Control device for radio access network, and computer-readable storage medium

Abstract

This control device for a RAN comprises: a generation means that generates a learning model by updating, on the basis of training data, a base learning model which has been received from another control device; a first control means that performs inference on the basis of the learning model generated by the generation means and that controls the RAN on the basis of an inference result; a measurement means that measures the inference speed of the learning model generated by the generation means; and a determination means that determines whether or not to update the learning model to increase the inference speed of the learning model, on the basis of a reference value of the inference speed and one or more measurement values of the inference speed which have been measured by the measurement means.

Description

Radio access network control device and computer-readable storage medium

[0001] The present disclosure relates to radio access network (RAN) control techniques.

[0002] Figure 1 shows the control configuration of the RAN proposed by the Open Radio Access Network (O-RAN) Alliance. As shown in Figure 1, a first control function that performs long-term control and a second control function that performs short-term control are defined. In the O-RAN, the first control function is called the non-real-time RAN intelligent controller (Non-RT RIC), and the second control function is called the near-real-time RAN intelligent controller (Near-RT RIC).

[0003] The first control function and the second control function are connected via an A1 interface. The second control function controls components of the RAN, such as a central unit (CU) and distributed units (DU), via an E2 interface. Controlling the CU and DU includes notifying and setting various parameter values ​​used by the CU and DU in their processing, and instructing the CU and DU to perform certain operations. Furthermore, in the following description, the term "RAN" is used as a general term for its components. Therefore, for example, "controlling the RAN" means controlling the CU and DU, which are components of the RAN. The first control function, the second control function, and the RAN are further connected via an O1 interface. The O1 interface can be used to transmit traffic data, performance data, fault data, etc., measured / detected and accumulated by the RAN.

[0004] Non-Patent Document 1 discloses a configuration that uses machine learning for RAN control. Specifically, Non-Patent Document 1 discloses performing machine learning based on various learning data collected from the RAN, etc., to generate a learning model, performing inference using the learning model, and controlling the RAN according to the inference results. In one of the configurations disclosed in Non-Patent Document 1, a first control function generates the learning model. Then, a second control function controls the RAN using the learning model generated by the first control function. Furthermore, in one of the configurations disclosed in Non-Patent Document 1, the first control function generates the learning model and distributes it to a second control function. The second control function updates the learning model received from the first control function by machine learning and controls the RAN using the learning model.

[0005] Non-Patent Documents 2 and 3 disclose various RAN control contents based on inference results from learning models. For example, the RAN control contents include beamforming control, radio resource allocation, traffic prediction, and CU and DU placement control using virtualization technology.

[0006] O-RAN Alliance, "AI / ML workflow description and requirements", O-RAN. WG2. AIML-v01.03, July 2021 M. E. Morocho-Cayamcela, et. al. , "Machine Learning for 5G / B5G Mobile and Wireless Communications: Potential, Limitations, and Future Directions", in IEEE Access, vol. 7, pp. 137184-137206, 2019 J. Kaur, et. al. , "Machine Learning Techniques for 5G and Beyond", in IEEE Access, vol. 9, pp. 23472-23488, 2021

[0007] For example, when the second control function controls a RAN based on a learning model received from the first control function or a modified learning model of the learning model, it is necessary to complete inference within a target time corresponding to the control content. However, there may be cases where the learning model received from the first control function or a modified learning model of the learning model cannot complete inference within the target time. For example, the second control function performs various RAN control operations using multiple learning models generated for each of multiple control contents. Therefore, due to limitations on the computer resources of the second control function, the inference speed using each learning model may vary. Furthermore, the inference speed may also vary depending on changes in the status of the RAN to be controlled. In the present disclosure, the inference speed may be defined as the time from inputting data into a learning model to obtaining an inference result. However, the inference speed may also be defined as the time from inputting data into a learning model to obtaining an inference result and completing control based on the inference result.

[0008] According to one aspect of the present disclosure, a control device of a radio access network (RAN) comprises a generation means for generating a learning model by updating a basic learning model received from another control device based on learning data, a first control means for performing inference based on the learning model generated by the generation means and controlling the RAN based on the inference results, a measurement means for measuring the inference speed using the learning model generated by the generation means, and a determination means for determining whether to update the learning model so that the inference speed of the learning model becomes faster based on one or more measured values ​​of the inference speed measured by the measurement means and a reference value of the inference speed.

[0009] According to the present disclosure, it is possible to suppress a decrease in inference speed due to a learning model used to control a RAN.

[0010] Other features and advantages of the present invention will become apparent from the following description taken in conjunction with the accompanying drawings, in which the same or similar elements are designated by the same reference numerals.

[0011] 1 is a control configuration diagram of a RAN according to the background art; 2 is a control configuration diagram of a RAN according to some embodiments; 3 is a configuration diagram of a first control device according to some embodiments; 4 is a configuration diagram of a second control device according to some embodiments; 5 is a sequence diagram according to some embodiments; 6 is a diagram showing an example of measurement results according to some embodiments; 7 is a configuration diagram of a second control device according to some embodiments;

[0012] Hereinafter, the embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the scope of the invention as claimed, and not all combinations of features described in the embodiments are necessarily essential to the invention. Two or more of the features described in the embodiments may be arbitrarily combined. Furthermore, the same reference numerals are used for the same or similar components, and redundant explanations will be omitted.

[0013] <First Embodiment> FIG. 2 shows the control configuration of a RAN according to this embodiment. The configuration of FIG. 2 is basically the same as the configuration of FIG. 1. The first control device 100 performs long-period control of the RAN and is, for example, a device that implements the function of a non-RT RIC of an O-RAN. The second control device 200 is a device that controls the RAN at a shorter period than the first control device 100 and is, for example, a device that implements the function of a near-RT RIC of an O-RAN. The first control device 100 is connected to one or more second control devices 200. Note that in the example of FIG. 2, the first control device 100 is connected to three second control devices 200, but this is merely an example, and the number of second control devices 200 connected to the first control device 100 can be any number equal to or greater than one.

[0014] In this embodiment, the first control device 100 performs machine learning to generate a learning model and distributes the generated learning model to each second control device 200. The first control device 100 generates a learning model corresponding to one or more control contents by each second control device 200 and distributes the learning model to each second control device 200. Each second control device 200 is associated with a geographical area and controls the components of the RAN located in the associated geographical area using one or more learning models acquired from the first control device 100. The second control device 200 controls a portion of the RAN of a mobile communication network that is in a geographical area associated with the second control device 200, but the portion of the RAN of the mobile communication network that is in a geographical area associated with the second control device 200 will also be referred to as the RAN.

[0015] FIG. 3 is a configuration diagram of the first control device 100. The communication unit 14 performs communication processing with the second control device 200 via the A1 interface and communication processing with the second control device 200 and the RAN via the O1 interface. The storage unit 12 stores learning data collected from the RAN via the O1 interface. Note that, in this embodiment, the learning data is collected from the RAN, but the learning data may be collected from the second control device 200. When collecting learning data from the second control device 200, the O1 interface or the A1 interface may be used. The learning unit 11 performs machine learning based on the learning data stored in the storage unit 12 to generate a learning model. The learning unit 11 transmits the generated learning model to each second control device 200 via the A1 interface of the communication unit 14. Note that the O1 interface can also be used to transmit the learning model to each second control device 200. The control unit 13 transmits and receives control messages to each second control device 200 via the communication unit 14. In this embodiment, the control unit 13 uses the O1 interface to transmit and receive control messages to and from each second control device 200, but may use the A1 interface.

[0016] 4 is a configuration diagram of the second control device 200. The inference control unit 21 performs inference based on the learning model received from the first control device 100 via the communication unit 24, and controls the RAN via the E2 interface. The control unit 23 transmits and receives control messages to the first control device 100 via the communication unit 24. For example, when the control unit 23 receives a measurement request from the first control device 100, it causes the measurement unit 22 to measure the inference speed using the learning model generated by the inference control unit 21. When the control unit 23 receives measurement data indicating the measurement result of the inference speed from the measurement unit 22, it transmits the measurement result to the first control device 100 using a control message.

[0017] The measurement unit 22 measures the inference speed based on a measurement instruction from the control unit 23 and outputs measurement data indicating the measurement result to the control unit 23. The measurement unit 23 is configured to measure the inference speed based on, for example, at least one of measurement method A and measurement method B. Measurement method A measures the inference speed when the inference control unit 21 actually controls the RAN. Measurement method B measures the inference speed by having the inference control unit 21 execute inference using test data as input to a learning model. The test data used in measurement method B is, for example, stored in advance in the measurement unit 22. Alternatively, the test data used in measurement method B is, for example, received from the first control unit 100 along with a measurement request. Note that in this embodiment, the first control unit 100 acquires the learning data from the RAN. However, when the first control unit 100 acquires the learning data from the second control unit 200, the second control unit 200 acquires the learning data from the RAN via, for example, the O1 interface or the E2 interface and transmits the learning data to the first control unit 100 via the A1 interface or the O1 interface.

[0018] FIG. 5 is a sequence diagram of a method according to this embodiment. In S1, the control unit 13 of the first control device 100 transmits a measurement request message to the second control device 200, instructing the second control device 200 to measure the inference speed using a learning model. If the second control device 200 uses multiple learning models corresponding to multiple control contents, the measurement request message may include information indicating a "target model," which is the learning model for which the inference speed is to be measured. The number of "target models" may be one or more. If the second control device 200 is capable of executing both measurement method A and measurement method B, the measurement request message may include information specifying the measurement method. If measurement method B is specified, the measurement request message may include test data. Furthermore, the measurement request message may include information indicating the measurement period for the inference speed.

[0019] In S2, the control unit 23 of the second control device 200 causes the measurement unit 22 to measure the inference speed of the target model in accordance with the measurement request message. In S3, the control unit 23 of the second control device 200 transmits the measurement result to the first control device 100.

[0020] FIG. 6 shows an example of measurement results transmitted in S3 when a measurement request for the inference speeds of two learning models, target model #1 and target model #2, is received in S1. The measurement unit 22 repeatedly measures the inference speed during the measurement period specified in the measurement request message. As shown in FIG. 6, the inference speed for target model #1 is measured at times T#11, T#12, T#13, etc. Note that in FIG. 6, the measured inference speeds at times T#11, T#12, and T#13 are 120 ms, 114 ms, and 150 ms. Similarly, as shown in FIG. 6, the inference speed for target model #2 is measured at times T#21, T#22, T#23, etc. Note that in FIG. 6, the measured inference speeds at times T#21, T#22, and T#23 are 40 ms, 48 ​​ms, and 50 ms.

[0021] For example, in the case of measurement method A, the measurement unit 22 measures the inference speed of the inference executed by the inference control unit 21 for each control cycle. In the case of measurement method B, the measurement unit 22 measures the inference speed using test data at a predetermined cycle.

[0022] Furthermore, the measurement results may also include information indicating the reference values ​​of the inference speeds of the target model #1 and the target model #2. According to FIG. 6 , the reference value of the inference speed for the target model #1 is 100 ms, and the reference value of the inference speed for the target model #2 is 50 ms. The reference values ​​of the inference speed may be stored in advance in the control unit 23, for example. Alternatively, the reference value of the inference speed may be dynamically determined by the control unit 23 based on some index, for example. Note that if the learning model is associated with the control content and the reference value of the inference speed determined according to the control content is also set in the first control device 100, it is not necessary to include the reference value of the inference speed in the measurement results.

[0023] In S4, the control unit 13 of the first control device 100 determines, based on the measurement results received in S3, whether or not it is necessary to distribute another learning model with the same control content but a faster inference speed for each of the target models #1 and #2 to the second control device 200. This determination can be made based on a comparison of a predetermined measurement value among one or more measurement values ​​of the inference speed of the target models acquired in S3, or an average value of the one or more measurement values, with a reference value for the inference speed of the target models.

[0024] As one example, the predetermined measurement value among the one or more measurement values ​​is the smallest measurement value among the one or more measurement values. In this case, if the smallest measurement value among the one or more measurement values ​​exceeds a reference value, the control unit 13 may determine that a different learning model with a faster inference speed for the same control content needs to be distributed to the second control device 200. As another example, the predetermined measurement value among the one or more measurement values ​​is the largest measurement value among the one or more measurement values. In this case, the control unit 13 may determine that a different learning model with a faster inference speed for the same control content needs to be distributed to the second control device 200 if the largest measurement value among the one or more measurement values ​​exceeds a reference value.

[0025] As yet another example, the predetermined measurement value among the one or more measurement values ​​is a measurement value at a predetermined rank in descending order of the one or more measurement values, or a measurement value corresponding to a predetermined percentile value among the one or more measurement values. In this case, too, when the predetermined measurement value exceeds the reference value, the control unit 13 may determine that it is necessary to distribute to the second control device 200 another learning model with the same control content but a faster inference speed.

[0026] Furthermore, if a value obtained by performing a predetermined calculation on one or more measurement values, such as the average value of one or more measurement values, rather than a predetermined measurement value among one or more measurement values, exceeds a reference value, the control unit 13 can be configured to determine that a different learning model with the same control content but a faster inference speed needs to be distributed to the second control device 200.

[0027] 6, the reference value for the inference speed of the target model #2 is 50 ms, and the measured values ​​of the inference speed are 40 ms, 48 ​​ms, 50 ms, etc., and only measured values ​​below the reference value exist. For example, when the smallest measured value exceeds the reference value and it is determined that a different learning model with a faster inference speed but the same control content needs to be distributed to the second control device 200, the control unit 13 determines that, for the target model #2, it is not necessary to distribute a different learning model with a faster inference speed but the same control content to the second control device 200.

[0028] On the other hand, according to FIG. 6 , the reference value for the inference speed of the target model #1 is 100 ms, and the measured values ​​of the inference speed are 120 ms, 114 ms, 150 ms, and so on, with only measured values ​​exceeding the reference value. For example, when the minimum measured value exceeds the reference value and it is determined that a different learning model with the same control content but a faster inference speed needs to be distributed to the second control device 200, the control unit 13 determines that, for the target model #1, a different learning model with the same control content but a faster inference speed needs to be distributed to the second control device 200. In this case, the control unit 13 instructs the learning unit 11 to transmit to the second control device 200 a different learning model with the same control content as the target model #1 but a faster inference speed. Based on this instruction, the learning unit 11 transmits the different learning model to the second control device 200 in S5.

[0029] The learning unit 11 may be configured to generate multiple learning models for a single control content in advance. The multiple learning models for a single control content have different inference speeds. Alternatively, the learning unit 11 may generate one learning model for a single control content in advance, and then generate another learning model with a faster inference speed using a compression technique, such as symmetric static quantization. Compressing a learning model using compression technology increases the inference speed of the learning model compared to the original learning model, but generally degrades the inference performance. Therefore, the control unit 13 may be configured to select the learning model with the highest inference performance as the learning model to be transmitted to the second control device 200 in S5, within a range where the average, minimum, and maximum inference speeds do not exceed reference values.

[0030] The above configuration makes it possible to suppress a decrease in the inference speed due to the learning model used to control the RAN.

[0031] Second Embodiment Next, a second embodiment will be described, focusing on differences from the first embodiment. The control configuration of the RAN in this embodiment is as shown in Fig. 2, and the configuration of the first control device 100 is as shown in Fig. 3.

[0032] 7 is a configuration diagram of the second control device 200 according to this embodiment. The second control device 200 of this embodiment has a storage unit 26 that stores learning data, and a learning unit 25. The storage unit 26 stores learning data collected from the RAN via the O1 interface or the E2 interface. The learning unit 25 updates the basic learning model received from the first control device 100 via the communication unit 24 by machine learning based on the learning data to generate a learning model. The inference control unit 21 performs inference based on the learning model learned by the learning unit 25, and controls the RAN via the E2 interface.

[0033] In this embodiment, the control unit 23 causes the measurement unit 22 to measure the inference speed at a predetermined timing even without a measurement request from the first control unit 100. Then, based on the measured inference speed and a reference value, the control unit 23 performs, for example, a compression process on the learning model so that the inference speed satisfies a predetermined condition. The predetermined condition may be the same as the condition used in S4 of FIG. 5 in the first embodiment to determine that a different learning model with a faster inference speed but the same control content does not need to be distributed to the second control unit 200. In other words, the predetermined condition may be, for example, that a predetermined measurement value or an average value of one or more measured values ​​of the measured inference speed does not exceed a reference value. Furthermore, the control unit 23 executes the process of FIG. 5 when triggered by receiving a measurement request from the first control unit 100.

[0034] Unlike the first embodiment, in this embodiment, the learning unit 25 can update the learning model using, for example, a compression process so that the inference speed satisfies the predetermined condition. However, depending on the basic learning model initially received from the first control unit 100, the processing in the learning unit 25 may not be able to update the learning model so that the inference speed satisfies the predetermined condition. In this case, the first control unit 100 can determine that the inference speed of the second control unit 200 does not satisfy the predetermined condition using the sequence in FIG. 5. If the first control unit 100 determines that the inference speed of the second control unit 200 does not satisfy the predetermined condition, in S5, the first control unit 100 transmits to the second control unit 200 another basic learning model with the same control content as the learning model that does not satisfy the predetermined condition but with a faster inference speed.

[0035] The above configuration can suppress a decrease in the inference speed due to the learning model used to control the RAN. In this embodiment, as in the first embodiment, the first control device 100 determines whether or not it is necessary to transmit another basic learning model with a faster inference speed to the second control device 200 based on one or more measured values ​​of the inference speed received from the second control device 200. However, if the second control device 200 determines that the basic learning model received from the first control device 100 cannot be updated to satisfy a predetermined condition, the second control device 200 may be configured to request another basic learning model with a faster inference speed from the first control device 100.

[0036] The first control device 100 and the second control device 200 may be realized by a single device, for example, a single computer, or may be realized by multiple devices, for example, multiple computers, that can communicate with each other.

[0037] The present disclosure further provides a program executable by one or more processors. The program includes instructions that, when executed by one or more processors of an apparatus, cause the apparatus to function as the first control device 100 or the second control device 200. The present disclosure also provides a non-transitory computer-readable storage medium storing the program. The present disclosure also provides a method executed by the first control device 100 or the second control device 200 to suppress a decrease in inference speed due to a learning model used to control the RAN, i.e., the method shown in FIG. 5 , etc. The present disclosure also provides a program for causing an apparatus having one or more processors to execute the method executed by the first control device 100 or the method executed by the second control device 200, and a non-transitory computer-readable storage medium storing the program.

[0038] The invention is not limited to the above-described embodiment, and various modifications and variations are possible within the scope of the gist of the invention.

[0039] This application claims priority based on Japanese Patent Application No. 2023-194467, filed November 15, 2023, the entire contents of which are incorporated herein by reference.

Claims

1. A control device for a radio access network (RAN), comprising: a generation means for generating a learning model by updating a basic learning model received from another control device based on learning data; a first control means for performing inference based on the learning model generated by the generation means and controlling the RAN based on the inference result; a measurement means for measuring the inference speed based on the learning model generated by the generation means; and a judgment means for determining whether to update the learning model so as to make the inference speed of the learning model faster based on one or more measured values ​​of the inference speed measured by the measurement means and a reference value of the inference speed.

2. The control device according to claim 1, wherein, when the determination means determines that the learning model should be updated, the generation means updates the learning model by compressing the learning model so as to increase the inference speed of the learning model.

3. A control device as described in claim 1 or 2, wherein the determination means determines to update the learning model when a specified measurement value among the one or more measurement values ​​exceeds the reference value, or when a value based on the one or more measurement values ​​exceeds the reference value.

4. A control device as claimed in any one of claims 1 to 3, further comprising a second control means for controlling the transmission of the one or more measurement values ​​measured by the measurement means to the other control device in response to a measurement instruction from the other control device.

5. The control device described in claim 4, wherein when the generation means receives another basic learning model from the other control device in response to the measurement means transmitting the one or more measurement values ​​measured by the measurement means to the other control device, the generation means generates a learning model by updating the other basic learning model based on learning data.

6. A control device according to claim 4 or 5, wherein said second control means transmits a reference value of said inference speed together with said one or more measured values ​​to said further control device.

7. The control device according to any one of claims 1 to 6, wherein the other control device is a device that implements a Non-RT RIC.

8. The control device according to any one of claims 1 to 7, wherein the control device is a device that implements a Near-RT RIC.

9. A computer-readable storage medium storing a program that, when executed by one or more processors of an apparatus having one or more processors, causes the apparatus to function as a control device according to any one of claims 1 to 8.

Images