネットワークであってもよく、保守用装置が接続されている保守用ネットワークであってもよい。 The generation unit 21 generates analysis target data used for analysis processing in the AI engine 25. For example, the generation unit 21 generates analysis target data using monitoring data collected from wireless devices and various devices on a communication network. Monitoring data may include, for example, sensor data when various devices on the communication network are sensor devices. Furthermore, the monitoring data may be data generated by various devices, and may be referred to as management data in various devices. Management data may include, for example, processor utilization in the device, data throughput, data transfer rate, and the like. Management data may also be called maintenance data. The communication network may be a mobile communication network connecting the radio access device and the core network device, etc., or may be a maintenance network to which the maintenance device is connected.
　生成部２１は、監視データを統計処理することによって分析対象データを生成してもよい。統計処理は、例えば、複数の監視データの平均値、中央値、最小値、最大値等の値の算出であってもよく、複数の監視データを組み合わせたデータの生成であってもよい。分析対象データは、監視データが加工された加工データと称されてもよい。 The generation unit 21 may generate analysis target data by statistically processing the monitoring data. Statistical processing may be, for example, calculation of values such as an average value, median value, minimum value, or maximum value of a plurality of pieces of monitoring data, or generation of data combining a plurality of pieces of monitoring data. Analysis target data may be referred to as processed data obtained by processing monitoring data.
　選択部１２は、監視データを生成する無線装置及び通信ネットワーク上の各種装置のうち、少なくとも一つの装置を選択してもよい。例えば、選択部１２は、選択した装置の識別情報を含む取得指示信号を生成し、取得信号を生成部２１へ出力してもよい。生成部２１は、取得指示信号に含まれる識別情報によって特定される装置から監視データを収集してもよい。取得指示信号に含まれる装置の識別情報は、管理装置２０を操作する管理者によって、入力インタフェースを介して選択部１２に入力されてもよい。 The selection unit 12 may select at least one device from the wireless device that generates the monitoring data and various devices on the communication network. For example, the selection unit 12 may generate an acquisition instruction signal including identification information of the selected device and output the acquisition signal to the generation unit 21 . The generation unit 21 may collect monitoring data from a device specified by identification information included in the acquisition instruction signal. The identification information of the device included in the acquisition instruction signal may be input to the selection unit 12 via the input interface by an administrator who operates the management device 20 .
　さらに、選択部１２は、複数のAIエンジン２５の中から、制御部１１が並行稼働する２以上のAIエンジン２５を選択する。制御部１１は、選択部１２において選択されたAIエンジン２５に対して、生成部２１において生成された分析対象データを出力する。選択部１２は、分析対象データの出力先であるAIエンジン２５の識別情報を含む出力先指示信号を生成し、出力先指示信号を制御部１１へ出力してもよい。制御部１１は、出力先指示信号に含まれる識別情報によって特定されるAIエンジン２５へ、分析対象データを出力してもよい。出力先指示信号に含まれるAIエンジンの識別情報は、管理装置２０を操作する管理者によって、入力インタフェースを介して選択部１２に入力されてもよい。 Furthermore, the selection unit 12 selects two or more AI engines 25 that the control unit 11 operates in parallel from among the plurality of AI engines 25 . The control unit 11 outputs the analysis target data generated by the generation unit 21 to the AI engine 25 selected by the selection unit 12 . The selection unit 12 may generate an output destination instruction signal including identification information of the AI engine 25 that is the output destination of the analysis target data, and output the output destination instruction signal to the control unit 11 . The control unit 11 may output the analysis target data to the AI engine 25 specified by the identification information included in the output destination instruction signal. The AI engine identification information included in the output destination instruction signal may be input to the selection unit 12 via the input interface by a manager who operates the management device 20 .
　変換部２２は、生成部２１において生成された分析対象データを、選択部１２において選択されたAIエンジン２５が分析可能な形態の分析対象データへ変換する。ソフトウェアバージョンの異なる複数のAIエンジン２５は、分析処理に使用するデータのデータ形式もしくはデータ構造等が異なる場合がある。また、異なるベンダによって生成された複数のAIエンジン２５も同様に、分析処理に使用するデータのデータ形式もしくはデータ構造等が異なる場合がある。例えば、変換部２２は、生成部２１において生成された分析対象データを、選択部１２において選択されたAIエンジン２５において分析可能なデータ形式、データ構造等へ変換する。言い換えると、変換部２２は、生成部２１において生成された分析対象データを、選択部１２において選択されたAIエンジン２５において読み取り可能なデータ形式、データ構造等へ変換する。 The conversion unit 22 converts the analysis target data generated by the generation unit 21 into analysis target data in a form that can be analyzed by the AI engine 25 selected by the selection unit 12 . A plurality of AI engines 25 with different software versions may differ in the data format or data structure of data used for analysis processing. Similarly, a plurality of AI engines 25 generated by different vendors may have different data formats or data structures of data used for analysis processing. For example, the conversion unit 22 converts the analysis target data generated by the generation unit 21 into a data format, data structure, etc. that can be analyzed by the AI engine 25 selected by the selection unit 12 . In other words, the conversion unit 22 converts the analysis target data generated by the generation unit 21 into a data format, data structure, etc. that can be read by the AI engine 25 selected by the selection unit 12 .
　制御部１１は、変換部２２において変換された分析対象データを、選択部１２において選択されたAIエンジン２５へ出力する。 The control unit 11 outputs the analysis target data converted by the conversion unit 22 to the AI engine 25 selected by the selection unit 12 .
　さらに、制御部１１は、並行稼働している複数のAIエンジン２５における分析結果を表示部２３へ出力する。表示部２３は、例えば、ディスプレイ装置であってもよい。表示部２３は、管理装置２０と一体として用いられるディスプレイ装置であってもよく、ケーブルもしくはネットワークを介して管理装置２０と接続されるディスプレイ装置であってもよい。 Furthermore, the control unit 11 outputs the analysis results of the multiple AI engines 25 operating in parallel to the display unit 23 . The display unit 23 may be, for example, a display device. The display unit 23 may be a display device used integrally with the management device 20, or may be a display device connected to the management device 20 via a cable or network.
　表示部２３は、受け取った分析結果を表示する。具体的には、表示部２３は、AIエンジン２５毎に、分析結果を表示する。管理装置２０の管理者は、表示部２３に表示される分析結果を確認し、無線装置３０の制御に用いるAIエンジン２５を選択する。例えば、管理者は、管理装置２０における入力インタフェースを介して、選択結果を管理装置２０へ入力する。管理者は、例えば、所望の分析結果に近い分析結果を出力したAIエンジン２５を選択する。 The display unit 23 displays the received analysis results. Specifically, the display unit 23 displays analysis results for each AI engine 25 . The administrator of the management device 20 confirms the analysis result displayed on the display unit 23 and selects the AI engine 25 to be used for controlling the wireless device 30 . For example, the administrator inputs the selection result to the management device 20 via the input interface of the management device 20 . The administrator selects, for example, the AI engine 25 that outputs an analysis result close to the desired analysis result.
　選択部１２は、管理者から入力された選択結果を、選択指示信号として受け取る。選択指示信号は、管理者が選択したAIエンジン２５の識別情報を含む。通信部１３は、選択部１２において選択されたAIエンジン２５の分析結果を無線装置３０へ送信する。 The selection unit 12 receives the selection result input by the administrator as a selection instruction signal. The selection instruction signal includes identification information of the AI engine 25 selected by the administrator. The communication unit 13 transmits the analysis result of the AI engine 25 selected by the selection unit 12 to the wireless device 30 .
　無線装置３０は、管理装置２０から受信した分析結果を用いて、例えば、リソースの最適化、自動制御、等を実行する。 Using the analysis results received from the management device 20, the wireless device 30 performs, for example, resource optimization, automatic control, and the like.
　ここで、図５を用いて、管理装置２０内において扱われるデータに関するデータフローについて説明する。図５は、管理装置２０内において扱われるデータが、データ取得、データ加工、AI/ML、及び措置実行の各工程を経て制御対象装置である無線装置３０へ送信されることが示されている。 Here, the data flow regarding data handled in the management device 20 will be described using FIG. FIG. 5 shows that data handled in the management device 20 is transmitted to the radio device 30, which is the device to be controlled, through each process of data acquisition, data processing, AI/ML, and action execution. .
　データ取得に関する工程において、生成部２１は、無線装置、さらに、通信ネットワーク上の各種装置等から監視データを収集する。生成部２１は、選択部１２において選択された無線装置もしくは通信ネットワーク上の各種装置から、監視データを収集してもよい。データ加工に関する工程において、生成部２１は、収集した監視データを用いて分析対象データを生成する。 In the data acquisition process, the generation unit 21 collects monitoring data from wireless devices and various devices on the communication network. The generation unit 21 may collect monitoring data from wireless devices selected by the selection unit 12 or from various devices on the communication network. In a process related to data processing, the generation unit 21 generates analysis target data using collected monitoring data.
　AI/MLに関する工程において、複数のAIエンジン２５に含まれる２つのAIエンジン２５が、分析対象データを用いた分析処理を実行する。図５においては、分析処理を実行するAIエンジン２５の数が２つであることが示されているが、分析処理を実行するAIエンジン２５の数は、２つに制限されない。また、例えば、分析処理を実行する２つのAIエンジン２５は、それぞれが異なるソフトウェアバージョンであってもよい。図５においては、ソフトウェアバージョン１（V1）のAIエンジン２５と、ソフトウェアバージョン２（V2）のAIエンジン２５とが、分析処理を実行することが示されている。例えば、ソフトウェアバージョン２が、ソフトウェアバージョン１よりも新しいバージョンであってもよい。例えば、図５のデータフローに従って、新たに提供されたソフトウェアバージョン２に関するAIエンジン２５の分析処理の精度を商用環境において確認するための試験が、行われてもよい。 In the process related to AI/ML, two AI engines 25 included in multiple AI engines 25 execute analysis processing using analysis target data. Although FIG. 5 shows that the number of AI engines 25 executing analysis processing is two, the number of AI engines 25 executing analysis processing is not limited to two. Also, for example, the two AI engines 25 that perform analysis processing may have different software versions. FIG. 5 shows that the AI engine 25 of software version 1 (V1) and the AI engine 25 of software version 2 (V2) execute analysis processing. For example, software version 2 may be a newer version than software version 1. For example, according to the data flow of FIG. 5, a test may be conducted in a commercial environment to confirm the accuracy of the AI engine 25 analysis process for the newly provided software version 2.
　AI/MLに関する工程において、選択部１２は、分析対象データを出力するAIエンジン２５を選択してもよい。つまり、選択部１２は、分析処理を実行するAIエンジン２５を選択してもよい。 In the AI/ML-related process, the selection unit 12 may select the AI engine 25 that outputs analysis target data. That is, the selection unit 12 may select the AI engine 25 that executes analysis processing.
　措置実行に関する工程において、通信部１３は、それぞれが異なるソフトウェアバージョンを利用する２つのAIエンジン２５における分析結果のうち、いずれかの分析結果を無線装置３０へ送信する。通信部１３は、２つのAIエンジン２５のうち、選択部１２において選択されたAIエンジン２５における分析結果を無線装置３０へ送信する。選択部１２は、管理装置２０の管理者によって入力された選択結果を受け付け、２つのAIエンジン２５のうち、いずれかのAIエンジン２５を選択する。選択部１２は、はじめに、データ取得から措置実行までのデータフローを決定し、生成部２１等の管理装置２０の構成要素は、選択部１２において決定されたデータフローに従って、処理を実行してもよい。 In the process related to action execution, the communication unit 13 transmits to the wireless device 30 one of the analysis results of the two AI engines 25 using different software versions. The communication unit 13 transmits the analysis result of the AI engine 25 selected by the selection unit 12 from among the two AI engines 25 to the wireless device 30 . The selection unit 12 receives a selection result input by the administrator of the management device 20 and selects one of the two AI engines 25 . The selection unit 12 first determines a data flow from data acquisition to action execution, and components of the management device 20 such as the generation unit 21 may execute processing according to the data flow determined by the selection unit 12. good.
　管理者は、表示部２３に表示された２つのAIエンジン２５における分析結果を確認し、ソフトウェアバージョン２によって動作する、新たに提供されたAIエンジン２５における分析結果が所望の結果を示しているか否かを確認する。管理者は、新たに提供されたAIエンジン２５における分析結果が所望の結果ではない場合、既に商用環境において動作していたソフトウェアバージョン１によって動作するAIエンジン２５における分析結果を利用することを決定してもよい。この場合、図５に示されるように、通信部１３は、ソフトウェアバージョン１によって動作するAIエンジン２５における分析結果を無線装置３０へ送信する。管理者は、新たに提供されたAIエンジン２５における分析結果が所望の結果である場合、ソフトウェアバージョン２によって動作するAIエンジン２５における分析結果を利用することを決定してもよい。 The administrator confirms the analysis results of the two AI engines 25 displayed on the display unit 23, and whether or not the analysis results of the newly provided AI engine 25 operated by software version 2 show the desired results. Check whether If the newly provided analysis result in the AI engine 25 is not the desired result, the administrator decides to use the analysis result in the AI engine 25 operating with software version 1 that has already been operating in the commercial environment. may In this case, as shown in FIG. 5, the communication unit 13 transmits the analysis result of the AI engine 25 operating with the software version 1 to the wireless device 30 . The administrator may decide to utilize the analysis results in the AI engine 25 running with software version 2 if the newly provided analysis results in the AI engine 25 are the desired results.
　管理装置２０の管理者は、図５のデータフローを適宜変更することが可能である。例えば、管理者は、分析対象データを生成するために必要な監視データを選択し、また、AI/MLに関する工程を実行するAIエンジンを選択し、さらに、分析結果を選択してもよい。選択部１２は、管理装置２０の入力インタフェースを介して管理者から入力された選択情報に従って、選択処理を実行する。例えば、表示部２３は、図５に示すデータフローを表示し、管理者は、表示部２３に表示されるデータフローの矢印を選択することによって、データフローを適宜変更してもよい。 The administrator of the management device 20 can change the data flow of FIG. 5 as appropriate. For example, the administrator may select monitoring data necessary to generate data to be analyzed, select an AI engine for executing processes related to AI/ML, and further select analysis results. The selection unit 12 executes selection processing according to selection information input by the administrator via the input interface of the management device 20 . For example, the display unit 23 may display the data flow shown in FIG. 5 and the administrator may change the data flow as appropriate by selecting the data flow arrows displayed on the display unit 23 .
　続いて図６を用いて実施の形態２にかかる分析結果の送信処理の流れについて説明する。はじめに、生成部２１は、選択部１２において選択された無線装置もしくは通信ネットワーク上の各種装置から、監視データを収集する（Ｓ２１）。監視データには、センサデータ、無線装置等の状態を示すデータ等が含まれてもよい。 Next, the flow of analysis result transmission processing according to the second embodiment will be described with reference to FIG. First, the generation unit 21 collects monitoring data from wireless devices selected by the selection unit 12 or various devices on the communication network (S21). The monitoring data may include sensor data, data indicating the state of the wireless device, and the like.
　次に、生成部２１は、収集した監視データを用いて分析対象データを生成する（Ｓ２２）。次に、選択部１２は、分析データの出力先となるAIエンジン２５を選択する（Ｓ２３）。 Next, the generation unit 21 generates analysis target data using the collected monitoring data (S22). Next, the selection unit 12 selects the AI engine 25 as the output destination of the analysis data (S23).
　次に、変換部２２は、選択部１２において選択されたAIエンジン２５が利用可能なように、分析対象データを変換する（Ｓ２４）。変換部２２は、分析対象データのデータ形式もしくはデータ構造を変換してもよい。データ形式は、例えば、データフォーマットと称されてもよい。データ構造は、例えば、データの配列等によって定められてもよい。変換部２２は、それぞれのAIエンジン２５と、それぞれのAIエンジン２５におけるデータ形式もしくはデータ構造とを予め関連付けて管理していてもよい。 Next, the conversion unit 22 converts the analysis target data so that it can be used by the AI engine 25 selected by the selection unit 12 (S24). The conversion unit 22 may convert the data format or data structure of the analysis target data. A data format may, for example, be referred to as a data format. The data structure may be defined by, for example, an array of data. The conversion unit 22 may manage each AI engine 25 and the data format or data structure in each AI engine 25 in association with each other in advance.
　次に、変換部２２において変換された分析対象データを受け取ったAIエンジン２５は、分析処理を実行する（Ｓ２５）。制御部１１は、複数のAIエンジン２５における分析処理の結果である分析結果を表示部２３へ表示してもよい。 Next, the AI engine 25 that has received the analysis target data converted by the conversion unit 22 executes analysis processing (S25). The control unit 11 may display analysis results, which are results of analysis processing in the plurality of AI engines 25 , on the display unit 23 .
　次に、選択部１２は、AIエンジン２５を選択する（Ｓ２６）。選択部１２は、管理者等から入力されたAIエンジン２５の選択結果に基づいて、AIエンジン２５を選択する。 Next, the selection unit 12 selects the AI engine 25 (S26). The selection unit 12 selects the AI engine 25 based on the selection result of the AI engine 25 input by the administrator or the like.
　次に、通信部１３は、選択部１２において選択されたAIエンジン２５の分析結果を無線装置３０へ送信する（Ｓ２７）。 Next, the communication unit 13 transmits the analysis result of the AI engine 25 selected by the selection unit 12 to the wireless device 30 (S27).
　以上説明したように、実施の形態２にかかる管理装置２０は、AIエンジン２５を並行稼働する。さらに、管理装置２０は、複数のAIエンジン２５における分析結果を確認した管理者から選択されたAIエンジン２５を選択し、選択したAIエンジン２５が生成した分析結果を、無線装置３０へ送信する。これにより、複数のAIエンジン２５を商用環境にて同時に動作させ、新しいAIエンジン２５が所望の分析結果を出力する場合に、商用環境にて稼働するAIエンジン２５を、新しいAIエンジン２５へ変更することができる。新しいAIエンジン２５は、新しいソフトウェアバージョンのAIエンジンもしくは現在稼働しているAIエンジン２５を生成したベンダとは異なるベンダから提供されたAIエンジンであってもよい。 As described above, the management device 20 according to the second embodiment operates the AI engine 25 in parallel. Furthermore, the management device 20 selects an AI engine 25 selected by the administrator who has confirmed the analysis results of the plurality of AI engines 25, and transmits the analysis results generated by the selected AI engine 25 to the wireless device 30. In this way, a plurality of AI engines 25 are simultaneously operated in the commercial environment, and when the new AI engine 25 outputs desired analysis results, the AI engine 25 operating in the commercial environment is changed to the new AI engine 25. be able to. The new AI engine 25 may be a new software version of the AI engine or an AI engine provided by a vendor different from the vendor that produced the currently running AI engine 25 .
　さらに、変換部２２は、分析対象データを、それぞれのAIエンジン２５が利用可能なようにデータ形式もしくはデータ構造を変換する。これより、管理装置２０においては、複数のAIエンジン２５が共通の分析対象データを用いて、分析処理を実行することができる。つまり、管理装置２０においては、AIエンジン２５毎に、分析対象データを生成する必要が無く、AIエンジン２５毎に分析対象データを生成する場合と比較して、分析対象データの生成処理を軽減することができる。 Furthermore, the conversion unit 22 converts the data format or data structure of the data to be analyzed so that each AI engine 25 can use it. Thus, in the management device 20, a plurality of AI engines 25 can use common analysis target data to execute analysis processing. In other words, in the management device 20, there is no need to generate analysis target data for each AI engine 25, and compared to the case where analysis target data is generated for each AI engine 25, the generation processing of analysis target data is reduced. be able to.
　図７は、管理装置１０及び管理装置２０（以下、管理装置１０等とする）の構成例を示すブロック図である。図７を参照すると、管理装置１０等は、ネットワークインタフェース１２０１、プロセッサ１２０２、及びメモリ１２０３を含む。ネットワークインタフェース１２０１は、ネットワークノードと通信するために使用されてもよい。ネットワークインタフェース１２０１は、例えば、IEEE 802.3 seriesに準拠したネットワークインタフェースカード（NIC）を含んでもよい。IEEEは、Institute of Electrical and Electronics Engineersを表す。 FIG. 7 is a block diagram showing a configuration example of the management device 10 and the management device 20 (hereinafter referred to as the management device 10 and the like). Referring to FIG. 7 , the management device 10 and the like include a network interface 1201 , a processor 1202 and a memory 1203 . Network interface 1201 may be used to communicate with network nodes. Network interface 1201 may include, for example, an IEEE 802.3 series compliant network interface card (NIC). IEEE stands for Institute of Electrical and Electronics Engineers.
　プロセッサ１２０２は、メモリ１２０３からソフトウェア（コンピュータプログラム）を読み出して実行することで、上述の実施形態においてフローチャートを用いて説明された管理装置１０等の処理を行う。プロセッサ１２０２は、例えば、マイクロプロセッサ、MPU、又はCPUであってもよい。プロセッサ１２０２は、複数のプロセッサを含んでもよい。 The processor 1202 reads and executes software (computer program) from the memory 1203 to perform the processing of the management device 10 and the like described using the flowcharts in the above embodiments. Processor 1202 may be, for example, a microprocessor, MPU, or CPU. Processor 1202 may include multiple processors.
　メモリ１２０３は、揮発性メモリ及び不揮発性メモリの組み合わせによって構成される。メモリ１２０３は、プロセッサ１２０２から離れて配置されたストレージを含んでもよい。この場合、プロセッサ１２０２は、図示されていないI/O（Input/Output）インタフェースを介してメモリ１２０３にアクセスしてもよい。 The memory 1203 is composed of a combination of volatile memory and non-volatile memory. Memory 1203 may include storage remotely located from processor 1202 . In this case, the processor 1202 may access the memory 1203 via an I/O (Input/Output) interface (not shown).
　図７の例では、メモリ１２０３は、ソフトウェアモジュール群を格納するために使用される。プロセッサ１２０２は、これらのソフトウェアモジュール群をメモリ１２０３から読み出して実行することで、上述の実施形態において説明された管理装置１０等の処理を行うことができる。 In the example of FIG. 7, memory 1203 is used to store software modules. The processor 1202 reads and executes these software modules from the memory 1203, thereby performing the processing of the management apparatus 10 and the like described in the above embodiments.
　図７を用いて説明したように、上述の実施形態における管理装置１０等が有するプロセッサの各々は、図面を用いて説明されたアルゴリズムをコンピュータに行わせるための命令群を含む１又は複数のプログラムを実行する。 As described with reference to FIG. 7, each of the processors included in the management device 10 and the like in the above-described embodiments includes one or more programs containing instructions for causing a computer to execute the algorithm described with reference to the drawings. to run.
　上述の例において、プログラムは、コンピュータに読み込まれた場合に、実施形態で説明された１又はそれ以上の機能をコンピュータに行わせるための命令群（又はソフトウェアコード）を含む。プログラムは、非一時的なコンピュータ可読媒体又は実体のある記憶媒体に格納されてもよい。限定ではなく例として、コンピュータ可読媒体又は実体のある記憶媒体は、random-access memory（RAM）、read-only memory（ROM）、フラッシュメモリ、solid-state drive（SSD）又はその他のメモリ技術、CD-ROM、digital versatile disc（DVD）、Blu-ray（登録商標）ディスク又はその他の光ディスクストレージ、磁気カセット、磁気テープ、磁気ディスクストレージ又はその他の磁気ストレージデバイスを含む。プログラムは、一時的なコンピュータ可読媒体又は通信媒体上で送信されてもよい。限定ではなく例として、一時的なコンピュータ可読媒体又は通信媒体は、電気的、光学的、音響的、またはその他の形式の伝搬信号を含む。 In the above examples, the program includes instructions (or software code) that, when read into a computer, cause the computer to perform one or more of the functions described in the embodiments. The program may be stored in a non-transitory computer-readable medium or tangible storage medium. By way of example, and not limitation, computer readable media or tangible storage media may include random-access memory (RAM), read-only memory (ROM), flash memory, solid-state drives (SSD) or other memory technology, CDs - ROM, digital versatile disc (DVD), Blu-ray disc or other optical disc storage, magnetic cassette, magnetic tape, magnetic disc storage or other magnetic storage device. The program may 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.
　なお、本開示は上記実施の形態に限られたものではなく、趣旨を逸脱しない範囲で適宜変更することが可能である。 It should be noted that the present disclosure is not limited to the above embodiments, and can be modified as appropriate without departing from the scope.
　上記の実施形態の一部又は全部は、以下の付記のようにも記載されうるが、以下には限られない。
　（付記１）
　第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働する制御部と、
　前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択する選択部と、
　前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果を利用する制御対象装置へ出力する通信部と、を備える管理装置。
　（付記２）
　前記第１のAIエンジン及び第２のAIエンジンは、それぞれのソフトウェアバージョンが異なる、付記１に記載の管理装置。
　（付記３）
　前記制御部は、
　前記第１のAIエンジンの分析結果及び前記第２のAIエンジンの分析結果を表示装置へ出力し、
　前記選択部は、
　前記第１のAIエンジンの分析結果及び前記第２のAIエンジンの分析結果を視認したユーザから入力された前記選択指示信号に基づいて、前記第１のAIエンジンもしくは前記第２のAIエンジンのいずれかを選択する、付記１又は２に記載の管理装置。
　（付記４）
　分析処理を実行するために、前記第１のAIエンジン及び前記第２のAIエンジンにおいて共通に用いられる分析対象データを生成する生成部をさらに備える、付記１乃至３のいずれか１項に記載の管理装置。
　（付記５）
　前記選択部は、
　前記第１のAIエンジン及び前記第２のAIエンジンを含む複数のAIエンジンの中から、少なくとも１以上のAIエンジンの識別情報を含む出力先指示信号に基づいて、生成された前記分析対象データの出力先を選択する、付記４に記載の管理装置。
　（付記６）
　前記選択部は、
　前記分析対象データを生成するために用いられる監視データを生成した複数の監視装置のうち、少なくとも１以上の監視装置の識別情報を含む取得指示信号に基づいて、前記監視データを取得する前記監視装置を選択する、付記４又は５に記載の管理装置。
　（付記７）
　前記分析対象データを、前記第１のAIエンジン及び前記第２のAIエンジンのそれぞれにおいて利用可能な形式のデータに変換する変換部をさらに備える、付記４乃至６のいずれか１項に記載の管理装置。
　（付記８）
　第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働する制御部と、前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択する選択部と、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を送信する通信部と、を備える管理装置と、
　前記分析結果を利用した自律制御を実行する制御対象装置と、を備える通信システム。
　（付記９）
　前記第１のAIエンジン及び第２のAIエンジンは、それぞれのソフトウェアバージョンが異なる、付記８に記載の通信システム。
　（付記１０）
　第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働し、
　前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択し、
　前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果を利用する制御対象装置へ出力する、制御方法。
　（付記１１）
　第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働し、
　前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択し、
　前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果を利用する制御対象装置へ出力することをコンピュータに実行させるプログラムを格納する非一時的なコンピュータ可読媒体。 Some or all of the above-described embodiments can also be described in the following supplementary remarks, but are not limited to the following.
(Appendix 1)
a control unit that operates in parallel a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine;
a selection unit that selects either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine;
The analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis of the second AI engine. and a communication unit that outputs the result to a controlled device that uses the result.
(Appendix 2)
The management device according to appendix 1, wherein the first AI engine and the second AI engine have different software versions.
(Appendix 3)
The control unit
outputting the analysis result of the first AI engine and the analysis result of the second AI engine to a display device;
The selection unit
Either the first AI engine or the second AI engine is selected based on the selection instruction signal input by the user who visually recognizes the analysis result of the first AI engine and the analysis result of the second AI engine. 3. The management device according to appendix 1 or 2, which selects:
(Appendix 4)
4. The method according to any one of Appendices 1 to 3, further comprising a generation unit that generates analysis target data that is commonly used in the first AI engine and the second AI engine in order to perform analysis processing. management device.
(Appendix 5)
The selection unit
of the data to be analyzed generated based on an output destination instruction signal containing identification information of at least one or more AI engines from among a plurality of AI engines including the first AI engine and the second AI engine; 5. The management device according to appendix 4, which selects an output destination.
(Appendix 6)
The selection unit
The monitoring device that acquires the monitoring data based on an acquisition instruction signal that includes identification information of at least one of the plurality of monitoring devices that generated the monitoring data used to generate the analysis target data. 6. The management device according to appendix 4 or 5, selecting
(Appendix 7)
7. The management according to any one of Appendices 4 to 6, further comprising a conversion unit that converts the analysis target data into data in a format that can be used by each of the first AI engine and the second AI engine. Device.
(Appendix 8)
A control unit that operates in parallel with a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine, and the first AI engine or the second AI a selection unit that selects either the first AI engine or the second AI engine based on a selection instruction signal containing engine identification information; a management device comprising a communication unit that transmits analysis results of a selected AI engine among the analysis results of the AI engines;
and a control target device that executes autonomous control using the analysis result.
(Appendix 9)
The communication system according to appendix 8, wherein the first AI engine and the second AI engine have different software versions.
(Appendix 10)
A first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine are operated in parallel,
selecting either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine;
The analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis of the second AI engine. A control method that outputs results to controlled devices that use them.
(Appendix 11)
A first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine are operated in parallel,
selecting either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine;
The analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis of the second AI engine. A non-transitory computer-readable medium that stores a program that causes a computer to output results to a controlled device that utilizes the results.
　１０　管理装置
　１１　制御部
　１２　選択部
　１３　通信部
　２０　管理装置
　２１　生成部
　２２　変換部
　２３　表示部
　２５　AIエンジン
　３０　無線装置 REFERENCE SIGNS LIST 10 management device 11 control unit 12 selection unit 13 communication unit 20 management device 21 generation unit 22 conversion unit 23 display unit 25 AI engine 30 wireless device

Claims (11)

  1. 　第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働する制御部と、
    　前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択する選択部と、
    　前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果を利用する制御対象装置へ出力する通信部と、を備える管理装置。     a control unit that operates in parallel a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine;
    a selection unit that selects either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine;
    The analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis of the second AI engine. and a communication unit that outputs the result to a controlled device that uses the result.
  2. 　前記第１のAIエンジン及び第２のAIエンジンは、それぞれのソフトウェアバージョンが異なる、請求項１に記載の管理装置。 The management device according to claim 1, wherein the first AI engine and the second AI engine have different software versions.
  3. 　前記制御部は、
    　前記第１のAIエンジンの分析結果及び前記第２のAIエンジンの分析結果を表示装置へ出力し、
    　前記選択部は、
    　前記第１のAIエンジンの分析結果及び前記第２のAIエンジンの分析結果を視認したユーザから入力された前記選択指示信号に基づいて、前記第１のAIエンジンもしくは前記第２のAIエンジンのいずれかを選択する、請求項１又は２に記載の管理装置。     The control unit
    outputting the analysis result of the first AI engine and the analysis result of the second AI engine to a display device;
    The selection unit
    Either the first AI engine or the second AI engine is selected based on the selection instruction signal input by the user who visually recognizes the analysis result of the first AI engine and the analysis result of the second AI engine. The management device according to claim 1 or 2, which selects whether or not.
  4. 　分析処理を実行するために、前記第１のAIエンジン及び前記第２のAIエンジンにおいて共通に用いられる分析対象データを生成する生成部をさらに備える、請求項１乃至３のいずれか１項に記載の管理装置。 4. The method according to any one of claims 1 to 3, further comprising a generation unit that generates analysis target data that is commonly used in the first AI engine and the second AI engine in order to perform analysis processing. management device.
  5. 　前記選択部は、
    　前記第１のAIエンジン及び前記第２のAIエンジンを含む複数のAIエンジンの中から、少なくとも１以上のAIエンジンの識別情報を含む出力先指示信号に基づいて、生成された前記分析対象データの出力先を選択する、請求項４に記載の管理装置。     The selection unit
    of the data to be analyzed generated based on an output destination instruction signal containing identification information of at least one or more AI engines from among a plurality of AI engines including the first AI engine and the second AI engine; 5. The management device according to claim 4, which selects an output destination.
  6. 　前記選択部は、
    　前記分析対象データを生成するために用いられる監視データを生成した複数の監視装置のうち、少なくとも１以上の監視装置の識別情報を含む取得指示信号に基づいて、前記監視データを取得する前記監視装置を選択する、請求項４又は５に記載の管理装置。     The selection unit
    The monitoring device that acquires the monitoring data based on an acquisition instruction signal that includes identification information of at least one of the plurality of monitoring devices that generated the monitoring data used to generate the analysis target data. 6. The management device according to claim 4 or 5, which selects .
  7. 　前記分析対象データを、前記第１のAIエンジン及び前記第２のAIエンジンのそれぞれにおいて利用可能な形式のデータに変換する変換部をさらに備える、請求項４乃至６のいずれか１項に記載の管理装置。 7. The method according to any one of claims 4 to 6, further comprising a conversion unit that converts the data to be analyzed into data in a format that can be used by each of the first AI engine and the second AI engine. management device.
  8. 　第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働する制御部と、前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択する選択部と、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を送信する通信部と、を備える管理装置と、
    　前記分析結果を利用した自律制御を実行する制御対象装置と、を備える通信システム。     A control unit that operates in parallel with a first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine, and the first AI engine or the second AI a selection unit that selects either the first AI engine or the second AI engine based on a selection instruction signal containing engine identification information; a management device comprising a communication unit that transmits analysis results of a selected AI engine among the analysis results of the AI engines;
    and a control target device that executes autonomous control using the analysis result.
  9. 　前記第１のAIエンジン及び第２のAIエンジンは、それぞれのソフトウェアバージョンが異なる、請求項８に記載の通信システム。 The communication system according to claim 8, wherein the first AI engine and the second AI engine have different software versions.
  10. 　第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働し、
    　前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択し、
    　前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果を利用する制御対象装置へ出力する、制御方法。     A first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine are operated in parallel,
    selecting either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine;
    The analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis of the second AI engine. A control method that outputs results to controlled devices that use them.
  11. 　第１のAIエンジンと、前記第１のAIエンジンと共通の目的の分析処理を実行する第２のAIエンジンと、を並行稼働し、
    　前記第１のAIエンジンもしくは前記第２のAIエンジンの識別情報を含む選択指示信号に基づいて、前記第１のAIエンジン及び前記第２のAIエンジンのいずれかを選択し、
    　前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果のうち、選択されたAIエンジンの分析結果を、前記第１のAIエンジンの分析結果もしくは前記第２のAIエンジンの分析結果を利用する制御対象装置へ出力することをコンピュータに実行させるプログラムを格納する非一時的なコンピュータ可読媒体。     A first AI engine and a second AI engine that executes analysis processing for a common purpose with the first AI engine are operated in parallel,
    selecting either the first AI engine or the second AI engine based on a selection instruction signal including identification information of the first AI engine or the second AI engine;
    The analysis result of the AI engine selected from the analysis result of the first AI engine or the analysis result of the second AI engine is used as the analysis result of the first AI engine or the analysis of the second AI engine. A non-transitory computer-readable medium that stores a program that causes a computer to output results to a controlled device that utilizes the results.
PCT/JP2021/048186 2021-12-24 2021-12-24 Management device, communication system, control method, and non-transitory computer-readable medium WO2023119623A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/JP2021/048186 WO2023119623A1 (en) 2021-12-24 2021-12-24 Management device, communication system, control method, and non-transitory computer-readable medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2021/048186 WO2023119623A1 (en) 2021-12-24 2021-12-24 Management device, communication system, control method, and non-transitory computer-readable medium

Publications (1)

Publication Number Publication Date
WO2023119623A1 true WO2023119623A1 (en) 2023-06-29

Family

ID=86901807

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2021/048186 WO2023119623A1 (en) 2021-12-24 2021-12-24 Management device, communication system, control method, and non-transitory computer-readable medium

Country Status (1)

Country Link
WO (1) WO2023119623A1 (en)

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
ANONYMOUS: "Adopting AI for mobile network optimization", KDDI, 26 June 2019 (2019-06-26), XP093073921, Retrieved from the Internet <URL:https://news.kddi.com/kddi/corporate/newsrelease/2019/06/26/3878.html> [retrieved on 20230816] *
SHIOZU, AKIHIRO: "Smart OPS that realizes further efficiency and sophistication of network operations", NTT DOCOMO TECHNICAL JOURNAL, JP, vol. 27, no. 1, 23 April 2019 (2019-04-23), JP, pages 28 - 35, XP009547262 *
TANII MASATO: "Technology to optimize networks and improve communication quality with AI to be introduced at au base stations", ITMEDIA, 26 June 2019 (2019-06-26), XP093073924, Retrieved from the Internet <URL:https://www.itmedia.co.jp/news/articles/1906/26/news087.html> [retrieved on 20230816] *
YAMAZAKI ATSUO, INADA MANABU, SUZUKI NAOHIRO, IMOTO NAOKI: "Digital solutions that accelerate value creation: Improving business efficiency with form recognition services using AI-OCR", HITACHI HYORON., HITACHI LTD., TOKYO., JP, vol. 102, no. 3, 20 July 2020 (2020-07-20), JP , pages 86 - 92, XP009547457, ISSN: 0367-5874 *
YOSHII, ISAMU: "Message to appeal using Inter BEE 2020", NEW MEDIA, vol. 38, no. 12, 1 December 2020 (2020-12-01), pages 36 - 37, XP009547456 *

Similar Documents

Publication Publication Date Title
US10956849B2 (en) Microservice auto-scaling for achieving service level agreements
Iftikhar et al. Deltaiot: A self-adaptive internet of things exemplar
D’Oro et al. OrchestRAN: Network automation through orchestrated intelligence in the open RAN
US10747568B2 (en) Systems and methods for managing a cloud computing environment
AU2021279510B2 (en) Network optimisation method, server, network side device, system, and storage medium
Brik et al. On predicting service-oriented network slices performances in 5G: A federated learning approach
EP3742669B1 (en) Machine learning in radio access networks
KR102245341B1 (en) Method for apply predictive model for workload distribution within the cloud edge
EP3780496B1 (en) Feature engineering programming method and apparatus
US11381463B2 (en) System and method for a generic key performance indicator platform
KR20120066116A (en) Web service information processing method and web service compositing method and apparatus using the same
KR102456900B1 (en) Data analysis system based on edge computing and method thereof
US10200240B2 (en) Method for configuring an M2M system
US10990068B2 (en) Technologies for providing cloud-based management of an electrical grid
US11310125B2 (en) AI-enabled adaptive TCA thresholding for SLA assurance
Benedetti et al. Reinforcement learning applicability for resource-based auto-scaling in serverless edge applications
KR20210056655A (en) Method for selecting predict-based migration candidate and target on cloud edge
US11556100B2 (en) Control method, related device, and system
KR20220042928A (en) A method of implementing an self-organizing network for a plurality of access network devices and an electronic device performing the same
WO2023119623A1 (en) Management device, communication system, control method, and non-transitory computer-readable medium
CN113114480A (en) Data reporting method and related equipment
KR20190114126A (en) Software-defined network controller for controlling message transmission to switch based on processing completion delay time of message and method thereof
EP3457634A1 (en) Collection of management plane performance data
WO2024038554A1 (en) Control system, control device, control method, and non-temporary computer-readable medium
CN112953844B (en) Network traffic optimization method and device

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21969054

Country of ref document: EP

Kind code of ref document: A1