O-ran-based performance optimization configuration method and device

Abstract

The present disclosure relates to an O-RAN-based network system, which selects different transmission modes for users in different scenarios through measurement and prediction of network load, user traffic, service type, mobility, etc., and realizes it through slice strategy, RRC configuration or MAC scheduling strategy, etc., so as to improve performance.

Description

Technical Field

[0001] This disclosure relates to O-RAN network systems, and more specifically, to methods and apparatus for performance optimization in O-RAN systems. Background Technology

[0002] The transformation of 5G (5th Generation) core networks has quietly begun. In the field of radio access networks (RAN), 5G RAN is characterized by numerous services, large bandwidth, and high frequency bands, inevitably leading to smaller single-site coverage, increased equipment complexity, and larger network deployment scale, resulting in huge network costs and increased investment return risks. Considering these specific characteristics and requirements of RAN, the RAN field must introduce a new R&D and design approach that integrates IT (Information Technology), CT (Communication Technology), and DT (Data Technology), which is consistent with the macro-evolutionary trend of the communications industry.

[0003] Based on this, China Mobile, together with operators such as AT&T, integrated the C-RAN Alliance and the xRAN Forum to take the lead in creating the O-RAN (open radio access network) industry alliance, proposing two core visions: "openness" and "intelligence." This aligns with the major development trend of the communications industry and represents another major network transformation led by operators. The aim is to leverage big data, machine learning, and artificial intelligence technologies to build an open and intelligent wireless network, while simultaneously combining open standards, white-box hardware, and open-source software to reduce wireless network costs.

[0004] Public content

[0005] This disclosure addresses at least the aforementioned problems and / or disadvantages and provides at least the following advantages. This disclosure relates to an O-RAN-based network system that, through measurement and prediction of network load, user traffic volume, service type, mobility, etc., selects different transmission methods for users in different scenarios, and specifically implements these methods through slicing strategies, RRC configuration, or MAC scheduling strategies, thereby improving base station system throughput, enhancing user experience, and helping terminals save power, among other performance improvements.

[0006] According to one aspect of this disclosure, a method for performance optimization in an O-RAN system includes: creating at least two network subslice instances (NSSIs); requesting information reporting through an interface between a management entity and a network function or within a network function; receiving the reported information and generating policy information based on the reported information; and distributing the policy information through the corresponding interface.

[0007] According to one aspect of this disclosure, different NSSIs correspond to corresponding selected transport modes, and wherein the corresponding selected transport mode is indicated by one of the following: a newly added separate information element (IE) in the NSSI; an NSSI name string rule; and an implicit association between the NSSI and the corresponding selected transport mode.

[0008] According to one aspect of this disclosure, receiving reported information includes at least one of the following: receiving user capability information through a corresponding interface; receiving cell capability information through a corresponding interface; and receiving slice measurement information through a corresponding interface.

[0009] According to one aspect of this disclosure, the method further includes receiving user scenario-related information from an external server, wherein the user scenario-related information includes at least one of the following: user mobility-related information, used for mobility estimation and prediction of each user in each cell of the slice; and user service traffic-related information, used for service traffic estimation and prediction of each user in each cell of the slice.

[0010] According to one aspect of this disclosure, receiving reported information and generating strategy information based on the reported information further includes: generating strategy information based on collected information related to user scenarios and using an artificial intelligence / machine learning (AI / ML) module.

[0011] According to one aspect of this disclosure, the policy information includes at least one of the following: user mobility-related information; user service traffic-related information; slice-level configuration information; user-level configuration information; and association information between slices and users.

[0012] According to one aspect of this disclosure, the method further includes: determining and sending RRC configuration parameters based on the generated policy information.

[0013] According to one aspect of this disclosure, the intelligent controller module for decision-making is deployed in one module or in multiple modules connected via an interface.

[0014] According to one aspect of this disclosure, based on the deployment method of the intelligent control module used for decision-making and the generated strategy information, corresponding information is added to the corresponding interface and the interface for receiving the strategy information is determined.

[0015] According to one aspect of this disclosure, a method for performance optimization in an O-RAN system includes: requesting information reporting through an interface between a management entity and a network function or within a network function; receiving the reported information and generating prediction information based on the reported information; and distributing the prediction information through a corresponding interface, wherein the prediction information is used by a base station to determine the user's transmission mode.

[0016] According to one aspect of this disclosure, receiving reported information includes at least one of the following: receiving user capability information through a corresponding interface; receiving cell capability information through a corresponding interface; and receiving slice measurement information through a corresponding interface.

[0017] According to one aspect of this disclosure, the method further includes receiving user scenario-related information from an external server, wherein the user scenario-related information includes at least one of the following: user mobility-related information, used for mobility estimation and prediction of each user in each cell of the slice; and user service traffic-related information, used for service traffic estimation and prediction of each user in each cell of the slice.

[0018] According to one aspect of this disclosure, the prediction information includes at least one of the following: user mobility prediction information; and user service traffic prediction information.

[0019] According to one aspect of this disclosure, receiving reported information and generating predictive information based on the reported information further includes: generating predictive information using an artificial intelligence / machine learning (AI / ML) module based on collected information related to the user scenario.

[0020] According to one aspect of this disclosure, the method further includes the base station determining and sending RRC configuration parameters based on the transmitted prediction information.

[0021] According to one aspect of this disclosure, the intelligent controller module for prediction is deployed in one module or in multiple modules connected via an interface.

[0022] According to one aspect of this disclosure, based on the deployment method of the intelligent control module used for prediction and the generated prediction information, corresponding information is added to the corresponding interface and the interface for receiving the prediction information is determined.

[0023] According to one aspect of this disclosure, receiving reported information and generating prediction information based on the reported information also includes receiving prediction feedback from the base station to update the prediction information.

[0024] According to one aspect of this disclosure, an apparatus for performance optimization in an O-RAN system includes: a network management entity that requests information reporting through an interface between the management entity and network functions or between network functions; a RAN intelligent controller, including a non-real-time RAN intelligent controller (Non-RT RIC) and a near-real-time intelligent controller (Near-RT RIC), configured to generate policy information or prediction information based on the reported information, and to distribute the policy information or prediction information through corresponding interfaces; and a base station configured to execute the distributed policy information or determine the transmission mode for users based on the distributed prediction information.

[0025] According to one aspect of this disclosure, an apparatus for performance optimization in an O-RAN system includes a memory and a processor, wherein the memory stores instructions that, when executed by the processor, implement the aforementioned method. Attached Figure Description

[0026] The above and other objects, features and advantages of this disclosure will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0027] Figure 1 An overall O-RAN framework according to an embodiment of the present disclosure is shown;

[0028] Figure 2 A flowchart illustrating an example of O-RAN slice physical resource allocation is shown;

[0029] Figure 3 This paper illustrates system performance optimization in a slice-based user mobility scenario according to embodiments of the present disclosure;

[0030] Figure 4 This paper illustrates system performance optimization in a slice-based user service prediction scenario according to embodiments of the present disclosure;

[0031] Figure 5 This illustrates system performance optimization based on slices in user mobility and business scenarios according to embodiments of the present disclosure;

[0032] Figure 6 Slice-based configuration parameter optimization according to embodiments of the present disclosure is illustrated;

[0033] Figure 7 The present disclosure illustrates system performance optimization in a predicted user mobility scenario according to an embodiment of the present disclosure;

[0034] Figure 8 System performance optimization based on prediction of business scenarios according to embodiments of the present disclosure is illustrated.

[0035] Figure 9 Prediction-based configuration parameter optimization according to embodiments of the present disclosure is illustrated;

[0036] Figure 10 A general flowchart illustrating a slice-based performance optimization implementation for RAN intelligent controller decision-making according to embodiments of the present disclosure is shown; and

[0037] Figure 11 A general flowchart is shown of a predictive, assisted RAN node performance optimization implementation method according to an embodiment of the present disclosure. Detailed Implementation

[0038] Embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the disclosure. Although certain embodiments and examples have been provided, it will be apparent to those skilled in the art, based on the content disclosed herein, that changes can be made to the illustrated embodiments and examples without departing from the scope of the disclosure. Those skilled in the art will understand that the principles of the present disclosure can be implemented in any suitably arranged system or apparatus.

[0039] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0040] Figure 1 An overall O-RAN framework according to an embodiment of this disclosure is shown. The design principles of the O-RAN reference architecture are based on the wireless network CU / DU (Centralized Unit / Distribute Unit) architecture and functional virtualization, introducing open interfaces and open hardware reference designs, while leveraging artificial intelligence to optimize the wireless control process. The following will refer to... Figure 1 Describe it.

[0041] 101 Service Management and Orchestration (SMO) is an entity that provides a variety of management services and network management functions.

[0042] 101-1 indicates a Non-Real Time RAN Intelligent Controller (Non-RT RIC) with functions such as microservice and policy management, wireless network analysis, and training of artificial intelligence models. The trained AI model is distributed to the near real-time RAN Intelligent Controller for online inference and execution through the A1 interface.

[0043] 102 indicates O-RAN network functions. Compared to non-O-RAN systems, it introduces a near-real-time RAN Intelligent Controller (Near-RT RIC), which includes entities such as the O-RAN Control Unit (O-CU), O-RAN Data Unit (O-DU), and O-RAN Ratio Unit (O-RU).

[0044] The network function part of O-RAN can be a gNB that supports the 5G protocol, or an eNB that supports the 4G (4th-Generation) and LTE (Long Term Evolution) protocols.

[0045] A significant difference between 5G and 4G protocols is that 5G supports network slicing, while 4G does not. Therefore, gNBs generally support network slicing, while existing eNBs do not.

[0046] 102-1 indicates Near-RT RIC, a near real-time wireless network intelligent controller component in the O-RAN architecture that is embedded within the CU. It can be understood as an enhanced functional entity for next-generation radio resource management (RRM) with embedded artificial intelligence technology.

[0047] 102-2 indicates that the O-CU, compared to the CU of the non-O-RAN system, adds support for the E2 interface.

[0048] 102-3 indicates that O-DUs, compared to DUs in non-O-RAN systems, add support for the E2 interface.

[0049] 103 indicates O-Cloud, which supports cloud systems with orchestratable tasks.

[0050] 104 indicates NG-core, which is the 5G core network.

[0051] 105 indicates that external systems, such as servers of various applications, can provide rich data to the SMO.

[0052] The O1 interface is used to connect SMO and O-RAN network functional entities.

[0053] The O2 interface is used to connect the SMO and the O-RAN cloud (O-cloud).

[0054] The A1 interface is used to embed non-real-time wireless intelligent controllers into network management functions, while near-real-time controllers are embedded in evolved Node B (eNB) / Next Generation Node B (gNB) base stations. Due to the introduction of artificial intelligence, the A1 management interface between network management and wireless network elements goes beyond the traditional network management's Fault, Configuration, Accounting, Performance and Security (FCAPS) functions, extending to new data information such as the distribution of base station operation policies and AI machine learning models.

[0055] The E2 interface is a standard interface between the near-RT RIC and the CU / DU protocol stack software. Analogous to the interface between Radio Resource Management (RRM) and Radio Resource Control (RRC) in traditional equipment, the near-RT RIC not only collects measurement information from various functional entities of the radio network through the E2 interface, but also sends control commands to the base station through this interface, ultimately controlling the base station's behavior. In an open software architecture, the standardization of the E2 interface enables the near-RT RIC functional software to evolve independently of traditional base station software versions, shortening the time-to-market for software features.

[0056] 5G supports network slicing. Existing O-RAN solutions use non-RT RIC to assess network congestion and support the allocation of physical resources for network slicing.

[0057] Figure 2 A flowchart illustrating an example of O-RAN slice physical resource allocation is shown. The following will refer to... Figure 2 Describe it.

[0058] Step 201 is triggered by external system 105. The external system is an external system that can provide rich data for SMO, such as an application server.

[0059] Step 202: The SMO assists the 101-1 Non-RT RIC in collecting regular internal and external data, which the Non-RT RIC then uses to assess slice congestion. This slice congestion assessment requires utilizing the Artificial Intelligence / Machine Learning (AI / ML) modules within the Non-RT RIC, based on external data. This external data may include, for example, messages related to natural disasters or other events on application servers that could cause network congestion.

[0060] Step 203: Based on the further congestion assessment requirements of the Non-RT RIC, the SMO completes further traffic-related data collection. These further congestion assessment requirements include assessments of congestion levels, etc. The traffic-related data can originate from measurement reports via the O1 interface or from external systems.

[0061] Step 204: Based on the collected relevant data, the Non-RT RIC completes the slice QoS (Quality of Service) measurement adjustment and configuration for the O-RAN network function entity based on AI / ML training. The adjustment includes adjusting the number of slice physical resource blocks (PRBs).

[0062] Step 205: Non-RT RIC completes congestion assessment.

[0063] If the congestion is relieved, then in step 206, the relevant congestion handling strategy is deleted.

[0064] The following will be referenced Figures 3 to 9 This describes system performance optimization based on ORAN according to embodiments of the present disclosure.

[0065] Massive Multiple Input Multiple Output (MIMO) technology is a key technology for 5G and 4.5G (4.5th Generation). In Time Division Duplexing (TDD) Massive MIMO systems, uplink channel information is measured using the Sounding Reference Signal (SRS), and downlink channel information is obtained through channel reciprocity. Alternatively, downlink channel information can be measured using the Channel State Information Reference Signal (CSI-RS), and users provide feedback on Channel State Information (CSI), including Precoding Matrix Index (PMI), Rank Indicator (RI), and Channel Quality Indicator (CQI). In a Frequency Division Duplexing (FDD) Massive MIMO system, since the uplink and downlink channels are not on the same frequency, the base station cannot obtain downlink channel information through SRS using channel reciprocity. However, it can obtain downlink channel information through user feedback by configuring CSI-RS.

[0066] After the base station obtains channel information, the gNB or eNB transmits signals through beamforming. Based on whether the signal transmitted by a base station on the same time-frequency resources belongs to one user or multiple users, Massive MIMO transmission methods are divided into Multi-User Multiple Input Multiple Output (MU-MIMO) and Single-User Multiple Input Multiple Output (SU-MIMO). Based on different methods of acquiring channel state information, the transmission methods can be further subdivided into SRS-based MU-MIMO, PMI-based MU-MIMO, SRS-based SU-MIMO, and PMI-based SU-MIMO. For PMI-based MU-MIMO, it can be further distinguished according to the type of codebook used for feedback, such as Type I codebook-based MU-MIMO, Type II codebook-based MU-MIMO, Type I codebook-based SU-MIMO, and Type II codebook-based SU-MIMO. If multiple stations are allowed to coordinate data transmission to users, the transmission method can also include Coordinated Multiple Point (CoMP) and other methods. In summary, the base station can select the specific transmission method for the user as needed.

[0067] In practical deployments, different transmission methods exhibit varying spectral efficiencies across different application scenarios. For instance, in real-world systems, downlink Massive MIMO systems can support up to 8 or 16 layers of multi-user data transmission. SRS-based MU-MIMO achieves the highest spectral efficiency under conditions of high SRS channel quality, stationary users, and large data packets. However, when SRS channel quality deteriorates, users move, or users only have small data packets, the transmission efficiency of SRS-based MU-MIMO can significantly decrease, potentially even falling below that of SU-MIMO.

[0068] In practical systems, base stations strive to select appropriate transmission modes for users. However, in existing non-O-RAN systems, while base stations can measure user CSI and Reference Signal Received Power (RSRP), they struggle to obtain information about user movement speed and predict user movement patterns. Similarly, while they can obtain current buffer occupancy (BO) information, they find it difficult to predict subsequent service usage. Furthermore, while they can obtain CSI feedback, they struggle to predict subsequent channel quality changes. Typically, configuring a user to use a transmission mode requires configuration via Radio Resource Control (RRC) signaling: for example, SRS-based MU-MIMO transmission requires configuring SRS parameters. Once a transmission mode is configured, it will be used for several subsequent Transport Time Intervals (TTIs). In other words, the transmission mode determined by the base station applies over several time intervals, not just the current moment. Therefore, determining the user's transmission mode based on predicted user and scenario information would be more appropriate.

[0069] In addition, O-RAN's Non-RT RIC can recommend optimized configuration parameters for the transmission mode of each base station based on information such as the number of users reported by each cell.

[0070] The SMO in O-RAN obtains rich user data from the application server. Based on this data, AI / ML in Non-RT RIC or Near-RT RIC can be used to predict user mobility, channel quality, and traffic volume. Using this information to help make decisions about user transmission methods will greatly improve the rationality of the decisions and the efficiency of system transmission.

[0071] On the one hand, 5G defines three major application scenarios: Enhanced Mobile Broadband (eMBB), Ultra Reliable Low Latency Communications (uRLLC), and Massive Machine Type Communications (mMTC). For example, eMBB application scenarios mainly refer to high-bandwidth consumer applications of mobile internet such as 4K / 8K high-definition video, augmented reality / virtual reality (AR / VR), and 3D holography.

[0072] In the 5G protocol definition, network slicing technology supports the coexistence of three major application scenarios: a single network entity can support multiple slices, each slice targeting a specific application scenario. Simultaneously, 5G networks also support customized slices, allowing services and implementations to define their own dedicated slices.

[0073] Meanwhile, by utilizing the slicing function, O-RAN can place users suitable for different transmission methods into different sub-slices, while informing the base station of the appropriate transmission method for each sub-slice. In this way, on the one hand, the transmission method used by the base station for users will be more appropriate; on the other hand, different slices use independent resources, further simplifying the base station's scheduling and transmission.

[0074] On the other hand, since 4G or 4.5G eNBs or gNBs do not support slicing and / or sub-slicing functions, information such as user attributes, service conditions, and mobility status can be added to the O1 or E2 interface and provided to the eNB. The eNB can then use this richer user and scenario information to select appropriate transmission methods and configuration parameters for users, which will greatly improve the rationality of decision-making and the efficiency of system transmission.

[0075] To address the aforementioned issues, this disclosure proposes a system performance optimization scheme based on O-RAN. By acquiring rich data provided by external systems, such as global positioning system (GPS) information provided by the application's server and business information collected by the service server, and combining it with internal monitoring information reported by the RAN node, the AI ​​module (AI / ML Function) in the Non-RT RIC or Near-RT RIC is used for prediction or decision processing. The prediction or decision results are then sent to the relevant functional modules in the RAN node, thereby achieving goals such as optimizing user transmission mode selection or configuration parameters (e.g., discontinuous reception (DRX) configuration).

[0076] Based on the different ways in which RICs (here, Non-RT RICs and Near-RT RICs are collectively referred to as RICs) process data, the embodiments disclosed in this disclosure can be divided into two categories: one where the RIC directly participates in decision-making and utilizes slicing, denoted as category A; and the other where the RIC performs predictions to assist modules in the RAN nodes in optimization processing, denoted as category B. According to different data sources and data classifications, they can be divided into mobility scenarios, service scenarios, and hybrid mobility and service scenarios. Therefore, based on classification and scenarios, this disclosure provides the following embodiments for illustration:

[0077] Example A-1: ​​System performance optimization in user mobility scenarios based on slicing;

[0078] Example A-2: System performance optimization in a business scenario based on slicing;

[0079] Example A-3: System performance optimization based on slice-based mobility and business scenarios;

[0080] Example A-4: Slice-based configuration parameter optimization;

[0081] Example B-1: System performance optimization based on predicted user mobility scenarios;

[0082] Example B-2: System performance optimization based on predictive business scenarios; and

[0083] Example B-3: Prediction-based configuration parameter optimization.

[0084] Those skilled in the art will understand that this disclosure is not limited to the above embodiments, and any case that uses or combines the innovative points of this disclosure is within the protection scope of this disclosure.

[0085] Example A-1

[0086] Figure 3 This paper illustrates system performance optimization in a slice-based user mobility scenario according to embodiments of the present disclosure. Reference will be made below. Figure 3 Describe it.

[0087] Step 301: Enable the slice optimization function during O-RAN slice operation. Unlike existing technologies, this disclosure is not triggered by congestion events; it can be configuration-triggered or run periodically. For example, the slice resource optimization function can be enabled based on the needs of the slice service provider or slice customer, or the slice optimization function can be continuously enabled within an O-RAN system.

[0088] Compared to existing congestion-triggered slice resource allocation methods, the slice resource optimization method disclosed herein can be performed periodically. For Non-RT RIC, it can support policy prediction at the second level, which is more dynamic than existing technologies. Specifically, existing technologies are triggered by congestion events, the event scale of which may be several days, several months, or several hours. However, the embodiments in this disclosure can be at the second level using Non-RT RIC and at the millisecond level using Near-RT RIC.

[0089] Step 302: Create a Network Slice Subnet Instance (NSSI) using the Network Slice Subnet Management Function (NSSMF).

[0090] Specifically, Single Network Slice Selection Assistance Information (S-NSSAI) defines an end-to-end network slicing, as detailed in 3GPP TS 23.003.

[0091] S-NSSAI is a 32-bit identifier containing 8 bits of slice / service type (SST) and 24 bits of slice differentiator (SD).

[0092] A collection of multiple S-NSSAIs constitutes Network Slice Selection Assistance Information (NSSAI). To support the implementation of NSSAI, the network deploys one or more Network Slice Instances (NSIs) and one or more NSSIs. An NSSI is all or part of an NSI. An NSSI can belong to one or more NSIs. An NSI can consist of one or more NSSIs.

[0093] This disclosure proposes that at least two NSSIs need to be created in an O-RAN system. In the network implementation, different NSSIs correspond to different preferred transmission modes, such as MU-MIMO or SU-MIMO.

[0094] The preferred transmission method described herein does not mean that all users in this NSSI only support this one transmission method and do not support other transmission methods.

[0095] For example, in NSSI where SRS-based MU-MIMO is the preferred transmission method, when a user's SRS information is incomplete, it may temporarily fall back to using SU-MIMO transmission.

[0096] The sub-slice management module creates at least two sub-slice instances on the RAN side, each supporting different transmission methods. Sub-slices can belong to the same slice instance or different slice instances, but all belong to the same S-NSSAI. Users do not need to update the S-NSSAI during sub-slice instance switching, thus reducing signaling overhead and enabling seamless soft handover, minimizing the impact on service transmission.

[0097] In existing technologies, there is no design for dividing sub-slices based on preferred transmission methods; therefore, information related to transmission methods is not included in the sub-slice attributes. In this disclosure, information representing the preferred transmission method for NSSI is added to the NSSI information interface of the O-RAN system.

[0098] Specifically, a separate Information Element (IE) can be added to the NSSI to identify the preferred transmission mode of this NSSI. Alternatively, the preferred transmission mode of this NSSI can be represented by the rules of the NSSI naming string, or it can be implicitly associated with the preferred transmission mode of the NSSI, such as the range of user movement speeds supported by the NSSI.

[0099] The advantages of using transmission mode-based sub-slices in the embodiments of this disclosure are that they physically separate users of the two transmission modes, eliminate mutual interference between different transmission modes, facilitate separate management and measurement of sub-slices for different transmission modes, and reduce the processing complexity of the base station. At the same time, notifying the base station of the preferred transmission mode through sub-slices can also save signaling overhead.

[0100] Step 303, Data Collection. The SMO module collects user mobility-related information (such as spatial coordinates, GPS, orientation information relative to the base station, surrounding environment distribution information, maps, etc.) from the application server. This information is used to estimate and predict the mobility of each user in each cell of the slice.

[0101] The SMO obtains user capability information, such as SRS reporting capabilities, supported demodulation reference signal (DMRS) types, and supported DMRS symbol count, through the network functions via the O1 interface, the Non-RT RIC via the A1 interface, or the Near-RT RIC via the E2 interface. This information is used to determine the set of supportable transmission modes for each user in each cell of the slice.

[0102] SMO obtains cell capability information, such as whether it supports MU-MIMO transmission, whether it supports SU-MIMO transmission, the maximum number of MU-MIMO transmission layers, the maximum number of SU-MIMO transmission layers, and whether it supports CoMP transmission, through the O1 interface from the network function, or through the A1 interface from the Non-RT RIC, or through the E2 interface from the Near-RT RIC. This information is used to determine the set of transmission modes that can be supported in each cell of the slice.

[0103] The SMO obtains slice-related measurement information, such as the number of online users, throughput, and physical resource block (PRB) utilization, from the network function via the O1 interface, the Non-RT RIC via the A1 interface, or the Near-RT RIC via the E2 interface. This information is used to determine the overall load of the slice.

[0104] After data collection, it is aggregated in the RIC module for AI processing.

[0105] In the case where the RIC module is a Non-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Near RT-RIC transmits the information collected by the SMO to the Non-RT RIC module through the A1 interface.

[0106] In the case where the RIC module is a Near-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Non-RT RIC then transmits the information to the Near-RT RIC module through the A1 interface. The network function transmits information to the Near-RT RIC module through the E2 interface.

[0107] In the existing technology, there are no use cases that specify the need to collect information related to the transmission mode decision, and there is no consideration of using the ability of RIC to estimate and measure user mobility to determine the transmission mode.

[0108] The advantages of using service information in the embodiments of this disclosure are that it allows for understanding the user's service model, predicting the timing of service occurrences, preparing resources for the user in advance, reducing processing latency, improving user experience, and meeting SLA requirements. The advantages of using user capability information are that it allows high-performance terminals to use MU-MIMO and low-performance terminals to use SU-MIMO, reducing unnecessary retransmissions and improving resource utilization. Different levels of terminals have different downlink data transmission capabilities. According to the 3GPP protocol, a terminal can support at least one layer of downlink data stream and at most eight layers of downlink data streams. High-performance terminals can eliminate interference from multiplexed users when demodulating downlink data, thus demodulating the original data.

[0109] Step 304: Determine the user sub-slice strategy. The RIC (Non-RT RIC and / or Near-RT RIC) selects the preferred transmission method for each user in each cell of the slice by running an AI module based on the collected data, and places the user into the corresponding sub-slice.

[0110] Specifically, this decision-making process can be completed by the Non-RT RIC. The decision is passed to the Near-RT RIC via the A1 interface, and then passed to the network function by the Near-RT RIC via the E2 interface.

[0111] Specifically, this decision-making process can be completed by the Non-RT RIC. The decision is shared with the SMO via the SMO's internal bus, and then the SMO passes it to the network function through the O1 interface.

[0112] Specifically, the Near-RT RIC can run an AI module to complete this decision-making process, and the decision is passed to the network function through the E2 interface.

[0113] Specifically, the decision-making process can be completed by running the AI ​​module separately by the Non-RT RIC and the Near-RT RIC. The Non-RT RIC transmits the decision and the completed model training to the Near-RT RIC through the A1 interface. The Near-RT RIC further runs the AI ​​module based on the information transmitted by the Non-RT RIC and the information transmitted through the E2 interface, completes the final decision, and transmits it to the network function through the E2 interface.

[0114] Specifically, the user transmission method after the decision is passed to the network function through sub-slices. For example, the user IDs belonging to each sub-slice are listed, or each user is labeled with the sub-slice ID to which they belong.

[0115] In existing technologies, there is no interface message for transmitting user-preferred transmission mode information. This disclosure introduces the above-mentioned interface message transmission to implement the sub-slice preferred transmission mode selection function.

[0116] Specifically, for the implementation of the Non-RT RIC running AI module to determine the preferred transmission method, information related to the preferred transmission method needs to be added to the A1 and E2 interfaces.

[0117] Specifically, to determine the preferred transmission method for the Near-RT RIC running AI module, information related to the preferred transmission method needs to be added to the E2 interface.

[0118] Specifically, for the implementation method where the AI ​​modules of Non-RT RIC and Near-RT RIC respectively complete this decision-making, it is necessary to add Non-RT RIC training results and model deployment information to the A1 interface, and to add information related to the preferred transmission method to the E2 interface. Specifically, the Non-RT RIC training results added to the A1 interface can be the preferred transmission method information for each user, the movement speed level information for each user, or the movement model information for each user, etc.

[0119] Step 305: Receive and apply the user sub-slice policy. After receiving the user's sub-slice information through the O1 interface or E2, the network function deletes the user information in the original sub-slice instance and creates the user information in the indicated sub-slice instance.

[0120] Specifically, whether to use the O1 interface or the E2 interface to receive policy information depends on the method used in step 304. For the Non-RT RIC decision-making method, the network function can receive the policy information through either the O1 interface or the E2 interface. For the Near-RT RIC decision-making method, or a joint Non-RT RIC and Near-RT RIC decision-making method, the network function can receive the policy information through the E2 interface.

[0121] After receiving the policy, the network function will complete the user's RRC configuration or Media Access Control (MAC) scheduling according to the sub-slice policy.

[0122] The embodiments of this disclosure are based on sub-slice instance operations, which can separate transmission methods from physical resources and eliminate mutual interference between different transmission methods. User operations during sub-slice switching are described.

[0123] To support the above process, other auxiliary steps can be added.

[0124] For example, a network function module can inform the SMO via the O1 interface whether it or the cell supports sub-slice functionality based on preferred transmission mode. This process is only performed on network function modules or cells that support this functionality. Figure 3The process in the middle.

[0125] For example, the network function module can further report some measurement data to help the RIC evaluate the effectiveness of the preferred transmission method strategy, thereby helping the RIC optimize the algorithm or terminate the sub-slice function based on the preferred transmission method.

[0126] According to the above embodiments, by measuring mobility, different transmission methods are selected for users in different scenarios, and this is specifically implemented through slicing strategies, RRC configuration, or MAC scheduling strategies, thereby improving the throughput of the base station system, enhancing user experience, and helping terminals save power, among other performance improvements.

[0127] Example A-2

[0128] Figure 4 This paper illustrates system performance optimization in a slice-based user traffic prediction scenario according to embodiments of the present disclosure. Reference will be made below. Figure 4 Describe it.

[0129] Step 401 is enabled during O-RAN slice operation, and is the same as step 301.

[0130] Step 402 is the same as step 302.

[0131] Step 403, Data Collection. The SMO module collects user traffic prediction information from the application server (e.g., the mobile application being used, whether the user is in the office or driving). This information is used to estimate and predict traffic for each user in each cell of the slice.

[0132] The SMO obtains user capability information, such as SRS reporting capability, supported DMRS types, and supported DMRS symbol count, from the network functions via the O1 interface, or from the Non-RT RIC via the A1 interface, or from the Near-RT RIC via the E2 interface. This information is used to determine the set of supported transmission modes for each user in each cell of the slice.

[0133] SMO obtains cell capability information, such as whether it supports MU-MIMO transmission, whether it supports SU-MIMO transmission, the maximum number of MU-MIMO transmission layers, the maximum number of SU-MIMO transmission layers, and whether it supports CoMP transmission, through the O1 interface from the network function, or through the A1 interface from the Non-RT RIC, or through the E2 interface from the Near-RT RIC. This information is used to determine the set of transmission modes that can be supported in each cell of the slice.

[0134] The SMO obtains slice-related measurement information, such as the number of online users, throughput, and PRB utilization, from the network function via the O1 interface, or from the Non-RT RIC via the A1 interface, or from the Near-RT RIC via the E2 interface. This information is used to determine the overall load of the slice.

[0135] After data collection, it is aggregated in the RIC module for AI processing.

[0136] In the case where the RIC module is a Non-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Near RT-RIC transmits the information collected by the SMO to the Non-RT RIC module through the A1 interface.

[0137] In the case where the RIC module is a Near-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Non-RT RIC then transmits the information to the Near-RT RIC module through the A1 interface. The network function transmits information to the Near-RT RIC module through the E2 interface.

[0138] In the existing technology, there are no use cases that specify the need to collect information related to transmission mode decisions, and there is no consideration of using the ability of RIC to estimate and measure user traffic to determine the transmission mode.

[0139] The advantages of using service information in the embodiments of this disclosure are that it allows for understanding the user's service model, predicting the timing of service occurrences, preparing resources for the user in advance, reducing processing latency, improving user experience, and meeting SLA requirements. The advantages of using user capability information are that it allows high-performance terminals to use MU-MIMO and low-performance terminals to use SU-MIMO, reducing unnecessary retransmissions and improving resource utilization. Different levels of terminals have different downlink data transmission capabilities. According to the 3GPP protocol, a terminal can support at least one layer of downlink data stream and at most eight layers of downlink data streams. High-performance terminals can eliminate interference from multiplexed users when demodulating downlink data, thus demodulating the original data.

[0140] Step 404 is the same as step 304.

[0141] Step 405 is the same as step 305.

[0142] To support the above process, other auxiliary steps can be added.

[0143] For example, a network function module can inform the SMO via the O1 interface whether it or the cell supports sub-slice functionality based on preferred transmission mode. This process is only performed on network function modules or cells that support this functionality. Figure 4 The process in the middle.

[0144] For example, the network function module can further report some measurement data to help the RIC evaluate the effectiveness of the preferred transmission method strategy, thereby helping the RIC optimize the algorithm or terminate the sub-slice function based on the preferred transmission method.

[0145] According to the above embodiments, by measuring the traffic of different service types, different transmission methods are selected for users in different scenarios, and this is specifically implemented through slicing strategies, RRC configuration, or MAC scheduling strategies, thereby improving the throughput of the base station system, enhancing user experience, and helping terminals save power, among other performance improvements.

[0146] Example A-3

[0147] Figure 5 This paper illustrates system performance optimization based on slice-based user mobility and service scenarios according to embodiments of the present disclosure. References will be made below. Figure 5 Describe it.

[0148] Step 501 is enabled during O-RAN slice operation, the same as step 301.

[0149] Step 502 is the same as step 302.

[0150] Step 503, Data Collection. The SMO module collects the following from the application server: user mobility-related information (e.g., spatial coordinates, GPS, orientation information relative to the base station, surrounding environment distribution information, maps, etc.), which is used to estimate and predict the mobility of each user in each cell of the slice; and user traffic prediction-related information (e.g., the mobile application being used, whether the user is in the office or driving, etc.), which is also used to estimate and predict the traffic of each user in each cell of the slice.

[0151] The SMO obtains user capability information, such as SRS reporting capability, supported DMRS types, and supported DMRS symbol count, from the network functions via the O1 interface, or from the Non-RT RIC via the A1 interface, or from the Near-RT RIC via the E2 interface. This information is used to determine the set of supported transmission modes for each user in each cell of the slice.

[0152] SMO obtains cell capability information, such as whether it supports MU-MIMO transmission, whether it supports SU-MIMO transmission, the maximum number of MU-MIMO transmission layers, the maximum number of SU-MIMO transmission layers, and whether it supports CoMP transmission, through the O1 interface from the network function, or through the A1 interface from the Non-RT RIC, or through the E2 interface from the Near-RT RIC. This information is used to determine the set of transmission modes that can be supported in each cell of the slice.

[0153] The SMO obtains slice-related measurement information, such as the number of online users, throughput, and PRB utilization, from the network function via the O1 interface, or from the Non-RT RIC via the A1 interface, or from the Near-RT RIC via the E2 interface. This information is used to determine the overall load of the slice.

[0154] After data collection, it is aggregated in the RIC module for AI processing.

[0155] In the case where the RIC module is a Non-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Near RT-RIC transmits the information collected by the SMO to the Non-RT RIC module through the A1 interface.

[0156] In the case where the RIC module is a Near-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Non-RT RIC then transmits the information to the Near-RT RIC module through the A1 interface. The network function transmits information to the Near-RT RIC module through the E2 interface.

[0157] In the existing technology, there are no use cases that specify the need to collect information related to transmission mode decisions, and there is no consideration of using the ability of RIC to estimate and measure user mobility and traffic to determine the transmission mode.

[0158] The advantages of using service information in the embodiments of this disclosure are that it allows for understanding the user's service model, predicting the timing of service occurrences, preparing resources for the user in advance, reducing processing latency, improving user experience, and meeting SLA requirements. The advantages of using user capability information are that it allows high-performance terminals to use MU-MIMO and low-performance terminals to use SU-MIMO, reducing unnecessary retransmissions and improving resource utilization. Different levels of terminals have different downlink data transmission capabilities. According to the 3GPP protocol, a terminal can support at least one layer of downlink data stream and at most eight layers of downlink data streams. High-performance terminals can eliminate interference from multiplexed users when demodulating downlink data, thus demodulating the original data.

[0159] Step 504 is the same as step 304.

[0160] Step 505 is the same as step 305.

[0161] To support the above process, other auxiliary steps can be added.

[0162] For example, a network function module can inform the SMO via the O1 interface whether the network function module or the cell supports sub-slice functionality based on preferred transmission mode. This process is only performed on network function modules or cells that support this functionality. Figure 4 The process in the middle.

[0163] For example, the network function module can further report some measurement data to help the RIC evaluate the effectiveness of the preferred transmission method strategy, thereby helping the RIC optimize the algorithm or terminate the sub-slice function based on the preferred transmission method.

[0164] According to the above embodiments, by measuring and predicting user traffic volume, service type, mobility, etc., different transmission methods are selected for users in different scenarios, and this is specifically implemented through slicing strategies, RRC configuration, or MAC scheduling strategies, thereby improving the throughput of the base station system, enhancing user experience, and helping terminals save power and other performance improvements.

[0165] Example A-4

[0166] Figure 6 Slice-based configuration parameter optimization according to embodiments of the present disclosure is illustrated. Reference will be made below. Figure 6 Describe it.

[0167] Step 601 is enabled during O-RAN slice operation, the same as step 301.

[0168] Step 602 is the same as step 302.

[0169] Step 603, Data Collection. The SMO module collects user scenario-related information from the application server, such as mobility prediction, traffic prediction, and network congestion prediction.

[0170] The SMO obtains user capability information, such as SRS reporting capability, supported DMRS types, and supported DMRS symbol count, from the network functions via the O1 interface, or from the Non-RT RIC via the A1 interface, or from the Near-RT RIC via the E2 interface. This information is used to determine the set of supported transmission modes for each user in each cell of the slice.

[0171] SMO obtains cell capability information, such as whether it supports MU-MIMO transmission, whether it supports SU-MIMO transmission, the maximum number of MU-MIMO transmission layers, the maximum number of SU-MIMO transmission layers, and whether it supports CoMP transmission, through the O1 interface from the network function, or through the A1 interface from the Non-RT RIC, or through the E2 interface from the Near-RT RIC. This information is used to determine the set of transmission modes that can be supported in each cell of the slice.

[0172] The SMO obtains slice-related measurement information, such as the number of online users, throughput, and PRB utilization, from the network function via the O1 interface, or from the Non-RT RIC via the A1 interface, or from the Near-RT RIC via the E2 interface. This information is used to determine the overall load of the slice.

[0173] After data collection, it is aggregated in the RIC module for AI processing.

[0174] In the case where the RIC module is a Non-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Near RT-RIC transmits the information collected by the SMO to the Non-RT RIC module through the A1 interface.

[0175] In the case where the RIC module is a Near-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Non-RT RIC then transmits the information to the Near-RT RIC module through the A1 interface. The network function transmits information to the Near-RT RIC module through the E2 interface.

[0176] Step 604: Determine the user sub-slice strategy. The RIC (Non-RT RIC and / or Near-RT RIC) optimizes the RRC configuration parameters for sub-slices with different transmission modes in the slice by running an AI module based on the collected data, or optimizes the RRC configuration parameters for each user of the sub-slices with different transmission modes in the slice.

[0177] Specifically, this decision-making process can be completed by the Non-RT RIC. The decision is passed to the Near-RT RIC via the A1 interface, and then passed to the network function by the Near-RT RIC via the E2 interface.

[0178] Specifically, this decision-making process can be completed by the Non-RT RIC. The decision is shared with the SMO via the SMO's internal bus, and then the SMO passes it to the network function through the O1 interface.

[0179] Specifically, the Near-RT RIC can run an AI module to complete this decision-making process, and the decision is passed to the network function through the E2 interface.

[0180] Specifically, the decision-making process can be completed by running the AI ​​module separately by the Non-RT RIC and the Near-RT RIC. The Non-RT RIC transmits the decision and the completed model training to the Near-RT RIC through the A1 interface. The Near-RT RIC further runs the AI ​​module based on the information transmitted by the Non-RT RIC and the information transmitted through the E2 interface, completes the final decision, and transmits it to the network function through the E2 interface.

[0181] Specifically, the RRC configuration parameters after the decision can be passed to the network function through sub-slices. For example, the corresponding RRC configuration parameter information can be listed for each sub-slice, such as DRX information for terminal nodes, SRS information for channel measurement, DMRS information for channel estimation, CSI-RS information for CSI measurement, and codebook configuration information.

[0182] Specifically, the RRC configuration parameters after the decision can be passed to the network function module in the form of RRC parameters for each user. For example, the corresponding RRC configuration parameter information for each user in each sub-slice can be listed, such as DRX information for terminal nodes, SRS information for channel measurement, DMRS information for channel estimation, CSI-RS information for CSI measurement, and codebook configuration information.

[0183] In existing technologies, there are no interface messages for transmitting user RRC configuration parameters. This disclosure introduces the above-mentioned interface message transmission to implement the function of sub-slice optimal RRC configuration parameters.

[0184] Specifically, for the implementation method of Non-RT RIC running AI module to determine the preferred RRC configuration parameters, it is necessary to add information related to the preferred transmission method to the A1 interface and E2 interface.

[0185] Specifically, to determine the preferred RRC configuration parameters for the Near-RT RIC running AI module, information related to the preferred transmission method needs to be added to the E2 interface.

[0186] Specifically, for the implementation method where the AI ​​modules of Non-RT RIC and Near-RT RIC respectively complete this decision-making, it is necessary to add Non-RT RIC training results and model deployment information to the A1 interface, and to add information related to the preferred RRC configuration parameters to the E2 interface. Specifically, the Non-RT RIC training results added to the A1 interface can be the preferred RRC configuration parameters for each user, the user's movement speed level information, or the user's movement model information, etc.

[0187] Step 605 is the same as step 505.

[0188] To support the above process, other auxiliary steps can be added.

[0189] For example, a network function module can inform the SMO via the O1 interface whether the network function module or the cell supports sub-slice functionality based on preferred transmission mode. This process is only performed on network function modules or cells that support this functionality. Figure 6 The process in the middle.

[0190] For example, the network function module can further report some measurement data to help the RIC evaluate the effectiveness of the preferred transmission method strategy, thereby helping the RIC optimize the algorithm or terminate the sub-slice function based on the preferred transmission method.

[0191] According to the above embodiments, by measuring user traffic, service type, mobility, etc., different transmission methods are selected for users in different scenarios, and specific implementation is achieved through slicing strategies, RRC configuration, or MAC scheduling strategies, thereby improving the throughput of the base station system, improving user experience, and helping terminals save power and other performance improvements.

[0192] Example B-1

[0193] Figure 7 This paper illustrates system performance optimization based on predicted user mobility scenarios according to embodiments of the present disclosure. Reference will be made below. Figure 7 Describe it.

[0194] Step 701 describes the triggering method. Unlike existing technologies, the embodiments of this disclosure are not event-triggered, but rather run periodically based on configuration. Configuration-triggered or periodically running methods can be used. For example, the function can be enabled based on the needs of a service provider or customer. Alternatively, this function can be continuously enabled within an O-RAN system.

[0195] Step 702, Data Collection. Compared to the prior art, this disclosure includes not only RAN internal data collected by the SMO using the O1 interface, but also rich data provided externally to the RAN.

[0196] The collected data mainly includes the user's moving speed, direction of movement, location, acceleration, etc., which are directly provided or calculated by GPS; if the user's navigation information, vehicle sensor information, etc. can be obtained, richer data can be acquired.

[0197] After data collection, it is aggregated in the RIC module for AI processing.

[0198] In the case where the RIC module is a Non-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Near RT-RIC transmits the information it collects to the Non-RT RIC module through the A1 interface.

[0199] In the case where the RIC module is a Near-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Non-RT RIC then transmits the information to the Near-RT RIC module through the A1 interface. The network function transmits information to the Near-RT RIC module through the E2 interface.

[0200] Step 703 involves training an AI / ML model for the RIC.

[0201] AI models can be deployed to Non-RT RICs and / or Near-RT RICs. The main difference between Non-RT RICs and Near-RT RICs is their real-time performance; Non-RT RICs operate on a second-level scale, while Near-RT RICs operate on a 10ms to 1s scale. Regardless of where the AI ​​model is deployed, it falls within the scope of this patent protection.

[0202] AI model training can be triggered periodically or by events, such as when the RAN's internal performance measurement data falls below a certain indicator, or when the number of users in a cell increases or decreases.

[0203] Step 704: Use an ML model for prediction.

[0204] AI models can be deployed to Non-RT RICs and / or Near-RT RICs. The main difference between Non-RT RICs and Near-RT RICs is their real-time performance; Non-RT RICs operate on a second-level scale, while Near-RT RICs operate on a 10ms to 1s scale. Regardless of where the AI ​​model is deployed, it falls within the scope of this patent protection.

[0205] The following description will use Near-RT RIC as an example. The processing of Near-RT RIC is similar and will not be described separately.

[0206] The SMO periodically sends data to the Near-RT RIC via the A1 interface. The Near-RT RIC uses an AI model for near real-time prediction. It is necessary to perform predictions for the next cycle instead of directly sending external data to the RAN node for the following reasons.

[0207] First, due to real-time reasons, neither Near-RT RIC (10ms~1s) nor Non-RT RIC (>1s) can guarantee absolute real-time performance. If data is sent directly to the RAN node, errors will occur due to time lag.

[0208] Second, AI / ML predictions can provide information that cannot be reflected in direct data, such as user acceleration, location at the next moment, or infer whether the service type is video or call based on the user's current business application layer information.

[0209] Third, data can be processed into the format required by RAN nodes, reducing the complexity of RAN processing.

[0210] Step 705: Functional modules in the RAN node use predictive information to assist in decision-making. Predictive information is used to determine the user transmission mode.

[0211] Specifically, RAN nodes receive mobility prediction information through the E2 interface.

[0212] The mobility prediction information may include detailed movement speed-related information such as the user's current movement speed, movement direction, acceleration, acceleration direction, and how long the movement speed will last.

[0213] The mobility prediction information may also include general information related to user movement speed, such as the user's mobility level, such as stationary, slow movement, medium speed movement, or high speed movement.

[0214] The mobility prediction information may also include information about the approximate range of user movement speed, such as the user's movement speed being below 1 km / h or between 1 km / h and 3 km / h.

[0215] Compared to existing technologies, this disclosure introduces user-level mobility-related information on the E2 interface.

[0216] Specifically, after receiving user mobility-related information, the RAN node will determine the user's preferred transmission method based on the above information.

[0217] For example, MU-MIMO transmission is preferred for low-speed users, while SU-MIMO mobility is preferred for high-speed users.

[0218] To support the above process, other auxiliary steps can be added.

[0219] For example, a network function module can inform the SMO via the O1 interface whether it or the cell supports functions based on received user mobility information. This process is only performed on network function modules or cells that support this function. Figure 7 The process in the middle.

[0220] For example, the network function module can further report some measurement data to help the RIC evaluate the effectiveness of the preferred transmission method strategy, thereby helping the RIC optimize the algorithm or terminate the function.

[0221] According to the above embodiments, by predicting mobility, different transmission methods are selected for users in different scenarios, and this is specifically implemented through slicing strategies, RRC configuration, or MAC scheduling strategies, thereby improving the throughput of the base station system, enhancing user experience, and helping terminals save power, among other performance improvements.

[0222] Example B-2

[0223] Figure 8 This illustrates system performance optimization based on predictive business scenarios according to embodiments of the present disclosure. Reference will be made below. Figure 8 Describe it.

[0224] Step 801 describes the triggering method. Unlike existing technologies, the embodiments of this disclosure are not event-triggered, but rather run periodically based on configuration. Configuration-triggered or periodically running methods can be used. For example, the function can be enabled based on the needs of a service provider or customer. Alternatively, this function can be continuously enabled within an O-RAN system.

[0225] Step 802, Data Collection. Compared to the prior art, this disclosure includes not only RAN internal data collected by the SMO using the O1 interface, but also rich data provided externally to the RAN.

[0226] The data collected mainly includes the type and traffic information of the user's current business, such as whether they are making a voice call, transferring files, or watching videos, and also obtains information such as the size of the transferred files and the duration of the videos.

[0227] After data collection, it is aggregated in the RIC module for AI processing.

[0228] In the case where the RIC module is a Non-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Near RT-RIC transmits the information collected by the SMO to the Non-RT RIC module through the A1 interface.

[0229] In the case where the RIC module is a Near-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Non-RT RIC then transmits the information to the Near-RT RIC module through the A1 interface. The network function transmits information to the Near-RT RIC module through the E2 interface.

[0230] Step 803 involves training an AI / ML model for the RIC.

[0231] AI models can be deployed to Non-RT RICs and / or Near-RT RICs. The main difference between Non-RT RICs and Near-RT RICs is their real-time performance; Non-RT RICs operate on a second-level scale, while Near-RT RICs operate on a 10ms to 1s scale. Regardless of where the AI ​​model is deployed, it falls within the scope of this patent protection.

[0232] AI model training can be triggered periodically or by events, such as when the RAN's internal performance measurement data falls below a certain indicator, or when the number of users in a cell increases or decreases.

[0233] Step 804 involves using an ML model for prediction.

[0234] AI models can be deployed to Non-RT RICs and / or Near-RT RICs. The main difference between Non-RT RICs and Near-RT RICs is their real-time performance; Non-RT RICs operate on a second-level scale, while Near-RT RICs operate on a 10ms to 1s scale. Regardless of where the AI ​​model is deployed, it falls within the scope of this patent protection.

[0235] The following description will use Near-RT RIC as an example. The processing of Near-RT RIC is similar and will not be described separately.

[0236] The SMO periodically sends data to the Near-RT RIC via the A1 interface. The Near-RT RIC uses an AI model for near real-time prediction. It is necessary to perform predictions for the next cycle instead of directly sending external data to the RAN node for the following reasons.

[0237] First, due to real-time reasons, neither Near-RT RIC (10ms~1s) nor Non-RT RIC (>1s) can guarantee absolute real-time performance. If data is sent directly to the RAN node, errors will occur due to time lag.

[0238] Second, AI / ML predictions can provide information that cannot be reflected in direct data, such as user acceleration, location at the next moment, or infer whether the service type is video or call based on the user's current business application layer information.

[0239] Third, data can be processed into the format required by RAN nodes, reducing the complexity of RAN processing.

[0240] Step 805: Functional modules in the RAN node use predictive information to assist in decision-making. Predictive information is used to determine the user transmission method.

[0241] Specifically, the RAN node receives traffic prediction information through the E2 interface.

[0242] The traffic prediction information may include the user's current buffer occupy (BO), service model, service arrival interval, average BO, and variance of BO.

[0243] The traffic information may also include general user traffic-related information, such as the user's traffic level, such as low-traffic, medium-traffic, or high-traffic services.

[0244] The traffic prediction information may also include information about the approximate traffic range, such as the user's traffic rate being between 10 bytes per millisecond and 300 bytes per millisecond.

[0245] Compared to existing technologies, this disclosure introduces user-level traffic-related information on the E2 interface.

[0246] Specifically, after receiving traffic-related information, the RAN node will determine the user's preferred transmission method based on the above information.

[0247] For example, SU-MIMO transmission is preferred for low-traffic users, while MU-MIMO mobility is preferred for high-traffic users.

[0248] To support the above process, other auxiliary steps can be added.

[0249] For example, a network function module can inform the SMO via the O1 interface whether it or the cell supports functions based on received user traffic information. This process is only performed on network function modules or cells that support this function. Figure 8 The process in the middle.

[0250] For example, the network function module can further report some measurement data to help the RIC evaluate the effectiveness of the preferred transmission method strategy, thereby helping the RIC optimize the algorithm or terminate the function.

[0251] According to the above embodiments, by predicting user traffic and service types, different transmission methods are selected for users in different scenarios, and this is specifically implemented through slicing strategies, RRC configuration, or MAC scheduling strategies, thereby improving the throughput of the base station system, enhancing user experience, and helping terminals save power, among other performance improvements.

[0252] Example B-3

[0253] Figure 9 Prediction-based configuration parameter optimization according to embodiments of the present disclosure is illustrated. Reference will be made below. Figure 9 Describe it.

[0254] Step 901 describes the triggering method. Unlike existing technologies, the embodiments of this disclosure are not event-triggered, but rather run periodically based on configuration. Configuration-triggered or periodically running methods can be used. For example, the function can be enabled based on the needs of a service provider or customer. Alternatively, this function can be continuously enabled within an O-RAN system.

[0255] Step 902, Data Collection. Compared to the prior art, this disclosure includes not only RAN internal data collected by the SMO using the O1 interface, but also rich data provided externally to the RAN.

[0256] The collection of user business-related data mainly includes the type and traffic information of the user's current business, such as whether the user is making a voice call, transferring files, or watching videos, and also obtains information such as the size of the transferred files and the duration of the videos.

[0257] After data collection, it is aggregated in the RIC module for AI processing.

[0258] In the case where the RIC module is a Non-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Near RT-RIC transmits the information collected by the SMO to the Non-RT RIC module through the A1 interface.

[0259] In the case where the RIC module is a Near-RT RIC, the SMO transmits the information collected by the SMO to the Non-RT RIC module through the internal bus, and the Non-RT RIC then transmits the information to the Near-RT RIC module through the A1 interface. The network function transmits information to the Near-RT RIC module through the E2 interface.

[0260] Step 903 involves training an AI / ML model for the RIC.

[0261] AI models can be deployed to Non-RT RICs and / or Near-RT RICs. The main difference between Non-RT RICs and Near-RT RICs is their real-time performance; Non-RT RICs operate on a second-level scale, while Near-RT RICs operate on a 10ms to 1s scale. Regardless of where the AI ​​model is deployed, it falls within the scope of this patent protection.

[0262] AI model training can be triggered periodically or by events, such as when the RAN's internal performance measurement data falls below a certain indicator, or when the number of users in a cell increases or decreases.

[0263] Step 904 involves using an ML model for prediction.

[0264] AI models can be deployed to Non-RT RICs and / or Near-RT RICs. The main difference between Non-RT RICs and Near-RT RICs is their real-time performance; Non-RT RICs operate on a second-level scale, while Near-RT RICs operate on a 10ms to 1s scale. Regardless of where the AI ​​model is deployed, it falls within the scope of this patent protection.

[0265] The following description will use Near-RT RIC as an example. The processing of Near-RT RIC is similar and will not be described separately.

[0266] The SMO periodically sends data to the Near-RT RIC via the A1 interface. The Near-RT RIC uses an AI model for near real-time prediction. It is necessary to perform predictions for the next cycle instead of directly sending external data to the RAN node for the following reasons.

[0267] First, due to real-time reasons, neither Near-RT RIC (10ms~1s) nor Non-RT RIC (>1s) can guarantee absolute real-time performance. If data is sent directly to the RAN node, errors will occur due to time lag.

[0268] Second, AI / ML predictions can provide information that cannot be reflected in direct data, such as user acceleration, location at the next moment, or infer whether the service type is video or call based on the user's current business application layer information.

[0269] Third, data can be processed into the format required by RAN nodes, reducing the complexity of RAN processing.

[0270] Step 905: Functional modules in the RAN node use predictive information to assist in decision-making. Predictive information is used to make decisions regarding user RRC configuration.

[0271] Specifically, the RAN node receives traffic prediction information through the E2 interface.

[0272] The traffic prediction information may include the user's current buffer occupy (BO), service model, service arrival interval, average BO, and variance of BO.

[0273] The traffic information may also include general user traffic-related information, such as the user's traffic level, such as low-traffic, medium-traffic, or high-traffic services.

[0274] The traffic prediction information may also include information about the approximate traffic range, such as the user's traffic rate being between 10 bytes per millisecond and 300 bytes per millisecond.

[0275] Compared to existing technologies, this disclosure introduces user-level traffic-related information on the E2 interface.

[0276] Specifically, after receiving traffic-related information, the RAN node will determine the user's RRC configuration parameters based on the above information.

[0277] For example, for low-traffic users, a longer DRX period and a longer SRS period configuration are preferred. For high-traffic users, a shorter DRX period and a shorter SRS period configuration are preferred.

[0278] To support the above process, other auxiliary steps can be added.

[0279] For example, the network function module can inform the SMO via the O1 interface whether the network function module or the cell supports functions based on received user traffic information. This process is only performed on network function modules or cells that support this function. Figure 9 The process in the middle.

[0280] For example, the network function module can further report some measurement data to help the RIC evaluate the effectiveness of the preferred transmission method strategy, thereby helping the RIC optimize the algorithm or terminate the function.

[0281] According to the above embodiments, by predicting user traffic volume, service type, and mobility, different transmission methods are selected for users in different scenarios, and this is specifically implemented through slicing strategies, RRC configuration, or MAC scheduling strategies, thereby improving the throughput of the base station system, enhancing user experience, and helping terminals save power, among other performance improvements.

[0282] In order to achieve system performance optimization according to the embodiments of the present disclosure described above, the following messages need to be introduced (some specific interface information is listed as examples).

[0283] A1 interface, downlink direction. If the Non-RT RIC does not execute the ML model, the following data is required: user GPS, navigation route, sensor information (vehicle or other mobile devices), user service software application layer information, such as service type, video size and duration, file size, etc. If the Non-RT RIC executes the ML model for prediction, the following prediction data is required: For Example B-1, mobility data, the speed of each user at all time points in the next period (e.g., 1 second period), direction of movement, acceleration, orientation, altitude, and other possible information; for Example B-2, service model data, including service type, transmission rate, and duration.

[0284] E2 interface, downlink direction. If Near-RT RIC executes the ML model, the following data is required: For Example B-1, mobility data, velocity (e.g., time points divided in 1ms), direction of movement, acceleration, orientation, altitude, and other possible information for each user at all time points in the next period (e.g., 100ms period); for Example B-2, service model data, including service type, transmission rate, and duration.

[0285] The E2 interface is for the uplink direction. It provides capability descriptions for the RAN functional modules, including a description of the MAC scheduling module's ability to process user mobility information and a description of the RRC module's ability to process user service model information.

[0286] The O1 interface, in the uplink direction, needs to support the reporting of relevant measurement data, such as the total system throughput, the average throughput of users, the user MIMO transmission mode, the average number of layers, etc.

[0287] Furthermore, in order to support the embodiments of this disclosure, the following modifications are required to the interface:

[0288] 1. Add a sub-slice instance information structure, defined as follows.

[0289]

[0290] This structure includes at least the transmission method and the sub-slice instance identifier. This structure is used in both measurement reporting and decision message distribution. The relevant interfaces are E2 and A1.

[0291] 2. Add user information about Massive MIMO capabilities.

[0292]

[0293]

[0294] 3. Mobility Data Definition (Downlink of A1 and E2 Interfaces)

[0295]

[0296] 4. Business Model Data Definition (Downlink of A1 and E2 Interfaces)

[0297]

[0298] The above is for reference only. Figures 3 to 9 Specific embodiments according to this disclosure are described. Reference will be made below. Figure 10 and Figure 11 This describes the general flow for both Class A (i.e., RIC directly participates in decision-making and utilizes slicing) and Class B (RIC performs prediction, assisting modules in the RAN node to perform optimization processing) implementations.

[0299] Figure 10 A general flowchart illustrating a slice-based performance optimization implementation for decision-making by a RAN intelligent controller according to an embodiment of the present disclosure is shown. Reference will be made below. Figure 10 Describe it.

[0300] The operation process for the Class A embodiment generally includes the following steps.

[0301] Step 1 involves creating sub-slices supporting MU (Multi-Use) and SU (Multi-Use) transmission modes, corresponding to step 301 (described using Example A-1 as an example). Based on these sub-slices, users with different transmission modes are separated from radio resources. This improves spectrum efficiency and facilitates subsequent optimized management of the sub-slices.

[0302] Step 2: The SMO requests the necessary data information from the E2 Nodes.

[0303] Step 3: The E2 Nodes report information to the SMO, corresponding to step 302. In the embodiments of this disclosure, Non-RTRIC needs to determine which type of transmission method the user should use based on the user's capability information provided by the E2 Nodes.

[0304] In steps 4 and 5, the SMO obtains GPS and service information from an external server. This corresponds to step 302. In this solution, the AI / ML module of the Non-RT RIC obtains the user's mobility model based on GPS information, providing information such as speed, direction of movement, spatial coordinates, and surrounding buildings for Non-RT RIC decision-making. The AI / ML module of the Non-RT RIC obtains the user's service model based on service information, providing information such as service type, service data volume, and service QoS requirements for Non-RT RIC decision-making.

[0305] Step 6: The Non-RT RIC allocates the optimal transmission method to the user based on mobility information, service information, user capability information, etc., corresponding to step 303. In the embodiments of this disclosure, the user dynamically adjusts their transmission method according to the scenario, thereby reducing unnecessary retransmissions and improving system throughput and resource utilization.

[0306] Steps 7 and 8 distribute the Non-RT RIC decision results to E2 Nodes (O-DU and O-CU-CP), corresponding to step 304. The O-DU is responsible for scheduling and radio resource management. The O-CU-CP is responsible for configuration parameter management.

[0307] The SMO obtains GPS (Global Positioning System) and service information from external application software or application servers, and acquires user capability information, slice-related measurement information, and user performance measurement information from E2 Nodes. The AI / ML module in the Non-RT RIC obtains the mobility model and service model based on the above information. The AI / ML module in the Non-RT RIC selects the optimal transmission mode (e.g., SU-MIMO or MU-MIMO) and optimal configuration parameters (DRX (Discontinuous Reception), SR (Scheduling Request), PDSCH (Physical Downlink Shared Channel), etc.) for the user based on the mobility model, service model, slice-related measurement information, user performance measurement information, and user capability information. The Non-RT RIC sends the configuration parameters to the O-CU-CP and informs the O-DU of the user's slice information. The O-CU-CP assigns the configuration parameters to the UE via signaling messages and simultaneously updates the user's configuration parameters on the O-DU side. Based on the received slice configuration notification, O-DU establishes the user's bearer information in the corresponding slice. After configuration, it then provides services for the user in that slice.

[0308] Figure 11A general flowchart illustrating a predictive, assisted RAN node performance optimization implementation method according to an embodiment of the present disclosure is shown. Reference will be made below. Figure 11 Describe it.

[0309] The operation process for the B-type embodiment generally includes the following steps.

[0310] Step 1: Periodically request the O1 interface to report internal RAN measurement data.

[0311] Step 2: The O1 interface reports internal RAN measurement data.

[0312] Step 3: Periodically request rich data from outside the RAN.

[0313] Step 4: Collect external RAN data, including mobility-related data such as GPS, and service model-related data such as transmission rate.

[0314] Step 5: Training the AI / ML model with the non-real-time intelligent controller is a necessary step. Performing ML inference to predict mobility and business models with the non-real-time intelligent controller is an optional step.

[0315] Step 6: Send the prediction information to the near real-time intelligent controller.

[0316] Step 7: The near real-time intelligent controller performs ML inference to predict mobility and business models. This is an optional step.

[0317] Step 8: The near real-time intelligent controller sends the prediction results to the functional modules in the RAN node, which is a necessary step.

[0318] Step 9: Performance measurement after O1 interface feedback of RAN parameter adjustment or MAC scheduling strategy adjustment.

[0319] Step 10: The near real-time intelligent controller acquires RAN performance measurement information and updated RAN external information.

[0320] Step 11: The non-real-time intelligent controller measures the impact of the predicted information on the system based on performance feedback, and then judges whether the ML model is reasonable.

[0321] Step 12: If the non-real-time intelligent controller deems the ML model unreasonable and needs to trigger retraining, this is an optional step.

[0322] Step 13: Update the ML model.

[0323] The above description is merely a preferred embodiment of this disclosure and is not intended to limit this disclosure. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

[0324] Those skilled in the art will understand that this disclosure includes devices for performing one or more of the operations described in this disclosure. These devices may be specifically designed and manufactured for the desired purpose, or may include known devices found in general-purpose computers. These devices have computer programs stored therein that can be selectively activated or reconfigured. Such computer programs may be stored in a device (e.g., a computer)-readable medium or in any type of medium suitable for storing electronic instructions and coupled to a bus, including but not limited to any type of disk (including floppy disks, hard disks, optical disks, CD-ROMs, and magneto-optical disks), ROM (Read-Only Memory), RAM (Random Access Memory), EPROM (Erasable Programmable Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory, magnetic cards, or optical cards. That is, a readable medium includes any medium by which a device (e.g., a computer) stores or transmits information in a readable form.

[0325] Those skilled in the art will understand that each block in these structural diagrams and / or block diagrams and / or flow diagrams, as well as combinations of blocks in these structural diagrams and / or block diagrams and / or flow diagrams, can be implemented using computer program instructions. Those skilled in the art will understand that these computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing method for implementation, thereby enabling the processor of the computer or other programmable data processing method to execute the schemes specified in the blocks or plurality of blocks of the structural diagrams and / or block diagrams and / or flow diagrams disclosed herein.

[0326] Those skilled in the art will understand that the steps, measures, and schemes in the various operations, methods, and processes discussed in this disclosure can be alternated, modified, combined, or deleted. Furthermore, other steps, measures, and schemes in the various operations, methods, and processes discussed in this disclosure can also be alternated, modified, rearranged, decomposed, combined, or deleted. Furthermore, steps, measures, and schemes in the prior art that are similar to those in the various operations, methods, and processes disclosed in this disclosure can also be alternated, modified, rearranged, decomposed, combined, or deleted.

[0327] The above description is only a partial embodiment of this disclosure. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles of this disclosure, and these improvements and modifications should also be considered within the scope of protection of this disclosure.

Claims

1. A method performed by a Radio Access Network (RAN) Intelligent Controller (RIC) entity in a wireless communication system, comprising: First information is obtained from the base station, the first information including at least one of capability information or measurement information, the capability information including at least one of user capability information or cell capability information, and the measurement information including at least one of slice measurement information or user performance measurement information; Obtain second information related to the terminal state from the application server, the second information including information related to user mobility; Based on the first information and the second information, at least one first terminal corresponding to the first sub-slice instance is determined from multiple terminals, and at least one second terminal corresponding to the second sub-slice instance is determined from the multiple terminals. as well as The determined result is sent to the base station. The first sub-slice instance is associated with single-user multiple-input multiple-output (SU-MIMO), and the second sub-slice instance is associated with multi-user multiple-input multiple-output (MU-MIMO).

2. The method according to claim 1, wherein, The second information also includes business traffic information. The Artificial Intelligence / Machine Learning (AI / ML) module was used for this determination.

3. The method according to claim 2, wherein, The determination includes: Based on the user mobility-related information, the AI / ML module is used to estimate the mobility of each of the plurality of terminals, and / or based on the service traffic information, the AI / ML module is used to estimate the service traffic usage of each of the plurality of terminals; and Based on the estimation, the at least one first terminal is assigned to the first sub-slice instance, and the at least one second terminal is assigned to the second sub-slice instance.

4. The method according to claim 1, further comprising: Determine a Radio Resource Control (RRC) configuration for each of the at least one first terminal and the at least one second terminal; as well as The determined RRC configuration is sent to the base station.

5. The method according to claim 1, wherein, The RIC entity is a non-real-time RIC entity, and the determined result is sent to the base station via a near-real-time RIC entity. Among them, the Radio Access Network (RAN) is an open RAN (O-RAN).

6. The method according to claim 1, wherein, The information related to user mobility includes Global Positioning System (GPS) information.

7. A method performed by a base station in a wireless communication system, comprising: Send first information to the Radio Access Network (RAN) Intelligent Controller (RIC) entity, the first information including at least one of capability information or measurement information, the capability information including at least one of user capability information or cell capability information, the measurement information including at least one of slice measurement information or user performance measurement information; The RIC entity receives policy information indicating sub-slice instances of multiple terminals, the sub-slice instances including a first sub-slice instance and a second sub-slice instance, the first sub-slice instance corresponding to at least one first terminal among the multiple terminals, and the second sub-slice instance corresponding to at least one second terminal among the multiple terminals. as well as Based on the policy information, resources are scheduled for the multiple terminals. The policy information is generated based on the first information and second information related to the terminal state. The second information is obtained from the application server and includes information related to user mobility. The first sub-slice instance is associated with single-user multiple-input multiple-output (SU-MIMO), and the second sub-slice instance is associated with multi-user multiple-input multiple-output (MU-MIMO).

8. The method according to claim 7, wherein, in, The second information also includes business traffic information. The Artificial Intelligence / Machine Learning (AI / ML) module is used to generate the strategy information based on the second information.

9. The method according to claim 8, wherein, Based on at least one of the following: the mobility of each of the plurality of terminals estimated by the AI / ML module based on information related to user mobility, or the service usage of each of the plurality of terminals estimated based on service traffic information, the at least one first terminal is assigned to a first sub-slice instance, and the at least one second terminal is assigned to a second sub-slice instance.

10. The method of claim 7, further comprising: Receive Radio Resource Control (RRC) configurations from each of the at least one first terminal and the at least one second terminal from the RIC entity. The RRC configuration is determined by the RIC entity based on the policy information.

11. The method according to claim 7, wherein, The RIC entity is a non-real-time RIC entity, and the policy information is received via a near-real-time RIC entity. Among them, the Radio Access Network (RAN) is an open RAN (O-RAN).

12. The method according to claim 7, wherein, The information related to user mobility includes Global Positioning System (GPS) information.

13. A Radio Access Network (RAN) Intelligent Controller (RIC) entity in a wireless communication system, comprising: transceiver; A controller, coupled to the transceiver and configured to perform the method as described in any one of claims 1-6.

14. A base station in a wireless communication system, comprising: transceiver; A controller, coupled to the transceiver and configured to perform the method as described in any one of claims 7-13.

Images