Implementing advanced sleep modes in telecommunications networks

Abstract

A system and method for implementing an advanced sleep mode in an O-RAN includes collecting measurement data for training an AI / ML model, training the AI / ML model based on the collected measurement data and deploying the AI / ML model in an nRT-RIC, activating the trained AI / ML model in the nRT-RIC, monitoring energy optimization data for the AI / ML model inference from an open radio unit (O-RU) via an E2 node, activating at least one advanced sleep mode (ASM) in the nRT-RIC, collecting data for temporarily suspending O-RU components based on the activation of the ASM, evaluating the collected data for temporarily suspending O-RU components based on the activated AI / ML model and the ASM, requesting initiation of the ASM to the O-RU based on the evaluation, and implementing the ASM by the E2 node and the O-RU based on the ASM initiation request.

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application is based on and claims priority to U.S. Provisional Patent Application No. 63 / 481,637, filed January 26, 2023, the disclosure of which is incorporated herein by reference in its entirety.

[0002] The present disclosure relates to implementing advanced sleep modes to reduce power consumption of radio base stations (BSs) in telecommunications networks. [Background technology]

[0003] The Radio Access Network (RAN) is a key component in telecommunications systems because it connects end-user devices (or user equipment) to the rest of the network. The RAN includes a combination of various Network Elements (NEs) that connect end-user devices to the core network. Traditionally, the hardware and / or software of a particular RAN is vendor-specific.

[0004] Open RAN (O-RAN) technology is emerging to enable multiple vendors to provide hardware and / or software for telecommunications systems. To this end, O-RAN decomposes RAN functions into an open centralized unit (O-CU), an open distributed unit (O-DU), and an open radio unit (O-RU). The O-CU is a logical node for hosting the RAN's Radio Resource Control (RRC), Service Data Adaptation Protocol (SDAP), and / or Packet Data Convergence Protocol (PDCP) sublayers. The O-DU is a logical node for hosting the O-RAN's Radio Link Control (RLC), Media Access Control (MAC), and Physical (PHY) sublayers. The O-RU is a physical node that converts radio signals from the antennas into digital signals that can be transmitted to the O-DU via the fronthaul (FH). These entities have open protocols and interfaces between them, allowing them to be developed by different vendors.

[0005] In the related art, the power consumption of a base station BS (e.g., an open radio unit (O-RU)) varies depending on the amount of cell traffic. As traffic increases, the physical components of the O-RU, such as the RF front-end module (RFFE) (e.g., power amplifier (PA), radio frequency (RF) transceiver (TRx), etc.), become the main energy consumers. However, in low-traffic scenarios, the main energy consumption comes from functional components (i.e., processing equipment) such as the digital RF front-end (DFE), lower PHY layer baseband processing functions, synchronization and fronthaul transport functions, etc. Furthermore, in the related art, the O-DU and / or O-CU (i.e., E2 node) may have to support several technologies (i.e., depending on the specific scenario and network deployment) that may affect energy consumption and may be load-dependent (e.g., network traffic, number of users, etc.).

[0006] To this end, because the E2 nodes (e.g., O-DU, O-CU, etc.) may not be aware of the internal architecture of the multi-vendor O-RU design (i.e., the capabilities of the O-RU's functional and physical components), the related technologies have the challenge that some parts of the O-RU (e.g., functional and / or physical components) continue to consume energy even when no transmission is taking place.

[0007] As a result, according to the related art, high-level (i.e., comprehensive) energy saving modes according to performance targets, such as those provided via A1 policies, performance triggers based on O-RAN key performance indicators (KPIs) (e.g., data traffic, data throughput), etc., are limited to energy saving models applied by the O-DU without knowledge of the O-RU capabilities to save energy (e.g., the O-RU's internal system architecture, such as the capabilities of the O-RU's physical and / or functional components). Therefore, energy saving models (i.e., comprehensive energy saving modes) applied to the O-RU according to the related art may cause unnecessary energy consumption due to poor application of these energy saving modes caused by lack of knowledge of the O-RU's component capabilities.

[0008] Furthermore, considering the local or network-wide impact of energy saving modes in O-RAN, there is a trade-off between system performance and energy savings. Therefore, according to related art, implementing energy saving modes while maintaining high overall network performance in O-RAN is a complex task. For example, other carriers and / or cells may have to cover (i.e., take over or service) additional network traffic, and network traffic changes over time.

[0009] To this end, related art lacks a level of abstraction that allows for an effective balance between system performance and energy savings.

[0010] As a result, according to the energy saving mode of the related technology, while saving energy in one O-RU can be maximized on a local basis, the overall network performance of the O-RAN may be reduced due to the complex task of balancing system performance and energy saving. Summary of the Invention [Means for solving the problem]

[0011] According to embodiments, the present disclosure relates to an implementation of an advanced sleep mode (ASM) that enables a near real-time radio access network intelligence controller (nRT-RIC) to consider multiple data dimensions, such as traffic load, user service type, and energy efficiency measurements, O-DU capability, and O-RU capability, and determine which ASM (i.e., which level or type of ASM) can be optimally applied by an energy saving (ES) function to achieve an optimal balance between system performance and energy savings. In this case, each level of ASM (i.e., type of ASM) comprises one or more intelligent scheduling functions associated with the respective ASM level / type, and each of the one or more intelligent scheduling functions applied by the O-DU and / or O-CU to the O-RU results in a different power consumption level of the O-RU components (i.e., physical and / or functional components of the O-RU).

[0012] In particular, the one or more intelligent scheduling functions may include scheduling of slots and / or symbols (e.g., through data convergence (traffic shaping)), and the intelligent scheduling functions (e.g., scheduling) associated with one or more ASM types / levels have the advantage of allowing temporary suspension (muting / switching off) of O-RU components to optimize power consumption of the O-RU.

[0013] As a result, the implementation of one or more ASM types / levels that take into account the capability data of the O-RU and / or O-DU allows for a level of abstraction of energy saving modes that allows for optimal operating energy efficiency of the O-RU while maintaining a high level of network performance across the O-RAN.

[0014] According to an embodiment, a system for implementing an advanced sleep mode in an open radio access network (O-RAN) includes a near-real-time radio access network (nRT-RIC) and a service management and orchestration (SMO) framework, where the SMO framework includes a non-real-time radio access network (RAN) intelligent controller (NRT-RIC). The system includes: collecting measurement data for training an artificial intelligence / machine learning (AI / ML) model by the SMO framework; training the AI / ML model by the NRT-RIC based on the collected measurement data; deploying the AI / ML model to the nRT-RIC; launching the trained AI / ML model in the nRT-RIC by the SMO framework; and transmitting the O-RAN data to an E2 node by the NRT-RIC. The method is configured to monitor energy optimization data for AI / ML model inference from an O-RU; activate at least one advanced sleep mode (ASM) in the nRT-RIC via the SMO framework; collect data from the O-RU via the E2 node by the nRT-RIC based on the activation of the at least one ASM to temporarily shut down one or more O-RU components; evaluate the collected data by the nRT-RIC based on the activated AI / ML model and the at least one ASM to temporarily shut down one or more O-RU components; request the O-RU via the E2 node to start the at least one ASM based on the evaluation; and implement the at least one ASM by the O-RU based on the ASM start request.

[0015] According to an embodiment, in a method for implementing an advanced sleep mode in an open radio access network (O-RAN), the method includes: collecting measurement data for training an artificial intelligence / machine learning (AI / ML) model by a service management and orchestration (SMO) framework; training the AI / ML model by a non-real-time radio access network (RAN) intelligent controller (NRT-RIC) based on the collected measurement data and deploying the AI / ML model to a near-real-time radio access network (RAN) intelligent controller (nRT-RIC); invoking the trained AI / ML model in the nRT-RIC by the SMO framework; receiving an AI / ML signal from an open radio unit (O-RU) via an E2 node by the NRT-RIC; The method includes monitoring energy optimization data for AI / ML model inference; activating at least one advanced sleep mode (ASM) in the nRT-RIC by the SMO framework; collecting data by the nRT-RIC from the O-RU via the E2 node based on the activation of the at least one ASM to temporarily shut down one or more O-RU components; evaluating the collected data by the nRT-RIC based on the activated AI / ML model and the at least one ASM to temporarily shut down one or more O-RU components; requesting, by the nRT-RIC, via the E2 node to start the at least one ASM based on the evaluation; and implementing, by the O-RU, the at least one ASM based on the ASM initiation request.

[0016] According to an embodiment, a non-transitory computer-readable storage medium having stored thereon instructions executable by at least one processor, the at least one processor being configured to implement a method for implementing an advanced sleep mode in an open radio access network (O-RAN), the method including: collecting measurement data for training an artificial intelligence / machine learning (AI / ML) model by a service management and orchestration (SMO) framework; training the AI / ML model by a non-real-time radio access network (RAN) intelligent controller (NRT-RIC) based on the collected measurement data and deploying the AI / ML model in a near-real-time radio access network (RAN) intelligent controller (nRT-RIC); invoking the trained AI / ML model in the nRT-RIC by the SMO framework; The method includes, by the IC, monitoring energy optimization data for AI / ML model inference from an open radio unit (O-RU) via the E2 node; activating at least one advanced sleep mode (ASM) in the nRT-RIC via the SMO framework; collecting data to temporarily shut down one or more O-RU components from the O-RU via the E2 node based on the activation of the at least one ASM; evaluating the collected data to temporarily shut down one or more O-RU components by the nRT-RIC based on the activated AI / ML model and the at least one ASM; requesting the O-RU via the E2 node to start the at least one ASM based on the evaluation; and implementing the at least one ASM by the O-RU based on the ASM initiation request.

[0017] Additional aspects will be set forth in part in the description that follows, and in part will be apparent from the description, or may be realized by practice of presented embodiments of the present disclosure.

[0018] Features, aspects, and advantages of certain exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, in which like reference numerals refer to like elements. [Brief explanation of the drawings]

[0019] [Figure 1] 1 shows the O-RAN architecture in the related art.

[0020] [Figure 2] FIG. 1 illustrates an example environment in which the systems and / or methods described herein may be implemented.

[0021] [Figure 3] FIG. 2 illustrates exemplary components of a device according to one embodiment.

[0022] [Figure 4] FIG. 1 illustrates an SMO / NRT-RIC framework system architecture configured to implement optimization of partial shutdown of O-RU components based on A1 policy, according to one embodiment.

[0023] [Figure 5] FIG. 1 is a flow diagram of a method for implementing an advanced sleep mode, according to one embodiment.

[0024] [Figure 6] FIG. 10 is a flow diagram of a method for collecting data to temporarily shut down one or more O-RU components based on activation of at least one ASM, according to another embodiment.

[0025] [Figure 7A] FIG. 1 is a flow diagram of a method for evaluating collected data to temporarily shut down one or more O-RU components, according to one embodiment.

[0026] [Figure 7B]FIG. 1 is a flow diagram of a method for evaluating collected data to temporarily shut down one or more O-RU components, according to one embodiment.

[0027] [Figure 8] FIG. 1 is a flow diagram of a method for implementing an advanced sleep mode, according to one embodiment.

[0028] [Figure 9] FIG. 1 illustrates several types of advanced sleep modes, according to one embodiment.

[0029] [Figure 10] 1 illustrates a method for scheduling at least one symbol to minimize the number of symbols in the time domain according to one embodiment.

[0030] [Figure 11] 1 illustrates a method for scheduling at least one slot to minimize the number of slots in the time domain, according to one embodiment.

[0031] [Figure 12A] FIG. 1 illustrates a method for collecting data over an O1 interface in a hierarchical O-RAN architecture, according to one embodiment.

[0032] [Figure 12B] FIG. 1 illustrates a method for collecting data over an O1 interface in a hybrid O-RAN architecture, according to one embodiment.

[0033] [Figure 13] FIG. 1 illustrates a method for training an AI / ML model within an nRT-RIC and deploying the AI / ML model, according to one embodiment.

[0034] [Figure 14]A diagram illustrating a method for invoking at least one advanced sleep mode (ASM) in an nRT-RIC by an SMO framework, according to one embodiment.

[0035] [Figure 15] FIG. 1 illustrates a method for collecting data to temporarily suspend one or more O-RU components based on activation of at least one ASM, according to one embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0036] The following detailed description of exemplary embodiments refers to the accompanying drawings. The foregoing disclosure provides illustration and description, but is not intended to be exhaustive or to limit implementations to the precise form disclosed. Modifications and variations are possible in light of the above disclosure or may be acquired from practice of implementations. Furthermore, one or more features or components of one embodiment may be incorporated into or combined with another embodiment (or one or more features of another embodiment). Additionally, in the flowcharts and descriptions of operations provided below, it should be understood that one or more operations may be omitted, one or more operations may be added, one or more operations may be performed (at least partially) concurrently, and the order of one or more operations may be rearranged.

[0037] It will be apparent that the systems and / or methods described herein may be implemented in different forms of hardware, firmware, or a combination of hardware and software. The actual dedicated control hardware or software code used to implement these systems and / or methods is not limiting of the implementation. Thus, the operation and behavior of the systems and / or methods are described herein without reference to specific software code. It should be understood that software and hardware can be designed to implement the systems and / or methods based on the description herein.

[0038] Although particular combinations of features are recited in the claims and / or disclosed herein, these combinations are not intended to limit the disclosure of possible implementations. Indeed, many of these features can be combined in ways not specifically recited in the claims and / or disclosed herein. Although each dependent claim listed below may depend directly on only one claim, the disclosure of possible implementations includes each dependent claim in combination with all other claims in the claim set.

[0039] No element, act, or instruction used herein should be construed as critical or required unless explicitly stated as such. Also, as used herein, the articles "a" and "an" are intended to include one or more items and may be used interchangeably with "one or more." Where only one item is intended, the term "one" or similar language is used. Also, as used herein, terms such as "has," "have," "having," "include," and "including" are intended to be open-ended terms. Furthermore, the phrase "based on" is intended to mean "based at least in part on," unless specifically stated otherwise. Furthermore, phrases such as "at least one of [A] and [B]" or "at least one of [A] or [B]" should be understood to include A only, B only, or both A and B.

[0040] The Radio Access Network (RAN) is a key component in telecommunications systems because it connects end-user devices (or user equipment) to the rest of the network. The RAN includes a combination of various network elements (NEs) that connect end-user devices to the core network. Traditionally, the hardware and / or software of a particular RAN is vendor-specific.

[0041] Open RAN (O-RAN) technology has emerged to enable multiple vendors to provide hardware and / or software for telecommunication systems. To this end, O-RAN divides RAN functions into a centralized unit (CU), a distributed unit (DU), and a radio unit (RU). The CU is a logical node for hosting the Radio Resource Control (RRC), Service Data Adaptation Protocol (SDAP), and / or Packet Data Convergence Protocol (PDCP) sublayers of the RAN. The DU is a logical node for hosting the Radio Link Control (RLC), Media Access Control (MAC), and Physical (PHY) sublayers of the RAN. The RU is a physical node that converts radio signals from the antenna into digital signals that can be transmitted to the DU via the fronthaul. These entities have open protocols and interfaces between them, allowing them to be developed by different vendors.

[0042] Figure 1 illustrates the O-RAN architecture of related technology. Referring to Figure 1, RAN functions in the O-RAN architecture are controlled and optimized by RICs. RICs are software-defined components that implement modular applications to facilitate the multi-vendor operability required in O-RAN systems. RICs also automate and optimize RAN operations. RICs are divided into two types: non-real-time RICs (NRT-RIC) and near-real-time RICs (nRT-RIC).

[0043] The NRT-RIC is the control point for non-real-time control loops and operates on sub-second timescales within a Service Management and Orchestration (SMO) framework. Its functions are implemented via modular applications called rApps (rApp 1, ..., rApp N) and include providing policy-based guidance and reinforcement over the A1 interface, which is the interface enabling communication between nRT-RICs and nRT RICs; performing data analytics, artificial intelligence / machine learning (AI / ML) training and inference for RAN optimization; and / or recommending configuration management actions over the O1 interface, which is the interface connecting the SMO to RAN management elements (e.g., nRT-RIC, O-RAN Centralized Unit (O-CU), O-RAN Distributed Unit (O-DU), etc.).

[0044] The nRT-RIC operates on a timescale of 10 milliseconds to 1 second and connects to the O-DU, O-CU (split into the O-CU Control Plane (O-CU-CP) and O-CU User Plane (O-CU-UP)), and Open evolved NodeB (O-eNB) via the E2 interface. The nRT-RIC uses the E2 interface to control the underlying RAN elements (E2 nodes / Network Functions (NFs)) via near-real-time control loops. The nRT-RIC monitors, pauses / stops, overrides, and controls E2 nodes (O-CU, O-DU, O-eNB) via policies. For example, the nRT-RIC sets policy parameters for activated functions in the E2 nodes. Additionally, the nRT-RIC hosts xApps to implement functions such as quality of service (QoS) optimization, mobility optimization, slicing optimization, interference mitigation, load balancing, and security. The two types of RICs work together to optimize O-RAN. For example, the NRT-RIC provides policies, data, and artificial intelligence / machine learning (AI / ML) models over the A1 interface that are implemented and used by the nRT-RIC to optimize the RAN, and the nRT-RIC returns policy feedback (i.e., how the policies set by the NRT-RIC are working).

[0045] The SMO framework, in which the NRT-RIC resides, manages and orchestrates RAN elements. Specifically, the SMO manages and coordinates what is called the O-Ran Cloud (O-Cloud). The O-Cloud is a collection of physical RAN nodes that host the RIC, O-CU, and O-DU, supporting software components (e.g., operating systems and runtime environments), and the SMO itself. In other words, the SMO manages the O-Cloud from within. The O2 interface is the interface between the SMO and the O-Cloud in which it resides. Through the O2 interface, the SMO provides infrastructure management services (IMS) and deployment management services (DMS).

[0046] On the other hand, the O-Cloud is a cloud computing platform that includes a collection of physical infrastructure nodes that meet the O-RAN requirements for hosting relevant O-RAN functions (e.g., nRT-RIC, O-CU-CP, O-CU-UP, O-DU, etc.), supporting software components (operating systems, virtual machine monitors, container runtimes, etc.), and appropriate management and orchestration functions.

[0047] The SMO framework in which the NRT-RIC resides manages and orchestrates RAN elements via four main interfaces: the A1 interface between the NRT-RIC and nRT-RIC within the SMO for RAN optimization, the O1 interface between the SMO and O-RAN network functions for FCAPS support, the open fronthaul (FH) M-Plane interface between the SMO and O-RU for FCAPS support in the hybrid model, and the O2 interface between the SMO and O-Cloud for platform resource and workload management.

[0048] 2 is a diagram of an example environment 200 in which the systems and / or methods described herein may be implemented. As shown in FIG. 2, environment 200 may include a user device 210, a platform 220, and a network 230. The devices in environment 200 may be interconnected via wired connections, wireless connections, or a combination of wired and wireless connections. In an embodiment, any of the functions and operations described with reference to FIG. 1 above may be performed by any combination of elements shown in FIG. 2.

[0049] User device 210 includes one or more devices capable of receiving, generating, storing, processing, and / or providing information related to platform 220. For example, user device 210 may include a computing device (e.g., a desktop computer, a laptop computer, a tablet computer, a handheld computer, a smart speaker, a server, etc.), a mobile phone (e.g., a smartphone, a wireless telephone, etc.), a wearable device (e.g., smart glasses or a smart watch), or a similar device. In some implementations, user device 210 may receive information from and / or transmit information to platform 220.

[0050] Platform 220 includes one or more devices capable of receiving, generating, storing, processing, and / or providing information. In some implementations, platform 220 may include a cloud server or a collection of cloud servers. In some implementations, platform 220 may be designed to be modular, such that particular software components can be swapped in or out depending on particular needs. Thus, platform 220 may be easily and / or quickly reconfigured for different uses.

[0051] In some implementations, as shown, platform 220 may be hosted in a cloud computing environment 222. Notably, although the implementations described herein describe platform 220 as being hosted within cloud computing environment 222, in some implementations platform 220 may not be cloud-based (i.e., may be implemented outside of a cloud computing environment) or may be partially cloud-based.

[0052] Cloud computing environment 222 includes an environment that hosts platform 220. Cloud computing environment 222 can provide services such as computing, software, data access, and storage, without requiring end-user (e.g., user device 210) knowledge of the physical location and configuration of the systems and / or devices that host platform 220. As shown, cloud computing environment 222 can include a collection of computing resources 224 (collectively referred to as “computing resources 224” and individually as “computing resource 224”).

[0053] Computing resources 224 include one or more personal computers, clusters of computing devices, workstation computers, server devices, or other types of computing and / or communication devices. In some implementations, computing resources 224 can host platform 220. Cloud resources may include compute instances running within computing resources 224, storage devices provided within computing resources 224, data transfer devices provided by computing resources 224, etc. In some implementations, computing resources 224 can communicate with other computing resources 224 via wired connections, wireless connections, or a combination of wired and wireless connections.

[0054] As further shown in FIG. 2, computing resources 224 include a collection of cloud resources, such as one or more applications (“APP”) 224-1, one or more virtual machines (“VM”) 224-2, virtualized storage (“VS”) 224-3, and one or more hypervisors (“HYP”) 224-4.

[0055] The application 224-1 includes one or more software applications that can be provided or accessed by the user device 210. The application 224-1 may eliminate the need for the software application to be installed and run on the user device 210. For example, the application 224-1 may include software associated with the platform 220 and / or any other software that can be provided via the cloud computing environment 222. In some implementations, one application 224-1 can send information to or receive information from one or more other applications 224-1 via the virtual machine 224-2.

[0056] Virtual machine 224-2 includes a software-implemented machine (e.g., a computer) that executes programs like a physical machine. Virtual machine 224-2 can be either a system virtual machine or a process virtual machine, depending on the application and the degree to which virtual machine 224-2 matches an actual machine. A system virtual machine may provide a complete system platform that supports the execution of a complete operating system (“OS”). A process virtual machine may execute a single program and support a single process. In some implementations, virtual machine 224-2 can run on behalf of a user (e.g., user device 210) and manage the infrastructure of cloud computing environment 222, such as data management, synchronization, or long-term data transfer.

[0057] Virtualized storage 224-3 includes one or more storage systems and / or one or more devices that use virtualization technology within the storage systems or devices of computing resources 224. In some implementations, in the context of storage systems, types of virtualization may include block virtualization and file virtualization. Block virtualization may refer to the abstraction (or separation) of logical storage from physical storage so that the storage system can be accessed regardless of the physical storage or heterogeneous structure. The separation may provide storage system administrators with flexibility in how they manage storage for end users. File virtualization may eliminate the dependency between data accessed at the file level and where the file is physically stored. This can enable performance optimization of storage usage, server consolidation, and / or nondisruptive file migration.

[0058] Hypervisor 224-4 may provide hardware virtualization technology that allows multiple operating systems (e.g., "guest operating systems") to run simultaneously on a host computer, such as computing resource 224. Hypervisor 224-4 may present a virtual operating platform to the guest operating systems and may manage the execution of the guest operating systems. Multiple instances of different operating systems may share virtualized hardware resources.

[0059] Network 230 may include one or more wired and / or wireless networks. For example, network 230 may include a cellular network (e.g., a fifth-generation (5G) network, a long-term evolution (LTE) network, a third-generation (3G) network, a code division multiple access (CDMA) network, etc.), a public land mobile network (PLMN), a local area network (LAN), a wide area network (WAN), a metropolitan area network (MAN), a telephone network (e.g., a public switched telephone network (PSTN)), a private network, an ad hoc network, an intranet, the Internet, an optical fiber-based network, etc., and / or a combination of these or other types of networks.

[0060] The number and arrangement of devices and networks shown in Figure 2 are provided as an example. In practice, there may be additional, fewer, different, or differently arranged devices and / or networks. Furthermore, two or more devices shown in Figure 2 may be implemented within a single device, or a single device shown in Figure 2 may be implemented as multiple distributed devices. Additionally, or alternatively, a set of devices (e.g., one or more devices) of environment 200 may perform one or more functions that are described as being performed by another set of devices in environment 200.

[0061] 3 is a diagram of example components of a device 300. The device 300 may correspond to a user device 210 and / or a platform 220. As shown in FIG. 3, the device 300 may include a bus 310, a processor 320, a memory 330, a storage component 340, an input component 350, an output component 360, and a communication interface 370.

[0062] The bus 310 includes components that enable communication between the components of the device 300. The processor 320 may be implemented in hardware, firmware, or a combination of hardware and software. The processor 320 may be a central processing unit (CPU), a graphics processing unit (GPU), an accelerated processing unit (APU), a microprocessor, a microcontroller, a digital signal processor (DSP), a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), or another type of processing component. In some implementations, the processor 320 includes one or more processors that can be programmed to perform functions. The memory 330 includes random access memory (RAM), read only memory (ROM), and / or another type of dynamic or static storage device (e.g., flash memory, magnetic memory, and / or optical memory) that stores information and / or instructions for use by the processor 320.

[0063] Storage component 340 stores information and / or software related to the operation and use of device 300. For example, storage component 340 may include a hard disk (e.g., a magnetic disk, optical disk, magneto-optical disk, and / or solid-state disk), a compact disc (CD), a digital versatile disc (DVD), a floppy disk, a cartridge, magnetic tape, and / or another type of non-transitory computer-readable medium, along with a corresponding drive. Input component 350 includes components that enable device 300 to receive information, such as via user input (e.g., a touchscreen display, a keyboard, a keypad, a mouse, buttons, switches, and / or a microphone). Additionally or alternatively, input component 350 may include sensors for sensing information (e.g., a global positioning system (GPS) component, an accelerometer, a gyroscope, and / or an actuator). Output component 360 includes components that provide output information from device 300 (e.g., a display, a speaker, and / or one or more light-emitting diodes (LEDs)).

[0064] The communication interface 370 includes transceiver-like components (e.g., a transceiver and / or a separate receiver and transmitter) that enable the device 300 to communicate with other devices via a wired connection, a wireless connection, or a combination of wired and wireless connections, etc. The communication interface 370 may enable the device 300 to receive information from another device and / or provide information to another device. For example, the communication interface 370 may include an Ethernet interface, an optical interface, a coaxial interface, an infrared interface, a radio frequency (RF) interface, a universal serial bus (USB) interface, a Wi-Fi interface, a cellular network interface, etc.

[0065] The device 300 may perform one or more processes described herein. The device 300 may perform these processes in response to the processor 320 executing software instructions stored by a non-transitory computer-readable medium, such as the memory 330 and / or the storage component 340. A computer-readable medium is defined herein as a non-transitory memory device. A memory device includes memory space within a single physical storage device or memory space spread across multiple physical storage devices.

[0066] Software instructions may be loaded into memory 330 and / or storage component 340 from another computer-readable medium or from another device via communications interface 370. The software instructions stored in memory 330 and / or storage component 340, when executed, may cause processor 320 to perform one or more processes described herein.

[0067] Additionally, or instead, hardwired circuitry may be used in place of or in combination with software instructions to implement one or more processes described herein. Thus, implementations described herein are not limited to any specific combination of hardware circuitry and software.

[0068] The number and arrangement of components shown in Figure 3 are provided as an example. In practice, device 300 may include additional, fewer, different, or differently arranged components than those shown in Figure 3. Additionally or alternatively, a set of components (e.g., one or more components) of device 300 may perform one or more functions that are described as being performed by another set of components of device 300.

[0069] In embodiments, any one of the operations or processes of Figures 4, 5, 6, 7, and 8 may be implemented by or using any one of the elements shown in Figures 1, 2, and 3. It will be appreciated that other embodiments are not limited thereto and may be implemented in a variety of different architectures (e.g., bare metal architectures, any cloud-based or deployment architectures such as Kubernetes, Docker, OpenStack, etc.).

[0070] FIG. 4 illustrates an SMO / NRT-RIC framework system architecture configured to implement advanced sleep modes, according to one embodiment.

[0071] 4, the NRT-RIC framework and the NRT-RIC within the NRT-RIC framework represent a subset of the functionality of the SMO framework. The NRT-RIC can access other SMO framework functions and thereby affect (i.e., control and / or execute) what is conveyed across the O1 and O2 interfaces (e.g., configuration management (CM) and / or performance management (PM) execution).

[0072] The SMO framework system architecture includes an SMO function that includes an O1 termination that enables communication between the SMO framework and E2 nodes (i.e., O-CU, O-DU, etc.) over the O1 interface.

[0073] The SMO framework system architecture includes an SMO function that includes an O2 termination that enables communication between the SMO framework and E2 nodes (i.e., O-CU, O-DU, etc.) over the O2 interface.

[0074] The NRT-RIC includes an NRT-RIC framework, which includes, among other functions, an R1 service exposure function that handles the provided R1 services according to an exemplary embodiment. In general, the NRT-RIC functions within the NRT-RIC framework support authorization, authentication, registration, discovery, communication support, etc. for rAPPs.

[0075] In general, R1 services may include a collection of services including, but not limited to, service registration and discovery services, authentication and authorization services, AI / ML workflow services, and A1, O1 and O2 interface related services.

[0076] An NRT-RIC application (rApp) is an application that leverages the functionality available in the NRT-RIC framework and / or SMO framework to provide value-added services related to RAN operation and optimization. The scope of rApps includes, but is not limited to, radio resource management, data analytics, etc., and information enrichment. In general, rApps refer to applications designed to consume and / or produce R1 services.

[0077] To this end, the NRT-RIC framework produces and / or consumes R1 services according to embodiments of the present invention via an R1 interface, which terminates at an R1 termination of the NRT-RIC framework. The R1 termination connects to the NRT-RIC framework and rApps via the R1 interface, allowing the NRT-RIC framework and rApps to exchange messages / data (i.e., requests and responses including data models) to access the R1 services via the R1 interface.

[0078] In general, the R1 interface is defined as the interface between an rApp and the NRT-RIC framework through which R1 services can be produced and consumed.

[0079] Additionally, the NRT-RIC framework includes A1-related functions, such as supporting A1 logical termination, A1 policy coordination and catalog, and A1-EI coordination and catalog.

[0080] In particular, the NRT-RIC framework and the A1-related functions therein provide A1 policies (i.e., sets of rules used to manage and control the change and / or maintenance of the state of one or more managed objects) based on guidance and enforcement over the A1 interface, which is an interface that enables communication between the NRT-RIC and the nRT-RIC. (i.e., A1 policies are a type of declarative policy expressed using formal descriptions that, according to related art, enable the NRT-RIC in the SMO to guide the nRT-RIC, and therefore the RAN, toward better realization of RAN intent (e.g., predetermined performance goals).)

[0081] Data management and exposure services within the NRT-RIC framework distribute data generated or collected by data producers to data consumers according to their needs (e.g., function management (FM) / consumption management (CM) / production management (PM) data to rApps, or CM changes from rApps to O-RAN via the O1 interface).

[0082] The NRT-RIC framework further comprises external terminations that support the exchange of data between the NRT-RIC framework and external AI / ML functions, Enrichment Information (EI) sources, or external monitoring, for example.

[0083] Within the NRT-RIC framework, AI / ML workflow services provide access to AI / ML workflows. For example, the AI / ML workflow services may assist in model training and monitoring of AI / ML models deployed to the NRT-RIC. As a result, the NRT-RIC framework and the AI / ML workflow services therein enable artificial intelligence and machine learning (AI / ML) training and inference for RAN optimization.

[0084] Additionally, the NRT-RIC framework provides A2 related functions, which support, for example, A2 logical termination, A2 policy coordination and catalogs, etc.

[0085] 4, the NRT-RIC framework (e.g., the NRT-RIC and / or at least one rApp hosted by the NRT-RIC framework) enables flexible parameters for temporarily suspending O-RU components (e.g., physical and functional components within the O-RU) within a cell or cluster of cells by providing an A1 policy over the A1 interface formulated by the NRT-RIC (e.g., by the NRT-RIC and / or at least one rApp hosted by the NRT-RIC framework assisted by machine learning (ML) techniques) toward the nRT-RIC. Via E2 interface actions, the nRT-RIC may implement deployment of the configured parameters for temporarily suspending O-RU components (e.g., physical and functional components within the O-RU) toward one or more E2 nodes.

[0086] 4, the SMO and NRT-RIC framework is configured to collect data (e.g., the SMO and NRT-RIC framework may include collection and control functions) to temporarily shut down one or more O-RU components, such as required cell configurations, performance indicators, measurement reports (e.g., cell load-related information and traffic information, energy efficiency EE and / or energy consumption EC measurement reports, geographic location information), etc. from E2 nodes (i.e., from E2 nodes such as O-CU, O-DU, etc.) and O-RUs (through E2 nodes or directly).

[0087] For example, data collected for temporarily shutting down one or more O-RU components may include relevant information of the corresponding O-RU: configuration, performance indicators, and measurement reports (e.g., cell load-related information and traffic information, energy efficiency EE and / or energy consumption EC measurement reports, geographic location information, etc.) that enable identification of the individual capabilities of the O-RU (e.g., the O-RU's internal system architecture, its performance, its functions, etc., for optimally controlling and / or configuring the O-RU components).

[0088] The collected data (e.g., measurement data for energy savings) may be used for (re)training and inference purposes of AI / ML models that support EE / ES functions to optimize energy efficiency EE and / or energy consumption EC towards nRT-RIC via an SMO / Non-RT RIC framework supported by machine learning (ML) techniques.

[0089] As mentioned above, in this specification, configuration, performance indicators, and measurement reports (i.e., data for performing temporary outage optimization of O-RU components) collected from E2 nodes (i.e., O-DU, O-CU, etc.) are intended to include relevant information of the corresponding O-RU components (e.g., O-RU capabilities of physical / functional components to implement advanced sleep modes).

[0090] According to an example embodiment, the data collected as described above may be collected from the O-RU via the E2 node via the O1-interface in accordance with a hierarchical O-RAN system architecture, and / or directly from the O-RU via the open FH M-Plane in accordance with a hybrid O-RAN system architecture.

[0091] For example, the data collected as described above may be collected from the O-RU via the E2 interface via the open FH M-Plane.

[0092] As a result, the SMO and NRT-RIC framework is configured to analyze data collected from E2 nodes (e.g., E2 nodes such as O-CU, O-DU, etc.) and / or O-RU to trigger advanced sleep mode (ASM).

[0093] Generally, energy efficiency EE is defined as the relationship between useful output and energy / electricity consumption, while energy consumption EC is defined as the integral of electricity consumption over time.

[0094] Further, the SMO and NRT-RIC framework is configured to formulate and / or provide optimization triggers, optimization targets (e.g., enabling traffic shaping to extend sleep mode at 50% peak power consumption), and A1 policies (e.g., intent-based policies) to the nRT-RIC (e.g., via the O1 interface and / or the A1 interface).

[0095] Additionally, the SMO and NRT-RIC framework may be configured to, for example, train, maintain, update, configure, etc., EE / ES AI / ML models within the NRT-RIC.

[0096] In an exemplary embodiment, the SMO framework (i.e., the NRT-RIC) may be configured to launch and deploy EE / ES AI / ML models within the nRT-RIC.

[0097] Further referring to FIG. 4, the E2 nodes (e.g., E2 nodes such as O-CU, O-DU, etc.) in FIG. 1 are configured to report required cell configurations, performance indicators and measurement reports (e.g., cell load-related and traffic information, EE / EC measurement reports), as well as advanced sleep mode O-DU and / or O-RU capabilities to the SMO via the O1 interface and to the nRT-RIC via the E2 interface and the O1 interface.

[0098] Furthermore, the E2 node (e.g., E2 node) is configured to apply ES functionality or execute advanced sleep mode on the O-RU based on a policy (E2 policy command) or recommendation (E2 control command) from the nRT-RIC, respectively.

[0099] Additionally, the E2 node (e.g., the E2 node) may be configured to perform traffic shaping (i.e., intelligent scheduling) to converge symbols and / or slots from the time domain to the frequency domain to extend outage periods of one or more O-RU components, and / or to perform actions such as, for example, adjusting remaining minimum system information (RMSI) broadcast intervals and common channels (e.g., physical downlink control channels), aligning discontinuous reception (DRX) cycles of user entities (UEs) within a cell, utilizing multi-carrier-based energy saving methods, etc.

[0100] For example, an O-RU as shown in FIG. 1 is configured to report advanced sleep mode related capability information to an O-DU via the M-Plane.

[0101] Additionally, in an exemplary embodiment, the O-RU is configured to perform a transition from an active state to a desired sleep mode (i.e., implement a temporary shutdown of the O-RU components based on the type of advanced sleep mode) as commanded by the E2 node (i.e., O-DU) via the CUS-Plane.

[0102] Furthermore, the O-RU is configured to report power consumption and sleep state related information to the E2 node (i.e., O-DU) and / or SMO framework via the open FH M-Plane in accordance with the hybrid O-RAN system architecture.

[0103] Referring to FIG. 4, the SMO framework and E2 node may extract input data to prepare and execute a method for implementing an advanced sleep mode in an O-RAN, where the input data may include per-cell and per-carrier load statistics (e.g., number of active users, average number of radio resource control (RRC) connections, average number of scheduled active users per transmission time interval (TTI), physical resource block (PRB) utilization, downlink DL and / or uplink UL cell / user throughput, precoding matrix indicator (PMI) reports and / or channel state information (CSI) reports, etc., among other O1-, O2-, E2-related data, etc., available in the O-RAN.

[0104] Additionally, the SMO framework and E2 node may also extract input data such as per-cell latency statistics (e.g., if Ultra-Reliable Low-Latency Communication (URLLC) slices are included, latency may be utilized as defined in the EE definition, TS 28.554), O-DU capabilities related to advanced sleep mode (e.g., O-DU capabilities due to a particular network scenario (e.g., traffic load), and network deployment related to advanced sleep mode) to prepare and execute methods for implementing advanced sleep mode in O-RAN.

[0105] Additionally, to prepare and execute a method for implementing an advanced sleep mode in an O-RAN, the O-RU may extract input data such as power consumption metrics, including, for example, average total and per carrier power consumption, power consumption by type of advanced sleep mode, among other power consumption metrics. Furthermore, the O-RU data may extract information about the O-RU in a supported advanced sleep mode along with (i.e., the O-RU's capabilities related to the advanced sleep mode) (e.g., the capabilities of the O-RU's physical and functional components related to the advanced sleep mode). For example, the information about the O-RU in a supported advanced sleep mode may include a sleep-to-active transition time (e.g., including the cell site / O-RU input power required for a particular energy efficiency (EE) KPI).

[0106] The extracted input data of the SMO framework, E2 node, and O-RU as described above may be collected by the NRT-RIC and / or nRT-RIC to process methods for implementing advanced sleep modes in the O-RAN.

[0107] Referring to FIG. 4 , communication between the NRT-RIC and the nRT-RIC for preparing and executing a method for implementing an advanced sleep mode in an O-RAN includes output data including O1-related configuration data (e.g., ES optimization triggers, ES optimization targets, etc.) and / or output data providing one or more A1 policies for executing the advanced sleep mode based on network operator (intent-based) policies, among other O1-related data, O2-related data, A1-related data, etc. available in the O-RAN.

[0108] Additionally, communications between the NRT-RIC and the E2 node (i.e., the E2 node) for preparing and executing a method for implementing an advanced sleep mode in an O-RAN include output data, such as E2 policy commands and / or E2 control commands, that provide, for example, policy and / or guidance regarding initiating traffic shaping (slot and / or symbol scheduling) to minimize active slot transmission time intervals (TTIs) and / or symbol TTIs, policy and / or guidance regarding performing a type of ASM for one or more O-RU components that references a hibernate sleep mode with expected latency greater than 10 ms, policy and / or guidance regarding configuration for common channel (i.e., PDCCH) and remaining minimum system information (RMSI) adjustments, and further from the E2 node to the O-RU.

[0109] To this end, the E2 policy commands and / or E2 control commands include policies and / or guidance regarding the initiation of advanced sleep modes to achieve an optimal balance between system performance and energy savings.

[0110] Furthermore, communication between an E2 node (i.e., an E2 node) and an O-RU to prepare and execute a method for implementing an advanced sleep mode in an O-RAN includes output data such as an advanced sleep mode execution command via a CUS-Plan. For example, the advanced sleep mode execution command may include an appropriate execution command that takes into account the capabilities of the O-RU and enables implementation of an advanced sleep mode that achieves an optimal balance between system performance and energy savings.

[0111] As a result, the O-RAN system architecture configured as in Figure 4 allows for an abstraction level of energy saving modes that take into account the capability data of the O-RU and / or O-DU to realize the implementation of one or more ASM types / levels, thereby optimizing the operational energy efficiency of the O-RU while maintaining a high level of network performance across the O-RAN.

[0112] FIG. 5 illustrates a flow diagram of a method for implementing an advanced sleep mode, according to one embodiment.

[0113] 5, a method for implementing an advanced sleep mode may include functions such as a service management and orchestration (SMO) framework, a near real-time radio access network (RAN) intelligent controller (nRT-RIC), a non-real-time radio access network (RAN) intelligent controller (NRT-RIC), and an NRT-RIC framework according to Figures 1 and 4. Furthermore, the above functions may be realized in a system for implementing an advanced sleep mode in an open radio access network (O-RAN), which may include a memory storing instructions and at least one processor configured to implement the near real-time radio access network (RAN) intelligent controller (nRT-RIC) and the service management and orchestration (SMO) framework, wherein the SMO framework includes a non-real-time radio access network (RAN) intelligent controller (NRT-RIC), and the at least one processor is configured to execute instructions in accordance with the method for implementing an advanced sleep mode of Figure 5.

[0114] In step 501, the SMO framework collects measurement data necessary to train an artificial intelligence / machine learning (AI / ML) model.

[0115] For example, required cell configurations, performance indicators, measurement reports (e.g., cell load-related information and traffic information, energy efficiency EE and / or energy consumption EC measurement reports, geographical location information), etc. from E2 nodes (i.e., from E2 nodes such as O-CU, O-DU, etc.) and from O-RU (via E2 nodes or directly).

[0116] In step 502, based on the collected measurement data, the NRT-RIC (e.g., retrains) the AI / ML model and deploys the AI / ML model to the nRT-RIC.

[0117] For example, the NRT-RIC retrieves the measurement data required to train artificial intelligence / machine learning (AI / ML) models from the SMO framework via the NRT-RIC framework, as shown in Figure 4.

[0118] In an exemplary embodiment, the input data for training the AI / ML model may include the following measurement data for monitoring the energy consumption and energy efficiency EC / EE of one or more E2 nodes and one or more O-RUs: Downlink Packet Data Convergence Protocol Service Data Unit (DL PDCP) SDU data volume per interface (DL data volume delivered from O-CU-UP to O-DU per public land mobile network (PLMN), per quality of service (QoS) level, per slice, per F1-U interface, per Xn-U interface, per X2-U interface); Uplink Packet Data Convergence Protocol Service Data Unit (UP PDCP) SDU data volume per interface (DL data volume delivered from O-CU-UP to O-DU per public land mobile network (PLMN), per quality of service (QoS) level, per slice, per F1-U interface, per Xn-U interface, per X2-U interface); SDU data volume (UL data volume delivered from O-CU-UP to O-DU per public land mobile network (PLMN), per quality of service (QoS) level, per slice, per F1-U interface, per Xn-U interface, per X2-U interface, reference signal received quality (RSRQ) measurements per synchronization signal block (SSB) per cell, reference signal received power (RSRP) measurements per SSB per cell, signal to interference and noise ratio (SINR) measurements per SSB per cell, energy consumption, power consumed by hardware components, transmit power, etc.

[0119] In step 503, the trained AI / ML model is launched in the nRT-RIC by the SMO framework.

[0120] In step 504, the NRT-RIC monitors energy optimization data for AI / ML model inference from the O-RU via the E2 node. For example, the NRT-RIC constantly (e.g., periodically or event-based) monitors the performance and energy consumption of the E2 node and the O-RU for inference. The performance and energy consumption (i.e., optimization data) may include data such as cell load-related data and traffic information, EE / EC measurement reports, and geographic location information.

[0121] In step 505, the SMO framework activates at least one advanced sleep mode (ASM) in the nRT-RIC. For example, the SMO may trigger the activation of at least one advanced sleep mode (ASM) in the nRT-RIC via the O1 interface. Alternatively, the NRT-RIC may provide a policy via the A1 interface that guides the nRT-RIC to apply EE / ES functions to activate at least one advanced sleep mode (ASM).

[0122] Further, in step 506, based on the activation of at least one ASM, the nRT-RIC collects data necessary for preparing and executing a temporary shutdown of one or more O-RU components via the E2 interface from the O-RU via the E2 node.

[0123] For example, the data required to prepare for and execute a temporary outage of one or more O-RU components may include E2 node data such as per-cell latency statistics (e.g., if an Ultra-Reliable Low Latency Communications (URLLC) slice is included, latency may be utilized as defined in the EE definition, TS 28.554), O-DU capabilities related to advanced sleep mode (e.g., O-DU capabilities due to a particular network scenario (e.g., traffic load), and network deployment related to advanced sleep mode).

[0124] Additionally, the data required to prepare for and execute a temporary outage of one or more O-RU components may include O-RU data such as, for example, power consumption metrics including, for example, average total and per carrier power consumption, power consumption by type of advanced sleep mode, among other power consumption metrics. Additionally, the O-RU data may extract information about the O-RU for supported advanced sleep modes (i.e., the O-RU's capabilities related to advanced sleep modes). For example, information about the O-RU for supported advanced sleep modes may include sleep-to-active transition times (e.g., including cell site / O-RU input power required for specific energy efficiency (EE) KPIs).

[0125] In step 507, the nRT-RIC evaluates the collected data necessary to prepare and execute a temporary shutdown of one or more O-RU components based on the invoked AI / ML model and at least one ASM. For example, based on the AI / ML inference in the nRT-RIC, the nRT-RIC considers the optimization policy provided in step 505 and analyzes the data collected in step 506 to prepare and execute a temporary shutdown of one or more O-RU components.

[0126] In step 508, based on the evaluation, the nRT-RIC requests the initiation of at least one ASM to the O-RU via the E2 node. For example, the nRT-RIC generates an E2 control command requesting the initiation of at least one ASM to the E2 node that complies with the capabilities of the O-RU and / or an E2 policy command for an energy saving (ES) function to be applied to the E2 node that complies with the capabilities of the O-RU. In this case, the nRT-RIC may transmit the E2 control command and / or the E2 policy command to the E2 node via the E2 interface.

[0127] For example, the E2 policy command and / or E2 control command may provide guidance and / or commands to implement an optimization policy that requests the E2 node to at least initiate an advanced sleep mode (ASM) by, for example, activating an intelligent scheduling function (e.g., slot and / or symbol scheduling) to converge symbols or slots, adjusting RMSI broadcast intervals and common channels, aligning DRX cycles of UEs in a cell, and utilizing multi-carrier-based energy saving methods to further extend the depth and frequency of the sleep mode (i.e., extending the inactive period of the O-RU by clearing (rescheduling) transmission symbols and / or slots in the time domain to transmission blocks in the frequency domain (i.e., resource blocks (RBs) for symbols and / or resource block groups (RBGs) for slots) according to various types / levels of the advanced sleep mode without affecting other transmissions to the O-RU).

[0128] According to an exemplary embodiment, the E2 policy command and / or E2 control command may provide, for example, guidance and / or commands to activate intelligent scheduling functions (e.g., slot and / or symbol scheduling) to minimize active slot TTI and / or symbol TTI, policy and / or guidance regarding the implementation of one or more O-RU components of a type of ASM that references a hibernate sleep mode with expected latency greater than 10 ms, policy and / or guidance regarding configuration for common channels (i.e., PDCCH) and remaining minimum system information (RMSI) adjustments.

[0129] To this end, the E2 policy commands and / or E2 control commands include policies and / or guidance regarding the initiation of advanced sleep modes to achieve an optimal balance between system performance and energy savings.

[0130] In step 509, based on the ASM initiation request, the O-RU implements at least one ASM. For example, the E2 node (i.e., the E2 node) may generate an advanced sleep mode execution command requesting implementation on the CUS-Plan. For example, the advanced sleep mode execution command may include appropriate execution commands that take into account the capabilities of the O-RU and enable implementation of an advanced sleep mode that achieves an optimal balance between system performance and energy savings.

[0131] As a result, the implemented type of ASM (i.e., among multiple types / levels of ASM) that takes into account O-RU and / or O-DU capability data along with data to temporarily shut down one or more O-RU components enables optimal operational energy efficiency of the O-RU while maintaining a high level of network performance across the O-RAN.

[0132] FIG. 6 illustrates a flow diagram of a method for collecting data to temporarily shut down one or more O-RU components based on activation of at least one ASM, according to another embodiment.

[0133] Referring to FIG. 6, in step 601, the nRT-RIC sends a data collection request for energy saving to the E2 node via the E2 interface.

[0134] In step 602, the E2 node receives a data collection request for energy savings from the nRT-RIC.

[0135] In step 603, the E2 node collects data for temporarily suspending one or more O-RU components from the O-RU via the open FH M-Plane interface.

[0136] In step 604, the E2 node transmits the collected data to the nRT-RIC via the E2 interface to temporarily deactivate one or more O-RU components.

[0137] As a result, the method of FIG. 6 enables the collection of data necessary to temporarily shut down one or more O-RU components extracted from input data by the SMO framework, providing an E2 node such as that shown in FIG. 4 with optimal precision to evaluate and request the initiation of a preferred type of advanced sleep mode for optimal energy efficiency of the O-RU's operation while maintaining a high level of network performance throughout the O-RAN.

[0138] FIG. 7A is a flow diagram of a method for evaluating collected data to temporarily shut down one or more O-RU components, according to one embodiment.

[0139] Referring to FIG. 7A, in step 701A, the nRT-RIC receives collected data for temporarily suspending one or more O-RU components from the E2 node via the E2 interface.

[0140] In step 702A, based on the AI / ML model inference, the nRT-RIC generates at least one E2 control command to request the initiation of at least one ASM to an E2 node that complies with the capabilities of the O-RU.

[0141] In step 703A, the nRT-RIC sends at least one E2 control command to the E2 node (eg, via the E2 interface).

[0142] FIG. 7B is a flow diagram of a method for evaluating collected data to temporarily shut down one or more O-RU components, according to another embodiment.

[0143] Referring to FIG. 7B, in step 701B, the nRT-RIC receives collected data for temporarily suspending one or more O-RU components from the E2 node via the E2 interface.

[0144] In step 702B, based on the AI / ML model inference, the nRT-RIC generates at least one E2 policy command for energy saving (ES) functions to be applied to E2 nodes that comply with the capabilities of the O-RU.

[0145] In step 703B, the nRT-RIC transmits at least one E2 policy command to the E2 node (eg, via the E2 interface).

[0146] As a result, according to the embodiment shown in Figures 7A and 7B, based on evaluating the collected data to temporarily shut down one or more O-RU components, the AI / ML model inference in the nRT-RIC provides one or more E2 control commands and / or at least one E2 policy command to the E2 node to enable optimal operating energy efficiency of the O-RU while maintaining a high level of network performance throughout the O-RAN.

[0147] 8 illustrates a flow diagram of a method for implementing an advanced sleep mode, according to one embodiment. Referring to FIG. 8, in step 801, based on the implementation, the NRT-RIC receives implementation feedback including a performance analysis of the AI / ML model via the SMO framework.

[0148] In step 802, the NRT-RIC analyzes the performance of the AI / ML model on the nRT-NIC.

[0149] In step 803, the NRT-RIC determines that at least one predetermined performance goal has not been achieved based on the performance of the AI / ML model.

[0150] In step 805, the NRT-RIC initiates a fallback mechanism associated with at least one predetermined performance goal.

[0151] For example, the NRT-RIC initiates a fallback mechanism that may include updating (e.g., via the O1 interface and / or the A1 interface) optimization triggers, optimization targets (e.g., traffic shaping allows for extended sleep mode at 50% peak power consumption), and A1 policies (e.g., intent-based policies) to the nRT-RIC.

[0152] As a result, the method for implementing the advanced sleep mode according to Figure 8 provides a determination to initiate an associated fallback mechanism (e.g., to use a more appropriate advanced sleep mode) based on an analysis of the performance of the AI / ML model in the NRT-NIC to at least one predetermined performance target. The initiation of the fallback mechanism (i.e., closed-loop control) enables adjustment of the ASM type / level to respond to changing O-RAN conditions in a closed-loop manner to achieve optimal operating energy efficiency of the O-RU while maintaining a high level of network performance across the O-RAN.

[0153] FIG. 9 illustrates multiple types of advanced sleep modes (i.e., ASM types / levels) according to one embodiment. Referring to FIG. 9, various types of (advanced) sleep modes (ASMs) can be implemented based on the type of control method for shutting down O-RU components and the shutdown timescale. These ASMs can cover most of the O-RU equipment. (i.e., O-RU components, such as physical and functional components of the O-RU, can be temporarily shut down.)

[0154] Furthermore, some multi-vendor O-RAN FH-based O-RUs may not necessarily need to support all types of sleep modes, but the O-RU specifications (O-RU capabilities) may need to enable various types of advanced sleep modes (i.e., the O-RU capabilities must support the types of control methods and outage timescales used by multiple types of (advanced) sleep modes (ASMs) to enable their implementation).

[0155] For this purpose, the nNRT-RIC and / or O-DU may acquire prior knowledge of the O-RU specifications (O-RU capabilities) in order to apply the ASMs supported by the O-RU.

[0156] As a result, the O-DU can execute (apply) various types of ASMs in the O-RU, for example, via the CUS-Plane, which indicates (messages) the number of slots in which the O-RU (i.e., O-RU component) can go to sleep.

[0157] According to a first embodiment, a first type of ASM can enable a microsleep mode (i.e., ASM 1) with a symbol-to-slot-based duration (i.e., temporary shutdown of O-RU components within a symbol-to-slot-based duration). For example, ASM 1 may be initiated (e.g., guided) from the nRT-RIC to the O-DU. For this purpose, the nRT-RIC may send E2 control commands and / or E2 policy commands to the O-DU, and the O-DU may instruct the O-RU to implement ASM 1.

[0158] According to ASM 1 (i.e., the micro sleep mode), when the E2 node (e.g., gNB) does not need to operate the TX / RX information within the next few orthogonal frequency division multiplexing (OFDM) symbols (i.e., truncated OFDM (TOFDM)-symbols <T≦slot [ms]), the E2 node may communicate the ability to apply ASM 1 (i.e., the micro sleep mode) to the nRT-RIC, and the nRT-RIC transmits E2 control commands and / or E2 policy commands to the E2 node, and for example, the E2 node instructs the O-RU that the radio frequency (RF) transceiver can be turned off for a predetermined period as described above (e.g., when restarting the O-RU and / or O-DU after power-on and / or maintenance). Alternatively, the nNRT-RIC provides an E2 policy command to guide the O-DU, and the O-DU instructs the O-RU according to the E2 policy command.

[0159] For this purpose, the sleep duration of ASM 1 (i.e., the micro sleep mode) is based on the O-RU capabilities that support the minimum granularity of the sleep interval (i.e., the finest subdivision of the stop state in the symbol-to-slot-based duration). The minimum granularity of the sleep interval requires the highest level of O-RU capabilities from the O-RU components to support the micro sleep mode.

[0160] As a result, due to the capabilities of the O-RU that support the micro sleep mode, the O-DU can make more information-based decisions (e.g., decisions that result in better energy efficiency) regarding the application of intelligent scheduling functions (e.g., traffic shaping such as data convergence of symbols and / or slots from the time domain to the frequency domain) to save energy.

[0161] ASM 1 (i.e., microsleep mode) has the advantage that the O-RU's ability to temporarily shut down O-RU components for symbol-to-slot-based durations, in accordance with the O-RU's and O-DU's ability to detect and respond to subtle changes in traffic patterns, allows the O-DU to take more precise energy-saving measures rather than relying on broad, blanket policies for energy conservation.

[0162] According to a second embodiment, another type of ASM can enable a light sleep mode (i.e., ASM 2) with an L-slot-based duration (1 slot to 10 ms) (i.e., temporary shutdown of O-RU components within an L-slot-based duration). For example, the light sleep mode (i.e., ASM 2) may be initiated (e.g., guided) from the nRT-RIC to the O-DU. For this purpose, the nRT-RIC may send E2 control commands and / or E2 policy commands to the O-DU. According to the light sleep mode (i.e., ASM 2), if the E2 node (e.g., gNB) does not need to operate TX / RX information within the next L slots (e.g., Slot ≦ T ≦ 10 ms), the E2 node may communicate the ability to apply the light sleep mode (i.e., ASM 2) to the nRT-RIC. The nRT-RIC sends E2 control commands and / or E2 policy commands to the E2 node, for example, instructing the O-RU that the E2 node can turn off its radio frequency (RF) transceiver and additional hardware components (i.e., other physical O-RU components) for a predetermined period as described above. Alternatively, the nNRT-RIC provides E2 policy commands that guide the O-DU, and the O-DU instructs the O-RU according to the E2 policy commands.

[0163] For this purpose, the sleep duration of ASM 2 (i.e., light sleep mode) is based on the O-RU capability level (i.e., O-RU capability to accommodate temporary shutdown of O-RU (physical) components) that supports sleep durations (shutdown intervals) from 1 slot to 10 ms.

[0164] As a result, in ASM 2 (ie, light sleep mode), the power consumption of the O-RU components is typically low compared to the power consumption in micro sleep mode (ie, ASM 1).

[0165] Furthermore, according to an exemplary embodiment, if there are no synchronization signal blocks (SSBs) on the downlink (DL) or physical random access channel (PRACH), or if there are no other uplink (UL) signals to be transmitted / received, the light sleep mode may be implemented by transmission blanking.

[0166] According to a third embodiment, another type of ASM can enable a deep sleep mode (i.e., ASM 3) with an M-slot-based duration (100 ms from a radio frame) (i.e., temporary shutdown of O-RU components within an M-slot-based duration). For example, the deep sleep mode (i.e., ASM 3) may be initiated (e.g., guided) from the nRT-RIC to the O-DU and O-RU. For this purpose, the nRT-RIC may send E2 control commands and / or E2 policy commands to the O-DU, and the E2 node may provide an O-RU implementation instruction command to the O-RU.

[0167] According to the deep sleep mode (i.e., ASM 3), if the E2 node (e.g., gNB) does not need to operate TX / RX information within the next M slots (e.g., 10 ms≦T≦100 ms), the E2 node may communicate the ability to apply deep sleep (i.e., ASM 3) to the nRT-RIC. The nRT-RIC sends an E2 control command and / or an E2 policy command to the E2 node, and the E2 node instructs the O-RU to implement the deep sleep mode (i.e., ASM 3). Alternatively, the nRT-RIC provides an E2 policy command to guide the O-DU, and the O-DU instructs the O-RU according to the E2 policy command.

[0168] For example, by initiating the deep sleep mode (i.e., ASM 3), additional hardware components (i.e., other physical O-RU components) can be turned off for the period specified above. However, according to the deep sleep mode (i.e., ASM 3), some hardware components of the O-RU, such as timing circuits, must be kept in standby or active to allow the O-RU to quickly transition to the active state (i.e., roll back immediately). The power consumption of the deep sleep mode (i.e., ASM 3) is lower than that of the light sleep mode (i.e., ASM 2).

[0169] The sleep duration of the deep sleep mode (i.e., ASM 3) may be based on the O-RU's ability to support a sleep duration (i.e., outage duration) of up to 100 ms from the radio frame. Due to a relatively long outage duration, the deep sleep mode (i.e., ASM 3) requires coordination with other neighboring cells because the user experience is affected when a cell is turned off during the outage duration according to ASM 3. Therefore, the O-DU may need to schedule and send an O-RU implementation instruction command to the O-RU.

[0170] According to a fourth embodiment, another type of ASM can enable a hibernate sleep mode (i.e., ASM 4) with an N-slot-based duration (e.g., 100 ms to several seconds) (i.e., temporary shutdown of O-RU components within an N-slot-based duration). For example, the type of ASM that enables hibernate sleep (i.e., ASM 4) may be initiated (e.g., guided) from the nRT-RIC to the O-DU and O-RU. For this purpose, the nRT-RIC may send E2 control commands and / or E2 policy commands to the O-DU, and the E2 node may provide an O-RU implementation instruction command to the O-RU.

[0171] According to the hibernate sleep mode (i.e., ASM 4), if the E2 node (e.g., gNB) does not need to operate TX / RX information within the next N slots (e.g., 100 ms to several seconds), the E2 node may communicate its ability to apply deep sleep (i.e., ASM 3) to the nRT-RIC. The nRT-RIC sends E2 control commands and / or E2 policy commands to the E2 node, which instructs the O-RU to implement the hibernate sleep mode (i.e., ASM 4).

[0172] For example, by initiating hibernate sleep mode (i.e., ASM 4), most of the hardware components (i.e., almost all physical O-RU components) can be turned off for a predetermined period of time as described above. However, according to hibernate sleep mode (i.e., ASM 4), some hardware components of the O-RU, such as timing circuits, must be kept in standby or active to allow the O-RU to quickly transition to the active state (i.e., roll back immediately). The power consumption of hibernate sleep mode (i.e., ASM 4) is low compared to the power consumption of deep sleep mode (i.e., ASM 3), and sleep durations of more than 1000 ms are not excluded.

[0173] The sleep duration for hibernate sleep mode (i.e., ASM 4) may be based on the O-RU's ability to support a sleep duration (i.e., outage duration) between 100 ms and several seconds. Due to a relatively long outage duration, hibernate sleep mode (i.e., ASM 4) requires coordination with other neighboring cells because user experience is affected when a cell is turned off for the outage duration according to ASM 4. Therefore, the O-DU may need to schedule and send an O-RU implementation instruction command to the O-RU.

[0174] Referring to FIG. 9, the sleep level (power saving) increases from level 1 to level 4 according to ASM 1 to ASM 4.

[0175] 10 illustrates a method for scheduling at least one symbol to minimize the number of symbols in the time domain according to one embodiment, which may be associated with an advanced sleep mode, for example, as shown in FIG.

[0176] 10, symbols may be allocated in resource blocks (RBs) across various frequency bands in the frequency domain. For example, symbols may be allocated across RB0 to RBn in the frequency domain. Furthermore, in the time domain, symbols may be allocated over time (i.e., symbol duration D, such as symbol TTI). For example, symbols may be allocated across symbol durations (e.g., D0 to D13), each having a predetermined duration (e.g., symbol TTI).

[0177] To minimize the number of symbols in the time domain (eg, symbols in a symbol sequence from D0 to D13), the symbols may be rearranged to resource blocks (eg, RB0 to RBn of each frequency band).

[0178] For example, symbols for the physical downlink control channel (PDCCH) are allocated across RB0 to RB7 in the frequency domain.

[0179] Furthermore, the symbols of the physical downlink shared channel (PDSCH) are allocated to three resource blocks (RBs) over a symbol sequence according to D0 to D13. For energy saving, the active time of the O-RU component can be shortened by converging the symbols from the time domain to the frequency domain. For this purpose, the symbol sequence D0 to D13 of the PDSCH symbols can be relocated in the frequency domain, for example, to RB0 to RB6. As a result, for the time domain sequences of PDSCH and PDCCH, six symbols can be cleared (i.e., the scheduled transmission can be emptied) for the remaining seven consecutive symbols for PDCCH and PDSCH broadcasting. As a result, the six cleared symbols in the time domain without data can be used as an energy saving slot (i.e., a slot in which one or more O-RU components are stopped).

[0180] As a result, the convergence of at least one symbol from the time domain to the frequency domain extends the stop period of one or more O-RU components. The extension of the stop period enables the E2 node (e.g., gNB) to more frequently recognize a state where the E2 node does not need to operate the TX / RX information within the number of orthogonal frequency division multiplexing (OFDM) symbols (i.e., truncated OFDM (TOFDM)-symbols <T ≦ ES slot [ms]) to which it is reallocated. This has the advantage that the E2 node can communicate, for example, the ability to apply ASM 1 and more frequently initiate a temporary stop of the O-RU component as shown in FIG. 9 while maintaining a high level of network performance across the O-RAN and saving energy.

[0181] FIG. 11 shows a method for scheduling at least one slot to minimize the number of slots in the time domain according to one embodiment.

[0182] Referring to FIG. 11, slots may be allocated to resource block groups (RBGs) across various frequency bands in the frequency domain. For example, slots may be allocated across RBG0 to RBGn in the frequency domain. Furthermore, in the time domain, slots may be allocated over time (i.e., slot time periods such as slot TTI). For example, slots may be allocated across slot durations (e.g., Slot 0 to Slot 10), each having a predetermined duration (e.g., slot TTI).

[0183] To minimize the number of slots in the time domain (eg, slots in a slot sequence by slot 0 to slot 10), the slots may be rearranged into resource block groups (eg, RBG0 to RBGn for each frequency band).

[0184] According to one example, synchronization signal block-master information block (SSB-MIB) slots are scheduled to be transmitted in RBG1-RBG2 in slot 0. System information block type 1 (SIB1) slots are scheduled to be transmitted in RBG3-RBG5 in slot 1. Physical downlink shared channel (PDSCH) slots are scheduled to be transmitted in RBG1 and RBG2 in slots 2 and 3.

[0185] Without being limited to this example, the concept of converging slots from the time domain to the frequency domain may be illustrated, for example, by a physical downlink shared channel (PDSCH) slot scheduled to be transmitted in slots 2 and 3 (i.e., otherwise empty slots 2 and 3). In this case, the slots may be converged (i.e., rescheduled) to be transmitted in RBG1-RBG2 in slots 4 and 5. Slots 4 and 5 have no transmission load in RBG1-RBG2 (i.e., RBG1 and RBG2 are empty in slots 4 and 5). As a result, slots 2 and 3 are cleared from transmission load in the time domain as well as the frequency domain and may be used for ASM activation (i.e., temporary deactivation of one or more O-RU components).

[0186] Additionally, other S1 block slots may be scheduled to be transmitted in RBG3-RBG4 in slots 4 and 5. These slots are not affected by the rescheduling described above.

[0187] In a second example, the concept of converging slots from the time domain to the frequency domain may be illustrated by, for example, a physical downlink shared channel (PDSCH) slot scheduled to be transmitted on RBG 5 over the sequence of slots 7 through 10. Slots 7 and 8 have no transmission load other than the PDSCH broadcast slot. To clear slots 7 and 8, the PDSCH slots may be rescheduled (i.e., converged) to RBG 2 and RBG 3 in slots 9 and 10, respectively. As a result, slots 7 and 8 are cleared of transmission load in both the time and frequency domains and may be used for ASM activation (i.e., temporary deactivation of one or more O-RU components).

[0188] 10 and 11, symbol scheduling and slot scheduling can be complemented by various traffic shaping techniques (i.e., intelligent scheduling functions). These intelligent scheduling functions are applied as shown in FIGS. 10 and 11 to make symbol scheduling and / or slot scheduling more flexible (i.e., adjusting the duration (symbol or slot duration) of a transmission to clear (empty) a maximum number of symbols or slots in the time domain without affecting other transmissions scheduled in parallel RBs or RBGs, respectively).

[0189] In a first traffic shaping example (i.e., according to a first intelligent scheduling function), a base station (e.g., O-RU) may dynamically adjust the number of paging frames (PFs) transmitted within a paging cycle (e.g., a paging cycle is referred to as a time interval T in a standard PF). For example, when the O-RAN operates under low or no load (i.e., O-RAN utilization is low), the number of PFs within a paging cycle may be reduced. Alternatively, when the network load is normal (i.e., O-RAN utilization is normal), the number of PFs within a paging cycle may be increased (rolled back).

[0190] As a result, dynamic adjustment of paging frames has the advantage of reducing the number of PFs in a paging cycle, thereby reducing the symbols and slots in the time domain, thereby facilitating the rescheduling (convergence) of symbols and slots from the time domain to the frequency domain, thereby clearing (emptying) the maximum number of symbols or slots in the time domain without affecting other transmissions scheduled in parallel RBs or RBGs.

[0191] In a second traffic shaping example (i.e., according to a second intelligent scheduling function), the base station (e.g., O-RU) may dynamically adjust the SSB periodicity. To this end, the O-DU and / or O-RU dynamically adjust the SSB periodicity between 5 ms and 120 ms based on the traffic load and requirements of performance-related KPIs such as latency and user throughput.

[0192] As a result, dynamic adjustment of the SSB period can provide an optimal duration that fits each RB and / or RBG, converging (rescheduling) SSB transmissions from each symbol and / or slot in the time domain to RBs and / or RBGs of other symbols or / or slots, thereby clearing (emptying) a maximum number of symbols or slots in the time domain without affecting other transmissions scheduled in other RBs or RBGs scheduled in parallel.

[0193] In a third traffic shaping example (i.e., according to a third intelligent scheduling function), a base station (e.g., an O-RU) may dynamically adjust the remaining minimum system information (RMSI). To this end, for example, the O-RU may dynamically adjust the periodicity of the RMSI and / or other system information (SI) based on traffic load and performance-related KPI requirements such as latency and user throughput.

[0194] As a result, by dynamically adjusting the period of RMSI and / or other system information (SI), an optimal period for transmitting RMSI and / or other system information (SI) can be provided based on O-RAN utilization, and a place can be reserved for symbols and slots to be rescheduled to respective RBs and / or RBGs in order to converge (reschedule) respective symbols and / or slots in the time domain to RBs and / or RBGs of other symbols or / or slots, thereby clearing (emptying) a maximum number of symbols or slots in the time domain without affecting other transmissions scheduled in parallel RBs or RBGs.

[0195] In a fourth traffic shaping example (i.e., according to a fourth intelligent scheduling function), a base station (e.g., an O-RU) may dynamically adjust the common channel periodicity. To this end, for example, the O-RU may dynamically adjust the periodicity of the physical random access channel (PRACH) and / or other common channels, such as the physical uplink control channel (PUCCH), (e.g., channel state information (CSI) and other reporting information), based on traffic load and performance-related KPI requirements such as latency and user throughput.

[0196] As a result, by dynamically adjusting the periodicity of RMSI and / or other system information (SI), an optimal periodicity for transmitting RMSI and / or other system information (SI) can be provided based on O-RAN utilization, reserving space for symbols and / or slots to be rescheduled to respective RBs and / or RBGs in order to converge (reschedule) respective symbols and / or slots in the time domain to RBs and / or RBGs of other symbols or / or slots, thereby clearing (emptying) a maximum number of symbols or slots in the time domain without affecting other transmissions scheduled in parallel RBs or RBGs.

[0197] In a fifth traffic shaping example (i.e., according to a fifth intelligent scheduling function), a base station (e.g., an O-RU) may dynamically adjust uplink scheduling requests (SRs). To this end, for example, the O-RU may dynamically adjust an uplink SR period based on ES policies and user performance requirements.

[0198] As a result, by dynamically adjusting the uplink SR period, the period for transmitting uplink SRs can be optimized, reserving space for symbols and / or slots to be rescheduled to respective RBs and / or RBGs in the time domain in order to converge (reschedule) each symbol and / or slot to the RBs and / or RBGs of other symbols or / or slots, thereby clearing (emptying) the maximum number of symbols or slots in the time domain without affecting other transmissions scheduled in parallel RBs or RBGs.

[0199] In a sixth traffic shaping example (i.e., according to a sixth intelligent scheduling function), a base station (e.g., an O-RU) may dynamically synchronize the discontinuous reception (DRX) cycles of multiple user entities (UEs) within a cell. To this end, the O-DU scheduling mechanism synchronizes the DRX cycles of multiple user entities (UEs) to minimize the O-RU awake time (i.e., the maximum level of empty symbols / slots available for deep sleep modes such as ASM 3). As a result, synchronizing the DRX cycles of multiple user entities (UEs) enables a minimum active time and optimal energy efficiency for the O-RU (i.e., the O-RU component) with extended outages of the O-RU component.

[0200] In a seventh traffic shaping example (i.e., according to a seventh intelligent scheduling function), a base station (e.g., an O-RU) may shift traffic of one or more carriers of the O-RU to one or more other carriers according to a multi-carrier scenario. For this purpose, the O-DU scheduling mechanism may apply the multi-carrier scenario to each of the one or more other carriers to extend the outage period of the O-RU components scheduled by the O-RU. As a result, the O-DU scheduling mechanism enables a maximum sleep mode duration and transition to a deeper sleep mode (e.g., from deep sleep mode (ASM 3) to hibernate sleep mode (ASM 4)).

[0201] As a result, the intelligent scheduling function (e.g., traffic shaping examples 1 through 7) implements the deepest (optimal) ASM level possible, allowing for the longest outage of one or more O-RU components and the lowest level of granularity of sleep intervals (i.e., the finest subdivision of outage states in symbol- to slot-based durations). This has the advantage of optimizing the abstraction level of the energy saving modes, enabling optimal operating energy efficiency of the O-RUs, while maintaining a high level of network performance across the O-RAN.

[0202] 12A-16, a method for implementing advanced sleep mode (ASM) in an O-RAN enables an ASM energy saving function in the O-RAN. To this end, the nRT-RIC applies configuration changes to the O-DU and / or O-RU components based on the ASM energy saving function and generates control actions and policy settings (e.g., E2 control commands and / or E2 policy commands) using an AI / ML-based solution. For example, the nRT-RIC provides AI / ML-based E2 policy commands to the E2 node (i.e., O-DU), and the O-DU translates the policies into O-RU implementation commands to instruct the O-RU to enforce (i.e., apply and implement) the ASM configuration provided by the nRT-RIC.

[0203] Additionally, the nRT-RIC controls the application of ASMs to the O-DU and / or O-RU. The E2 node and O-RU configure their respective components (e.g., physical and functional O-RU components) and implement control actions (e.g., E2 control commands, O-RU implementation commands) and enforcement policies (e.g., E2 policy commands).

[0204] The SMO and NRT-RIC framework develops policies to trigger and guide ES optimization in the nRT-RIC.

[0205] 12A-16, a method for implementing advanced sleep mode (ASM) begins when the O-RAN becomes operational. (That is, the method begins when the A1 interface connection between the NRT-RIC and the nRT-RIC, the SMO, the O1 interface connection between the NRT-RIC and the E2 node, the E2 interface connection between the nRT-RIC and the E2 node, and the open FH M-Plane interface connection between the upper layer network functions (SMO, NRT-RIC) and the lower layer network functions (O-DU and / or O-RU) are established.)

[0206] For example, an open FH M-Plane interface connection may be established between the E2 node and the O-RU (i.e., an open FH M-Plane connection with a hierarchical O-RAN architecture) or between the O-RU and the SMO (i.e., an open FH M-Plane connection with a hybrid O-RAN architecture).

[0207] Furthermore, depending on the O-RAN conditions under which it operates, the nRT-RIC has knowledge of overlapping carriers / cells and their coverage (e.g., which carriers / cells are coverage layers and which carriers / cells are capacity layers). Furthermore, depending on the O-RAN conditions under which it operates, the nRT-RIC obtains (performance) targets (e.g., performance policies / goals) as provided by the NRT-RIC framework and / or the SMO framework.

[0208] To this end, the network operator may set performance targets for the ES functions in the NRT-RIC (i.e., predetermined performance parameters for the EE / ES in the O-RAN, e.g., one or more predetermined performance targets for the EE / EC in the O-RAN). The targets in the NRT-RIC may be formulated in the respective A1 policies.

[0209] As a result, the method for implementing ASM may begin when a network operator enables collection and control capabilities within the SMO framework (i.e., optimization rApp) along with an initial AI / ML model.

[0210] Furthermore, based on the performance target, the nRT-RIC obtains capability data of the O-DU and / or O-RU components to support the implementation of ASM to achieve the above performance target.

[0211] 12A illustrates a method for collecting data via an O1 interface in a hierarchical O-RAN architecture according to one embodiment. Referring to FIG. 12A, operations 1-5 illustrate data collection for a method for implementing ASM in a hierarchical O-RAN architecture. In operation 1, a collection and control function within an SMO framework (e.g., rApp) requests collection of measurement data for training to an AI / ML model (e.g., NRT-RIC) via the O1 interface.

[0212] In operation 2, upon receiving a request from the SMO framework and / or the NRT-RIC framework, an E2 node (i.e., an E2 node such as an O-CU, an O-DU, etc.) requests measurement data from the O-RU via the open FH M-Plane.

[0213] In operation 3, the O-RU transmits measurement data to the E2 node via the open FH M-Plane interface. In one embodiment, the measurement data may include status and / or capability data of the O-RU for formulating targets for ES functionality with ASM capabilities.

[0214] In operation 4, the E2 node (i.e., O-CU, O-DU, etc.) sends measurement data, such as the configuration data and configured measurement data described above, to the SMO framework via the O1 interface.

[0215] In operation 5, the NRT-RIC retrieves the measurement data to train the AI / ML model within the NRT-RIC.

[0216] 12A, the SMO initiates measurement data collection requests to the E2 node and the O-RU for AI / ML model training for energy saving optimization. This has the advantage that specific measurement data can be extracted from the O-RU, for example, upon initialization and reconfiguration of the O-RU (e.g., after an update, maintenance, etc.).

[0217] 12B illustrates a method for data collection according to another embodiment. Referring to FIG. 12B, operations 1-3 illustrate alternative data collection for a method for implementing ASM in a hybrid O-RAN architecture. In operation 1, the SMO requests the O-RU directly via the open FH M-Plane to collect measurement data for AI / ML model training.

[0218] In operation 2, upon receiving a data collection request from the SMO framework, the O-RU sends measurement data to the SMO framework via the open FH M-Plane. In one embodiment, the optimization data may include status and / or capability data of the O-DU and / or O-RU for formulating performance targets including ASM functionality. For example, the SMO framework may receive measurement data from the O-RU periodically or on an event basis.

[0219] In operation 3, the NRT-RIC retrieves the measurement data to train the AI / ML model within the NRT-RIC.

[0220] 12, according to an exemplary embodiment, the E2 node (i.e., O-DU) and O-RU transmit configured measurement data to the SMO for NRT-RIC processing on a periodic or event-based basis. The periodic communication of configured measurement data based on NRT-RIC processing has the advantage that the NRT-RIC can periodically adjust performance targets (e.g., A1 policies) or train AI / ML models in the NRT-RIC based on the measurement data.

[0221] 13 shows an AI / ML model training flow in the NRT-RIC according to one embodiment. Referring to FIG. 13, in operation 1, the NRT-RIC trains an AI / ML model.

[0222] In operation 2, the NRT-RIC deploys the trained AI / ML model to the nRT-RIC, and the SMO framework launches the AI / ML model in the nRT-RIC via the O1 and A1 interfaces.

[0223] In operation 3, the NRT-RIC constantly monitors the performance of the trained AI / ML model (i.e., the NRT-RIC analyzes the performance of the retrained AI / ML model). For example, the NRT-RIC constantly monitors the performance and energy consumption of the E2 node and O-RU, such as cell load and traffic-related information, EE / EC measurement reports, and geolocation information, for AI / ML model inference.

[0224] 14 illustrates a method for initiating an ASM according to one embodiment. Referring to FIG. 13, in operation 1, the SMO may trigger the initiation of an advanced sleep mode in the NRT-RIC via the O1 interface as a performance target / goal.

[0225] Alternatively, in operation 2, the NRT-RIC may provide, via the A1 interface, a policy that guides the nRT-RIC to apply the EE / ES function in advanced sleep mode.

[0226] FIG. 15 illustrates a method for invoking an ASM according to one embodiment.

[0227] 15, in operations 1 to 7, when the nRT-RIC receives either a performance target via the O1 interface or an A1 policy via the A1 interface, the nRT-RIC considers the optimization policy (performance trigger) via the O1 interface or the A1 policy via the A1 interface based on AI / ML inference in the nRT-RIC. Furthermore, based on evaluation of measurement data and data (such as O-DU and / or O-RU capability data) to temporarily suspend one or more O-RU components, the nRT-RIC may request the E2 node to prepare and execute an AMS (i.e., generate an E2 control command and / or an E2 policy command).

[0228] In operation 1, for example, based on policy guidance from the nRT-RIC, the nRT-RIC requests measurement data from the E2 node via the E2 interface.

[0229] In operation 2, the E2 node requests the O-RU to provide measurement data (eg, including data for temporarily shutting down one or more O-RU components).

[0230] In operation 3, the O-RU transmits measurement data (e.g., including data for temporarily stopping one or more O-RU components) to the E2 node. The E2 node transmits measurement data (e.g., including data for temporarily stopping one or more O-RU components and capability data of the O-RU and / or O-DU) to the nNRT-RIC.

[0231] In operation 5, the nRT-RIC evaluates the data collected from the E2 node in operation 4 based on AI / ML inference in the nRT-RIC (taking into account the optimization policy) (i.e., by analyzing the traffic and implementing techniques such as traffic shaping for data convergence in symbols or slots, adjusting RMSI broadcast intervals and common channels, and aligning the DRX cycles of UEs in the cell).

[0232] In operation 6, the nRT-RIC requests the E2 node to enter Advanced Sleep Mode (ASM). To this end, the nRT-RIC may also send control actions and / or policies (E2 control commands and / or E2 policy commands) to direct user traffic to different carriers to further extend the depth and frequency of ASM.

[0233] In operation 7, based on the ASM initiation request, the E2 node (i.e., the E2 node) sends an O-RU implementation command to the O-RU. Upon receiving the O-RU implementation command, the O-RU implements at least one ASM (i.e., an ASM type).

[0234] In operation 8, as shown in operation 2 of FIG. 13, the NRT-RIC monitors and analyzes the performance of the AI / ML model (e.g., the NRT-RIC may constantly monitor and analyze the performance of the AI / ML model). For example, the NRT-RIC monitors the performance and energy consumption of the E2 node, the energy consumption of the O-RU, etc.

[0235] In an example embodiment, the input data used in AI / ML model training may include the following measurement data for monitoring energy consumption and energy efficiency EC / EE of one or more E-Nodes and one or more O-RUs: Downlink Packet Data Convergence Protocol Service Data Unit (DL PDCP) SDU data volume per interface (DL data volume delivered from O-CU-UP to O-DU per public land mobile network (PLMN), per quality of service (QoS) level, per slice, per F1-U interface, per Xn-U interface, per X2-U interface), Uplink Packet Data Convergence Protocol Service Data Unit (UP PDCP) per interface SDU data volume (UP data volume delivered from O-CU-UP to O-DU per public land mobile network (PLMN), per quality of service (QoS) level, per slice, per F1-U interface, per Xn-U interface, per X2-U interface, reference signal received quality (RSRQ) measurements per synchronization signal block (SSB) per cell, reference signal received power (RSRP) measurements per SSB per cell, signal to interference and noise ratio (SINR) measurements per SSB per cell, energy consumption, power consumed by hardware components, transmit power, etc.

[0236] In operation 9, based on the monitoring, the NRT-RIC may determine that a predetermined performance goal is not being achieved, in which case the rApp initiates at least one fallback mechanism associated with the predetermined performance goal.

[0237] For example, if the performance of the AI / ML model is below performance (not achieving a predetermined performance target), the NRT-RIC may modify the A1 policy. To this end, the NRT-RIC may send the modified A1 policy to the NRT-RIC. For example, the NRT-RIC may update or delete the A1 policy of the ASM energy function and send the modified A1 policy to the NRT-RIC.

[0238] Alternatively, the NRT-RIC may initiate an AI / ML model update and / or AI / ML model retraining based on predetermined performance goals received by the SMO framework.

[0239] The method for implementing the advanced sleep mode may terminate when the E2 node becomes non-operational or when the operator disables the SMO framework. However, as long as the SMO framework is enabled, the nRT-RIC continues to monitor the ES functionality in the E2 node and the O-RU, and the E2 node and the O-RU operate using updated parameters / models (i.e., implement the most recent commands from the nRT-RIC, including the most recent E2 control commands and / or the most recent E2 policy commands to prepare and execute ASM).

[0240] As a result, with the implementation of the latest commands from the nRT-RIC, including the latest E2 control commands and / or the latest E2 policy commands to prepare and execute the ASM while taking into account the capability data of the O-RU and / or O-DU, the applied ASM type / level enables optimal energy efficiency of the operation of the O-RU while maintaining a high level of network performance across the O-RAN.

[0241] Use cases according to one or more exemplary embodiments are described below. [Table 1]

[0242] 7. Advanced Sleep Mode

[0243] Editor's note: This chapter includes a study of potential impacts and enhancements to O-RAN interfaces, counters, and KPIs in the case of advanced sleep modes.

[0244] 7.1 Problem Statement, Solution, and Value Proposition

[0245] The power consumption of a radio base station (BS) varies depending on the amount of cell traffic. As traffic increases, the power amplifier (PA) becomes the main energy consumer. However, in low-traffic scenarios, the main energy consumption comes from the processing unit. Even when no transmission is taking place, parts of the BS continue to consume energy, providing an opportunity to reduce unnecessary energy consumption by shutting down unused components.

[0246] Advanced Sleep Mode (ASM) allows the BS to enter a low-power state when not in use, directly saving energy through intelligent shutdown of O-RU subcomponents. The decision to implement ASM is a complex task that balances system performance and energy savings. The Near-RT RIC is responsible for configuring cell parameters such as SSB periodicity and must consider multiple factors, such as traffic load, user service type, and energy efficiency measurements, to make an informed decision on whether to implement ASM. The impact of ASM on O-DU and O-CU energy consumption is flexible and depends on the specific scenario and network deployment. The goal is to achieve energy savings while maintaining a high level of network performance.

[0247] 7.1.1 Background

[0248] The solution involves implementing short DRX (Discontinuous Reception) sleep cycles, long DRX sleep cycles, and state transitions in the gNodeB (gNB), similar to how they are currently implemented in the UE (User Equipment). The sleep state transitions are those of the O-RU that are initiated by the O-DU based on policy or control provided by the Non-RT or Near-RT RIC.

[0249] Deeper sleep modes can offer higher energy savings because more devices can be turned off during longer transition times. However, it is important to understand that deeper sleep modes also have higher transition times to move from active to sleep and vice versa, so a balance must be struck between energy savings and network user performance.

[0250] 7.1.2 Advanced Sleep Mode Types

[0251] Based on the control methods and time scales that define the four sleep modes, this should be sufficient to cover the majority of devices.

[0252] It should be noted that an O-RAN fronthaul-based O-RU does not necessarily need to support all sleep modes, but the specification must enable all possible sleep modes.

[0253] The O-DU can implement the sleep mode in the O-RU via the CUS-Plane, indicating the number of slots in which the O-RU can go to sleep. [Table 2] TIFF2026500240000004.tif157157

[0254] 7.1.3 Intelligent Scheduling for Energy Savings Related to Advanced Sleep Modes

[0255] 7.1.3.1 Data Convergence at Symbols

[0256] To optimize the transmission of PDSCH data, the base station can use a technique known as "frequency domain scheduling." This technique allows the base station to group the PDSCH data to be transmitted onto fewer time-domain resources but spread them over more frequency-domain resources. This increases the number of symbols without data transmission, resulting in more efficient use of available resources and higher data rates.

[0257] Figure 7.1.3-1, Data Convergence at Symbol

[0258] This function supports the extension of SM1 (symbol ~ slot).

[0259] 7.1.3.2 Data Convergence in Slots

[0260] By concentrating the scheduling of PDSCH data in the time domain, the base station can increase the number of symbols without data transmission, but also increases the scheduling delay to some extent.

[0261] One way to do this is to pack data from multiple users into the fewest slots possible, improving energy efficiency, while accepting increased user data latency.

[0262] Transmitting PDSCH data in specific slots, such as Master Information Block (MIB), System Information Block 1 (SIB1), Other System Information (OSI), or paging, may also be one way to conserve the number of transmission opportunities.

[0263] Figure 7.1.3-2, Data convergence in slots

[0264] This function supports the extension of SM1 (symbol ~ slot).

[0265] 7.1.3.3 Dynamic SSB period adjustment

[0266] Dynamically adjusting the SSB period between 5ms and 120ms based on traffic load and performance KPI requirements such as latency and user throughput allows for deeper sleep modes with longer sleep durations, thus saving more energy.

[0267] It should be noted that the multiple long SS block periods that enable deep mode can negatively impact user experience. Devices performing cell reselection and handover may experience delays in RRM measurements based on the SS blocks of NR cells. Furthermore, because the default period assumed by devices is 20 ms, they have a very low probability of detecting cells with SS block periods higher than 20 ms. In LTE systems, this is not an issue because PSS / SSS are always transmitted periodically, every 5 ms.

[0268] However, delays can be avoided thanks to the sparse synchronization raster supported by NR. Because many frequency bands with wide bandwidths can be deployed in NR, devices must spend an unreasonable amount of time on the cell search process along all potential carrier locations. This is much worse than in LTE when there are multiple beams on both the network and UE sides that must be acquired and paired. The sparse synchronization raster specifically defines the locations in the frequency domain where devices must search for SS blocks. These locations do not always coincide with the center frequency, as in LTE. However, as soon as a device detects an SS block, it receives all the system information necessary to establish a connection to the cell from the PBCH and SIB1, including information about the subcarrier spacing used for transmission.

[0269] Furthermore, 5G NR SA cells with a longer than default SS block period of 20 ms can be discovered for cell reselection in idle / inactive mode if the correct SMTC is indicated in the serving cell's system information (SI). Two SMTCs for idle / inactive mode are required, coexisting in the network in this case: one for SA cells with a 20 ms SS block period that can be detected in initial cell search, and another for sleeper cells with a higher SS block period than 20 ms that can be detected during cell reselection.

[0270] This feature supports extended deep sleep mode as well as hibernate sleep mode.

[0271] 7.1.3.4 Dynamic RMSI Adjustment

[0272] The period of RMSI and other SIs is dynamically adjusted based on traffic load and performance KPI requirements such as latency and user throughput.

[0273] This feature supports extended deep sleep mode as well as hibernate sleep mode.

[0274] 7.1.3.5 Dynamic Adjustment of Paging Frames

[0275] The base station adjusts the frequency at which it transmits paging frames (PFs) based on network traffic. When traffic is low, the base station transmits fewer PFs per paging cycle (PC) to increase the number of symbols without data transmission. When traffic returns to normal levels, the base station restores the original frequency of transmitting PFs.

[0276] This feature supports all ASM extensions.

[0277] 7.1.3.6 Dynamic Adjustment of Common Channel Period

[0278] Dynamically adjust the periodicity of PRACH, and other common channels (CSI and other reporting) such as PUCCH, based on traffic load, signaling load, and performance KPI requirements such as latency, jitter, user throughput, packet error rate, and access delay.

[0279] This feature supports all ASM extensions.

[0280] 7.1.3.7 Dynamic adjustment of uplink scheduling requests

[0281] The uplink scheduling request (SR) period is dynamically adjusted based on ES policy and user performance requirements.

[0282] The purpose is to enable a deeper sleep mode by increasing the SR period for the UE.

[0283] This feature supports all ASM extensions.

[0284] 7.1.3.8 Synchronizing UE DRX cycles

[0285] The UE DRX periods are synchronized to minimize the network awake time, which can be done via the O-DU scheduling mechanism.

[0286] This feature supports all ASM extensions.

[0287] 7.1.3.9 Intelligent Scheduling in Multi-Carrier Scenarios

[0288] Shifting users' traffic to other carriers in a multi-carrier scenario to achieve maximum sleep duration with a minimum number of cells can also result in improved energy efficiency of the network.

[0289] This feature supports all ASM extensions.

[0290] 7.2 Architecture / Deployment Options

[0291] 7.1.4 Solution 1: ASM with Near-RT RIC Deployment

[0292] 7.1.4.1 Description and UML Diagrams [Table 3] TIFF2026500240000006.tif159157

[0293] @startuml skin rose skinparam defaultFontSize 15 autonumber Box "Service Management & Orchestration Framework" #gold Participant “Collection & Control” as smo Participant "Non-RT RIC" as NRTRIC End box Box "O-RAN Nodes" #lightpink Participant "Near-RT RIC" as RTRIC Participant "E2-Nodes" as E2NODES Participant "O-RUs" as ORUs End box autonumber 1.1 Group Data Collection alt via O1 smo -> E2NODES : < <o1>> Data collection request for Energy Saving E2NODES -> ORUs : < <fh>> Data collection request Else via OFH-MP smo -> ORUs : < <ofh-mp>> Data collection request end autonumber 2.1 alt Via O1 ORUs -> E2NODES : < <fh>> Measurement Data Collection E2NODES -> smo : < <o1>> Measurement Collection for Energy Saving Else via OFH-MP ORUs -> smo : < <fh>> Measurement Data Collection E2NODES -> smo : < <o1>> Measurement Collection for Energy Saving End autonumber 3 smo -> NRTRIC : < <o1>> Data Retrieval end autonumber 4.1 group AI / ML workflow NRTRIC -> NRTRIC : AI / ML Model training NRTRIC -> NRTRIC : Monitoring & Analysis of Energy Efficicincy \n & Consumption (E2 Nodes & O-RU)(s NRTRIC -> RTRIC : < <o1> > or < <o2>> Deploy AI / ML model end autonumber 5.1 group Optimization Trigger and Policy alt via O1 smo -> RTRIC : < <o1>> Optimization Trigger / Target else via A1 NRTRIC --> RTRIC : < <a1>> Intent based Policy end autonumber 6.1 group Actor Data Collection & Decision Making RTRIC -> E2NODES : < <e2>> Data collection request for Energy Saving E2NODES -> ORUs : < <fh>> Data collection request for Energy Saving ORUs -> E2NODES : < <fh>> Measurement Data Collection for Energy Saving E2NODES -> RTRIC : < <e2>> Measurement Data Collection for Energy Saving RTRIC -> RTRIC: AI / ML model inference RTRIC -> E2NODES: < <e2>> Advance Sleep Modes and related energy savings methods \nenforecement for Energy saving E2NODES -> ORUs: < <fh>> Updated O-RU Configurations autonumber 6.1 group AI / ML workflow NRTRIC -> NRTRIC : Performance analysis of AI / ML model \n(with possible actions, e.g. fallback, re-training) NRTRIC -> RTRIC : < <o1> > or < <o2>> Update AI / ML model end @enduml

[0294] Figure 7.1.1-1, Advanced Sleep Mode Function Flow

[0295] 7.1.4.2 Roles of O-RAN entities

[0296] 1) SMO (including non-RT RIC)

[0297] Collecting necessary cell configurations, performance indicators and measurement reports (e.g., cell load and traffic related information, EE / EC measurement reports and geographic location information) from E2 nodes and O-RUs for the purpose of training AI / ML models supporting EE / ES functions.

[0298] Trigger and run EE / ES AI / ML model training / retraining.

[0299] Analyzes data received from E2 nodes and O-RUs to trigger advanced sleep mode functions

[0300] Provides optimization triggers, optimization targets, and A1 policies (e.g., traffic shaping allows for extended sleep mode at 50% peak power consumption) to the Near-RT RIC via the O1 or A1 interface.

[0301] 2) Near-RT RIC

[0302] Collects necessary cell configurations, performance indicators and measurement reports (e.g., cell load related and traffic information, EE / EC measurement reports) from E2 nodes and O-RUs.

[0303] Receive EE / ES AI / ML models for deployment via O1.

[0304] Receives EE / ES related configuration management via the O1 interface and / or policies via the A1 interface for consideration during optimization.

[0305] Analyzes data received from E2 node and performs AI / ML model inference to determine advanced sleep mode actions for EE / ES (e.g., traffic shaping for data convergence in symbols and / or slots, adjusting RMSI and common channel broadcast intervals, aligning DRX cycles of UEs in a cell) to be performed taking into account optimization targets / policies.

[0306] Provide policies and / or necessary information via the E2 interface to trigger actions for EE / ES optimization.

[0307] 3) E2 node

[0308] The required cell configuration, performance indicators and measurement reports (e.g., cell load related and traffic information, EE / EC measurement reports), as well as advanced sleep mode capabilities are reported to the SMO via the O1 interface and to the Near-RT RIC via the E2 or O1 interface.

[0309] Implementing advanced sleep mode for O-RU based on policy or recommendation from Near-RT RIC.

[0310] Performing actions such as shaping for data convergence in symbols or slots, adjusting RMSI broadcast intervals and common channels, aligning DRX cycles of UEs within a cell, and utilizing multi-carrier based energy saving methods to increase the frequency and depth of O-RU sleep.

[0311] 4) O-RU

[0312] Reports capability information regarding advanced sleep modes to the O-DU via the M-Plane.

[0313] Transition from active to desired sleep mode commanded by O-DU via CUS-Plane.

[0314] Reports power consumption and sleep state related information to O-DU or SMO via M-Plane.

[0315] 7.1.4.3 Input / Output Data Requirements

[0316] 7.1.4.3.1 Summary

[0317] Input data

[0318] 1) SMO and E2 nodes

[0319] Per cell and per carrier load statistics such as number of active users, average number of RRC connections, average number of scheduled active users per TTI, PRB utilization, DL / UL cell / user throughput, PMI / CSI reporting

[0320] Latency statistics per cell (when URLLC slices are included, latency is used in the EE definition, TS 28.554)

[0321] O-DU capabilities related to advanced sleep modes

[0322] 2) O-RU

[0323] Power consumption metrics: Total power consumption and average power consumption per carrier

[0324] Power consumption per sleep mode

[0325] O-RU information for supported advanced sleep modes along with sleep-to-active transition time (site / O-RU input power required for specific EE KPIs)

[0326] Output Data

[0327] 1) Non-RT RIC to Near-RT RIC

[0328] O1 configuration (i.e., ES optimized trigger / target) OR

[0329] A1 policy for implementing advanced sleep modes based on operator policy

[0330] 2) Near-RT RIC to E2 node

[0331] Policy and guidance on invoking traffic shaping to minimize active symbols and slots (TTIs)

[0332] Policy and guidance for performing hibernate sleep (due to expected latency greater than 10ms)

[0333] Policy and guidance for common channel and RMSI configuration

[0334] 3) From E2 node to O-RU

[0335] a. Appropriate sleep mode execution command via CUS-Plane

[0336] 7.2 Impact Analysis for O-RAN Working Group

[0337] 7.3 Relationship and Impact to 3GPP (registered trademark) Specifications

[0338] 7.4 Gain analysis

[0339] 7.5 Feasibility Analysis

[0340] 7.5.1 Impact on Continuous Operation During Advanced Sleep Mode

[0341] ASM limited to the symbol level is not expected to have a significant impact on user performance. However, longer sleep modes, which involve shutting down more components or reducing their activity for longer periods of time, may have a significant impact on user performance, as they may cause service degradation such as increased latency or reduced data rates.

[0342] 7.5.2 Impact on Coverage

[0343] If the O-RU is performing deep sleep mode by adjusting the SSB period over 20 ms in NR, the UE may not be able to connect to the gNB due to missing SSB measurements, and therefore coverage within the area may be affected. However, this can be addressed with a properly configured sparse synchronization raster when multiple carriers are present.

[0344] 7.5.3 Impact and Implications for Vendor-Specific Scheduling and Beamforming Algorithms

[0345] Advanced sleep modes can impact cell and user performance.

[0346] It is up to the proprietary scheduler algorithms to handle such events most efficiently. Scheduling (e.g., user selection, resource allocation), adaptive SU-MIMO and MU-MIMO (e.g., MIMO modes, spatial streams and layers), common and shared channel scheduling are coordinated by the base station according to policies set by the Near-RT RIC.

[0347] 7.5.4 Limited O-RU / O-DU Capabilities

[0348] O-RU and O-DU may have limitations in implementing sleep mode capabilities. Even if an O-RU supports ASM functionality, it is unlikely to support all ASMs through a comprehensive specification framework. This may cause multi-vendor interoperability issues.

[0349] Figure 7.1.3.1: Data convergence at the previously referenced symbols is according to Figure 10.

[0350] Figure 7.1.3-2: Data convergence in the slots referenced above is according to Figure 11.

[0351] FIG. 7.1.1-1: The pre-sleep mode function flow referred to above is according to FIGS. 12A, 12B and 13 to 15.

[0352] The foregoing disclosure provides illustration and description, but is not intended to be exhaustive or to limit implementations to the precise form disclosed. Modifications and variations are possible in light of the above disclosure or may be acquired from practice of the implementations.

[0353] Some embodiments may relate to systems, methods, and / or computer-readable media at any possible level of technical detail. Furthermore, one or more of the above components described above may be implemented as instructions stored on a computer-readable medium and executable by at least one processor (and / or may include at least one processor). The computer-readable medium may include computer-readable non-transitory storage medium(s) having computer-readable program instructions for causing a processor to perform operations.

[0354] A computer-readable storage medium may be a tangible device capable of retaining and storing instructions for use by an instruction execution device. A computer-readable storage medium may be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of computer-readable storage media includes portable computer diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanically encoded devices such as punch cards or ridge-in-groove structures with instructions recorded thereon, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium should not be construed as a transitory signal itself, such as an electric wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse passing through a fiber optic cable), or an electrical signal transmitted through a wire.

[0355] The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to each computing / processing device or to an external computer or storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network may include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface within each computing / processing device receives the computer-readable program instructions from the network and transfers the computer-readable program instructions for storage in a computer-readable storage medium within the respective computing / processing device.

[0356] The computer-readable program code / instructions for carrying out operations may be either source code or object code written in any combination of one or more programming languages, including assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state-setting data, configuration data for integrated circuits, or object-oriented programming languages ​​such as Smalltalk, C++, and procedural programming languages ​​such as the "C" programming language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry including, for example, a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA) can execute computer-readable program instructions by utilizing state information of the computer-readable program instructions to personalize the electronic circuitry to perform aspects or operations.

[0357] These computer-readable program instructions may be provided to a processor of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, executed by the processor of the computer or other programmable data processing apparatus, form means for performing the functions / acts specified in the flowchart and / or block diagram blocks. These computer-readable program instructions may also be stored on a computer-readable storage medium that can instruct a computer, programmable data processing apparatus, and / or other device to function in a particular manner, such that the computer-readable storage medium having instructions stored therein comprises an article of manufacture containing instructions that implement aspects of the functions / acts specified in one or more blocks of the flowchart and / or block diagram.

[0358] The computer-readable program instructions may also be loaded into a computer, other programmable data processing apparatus, or other device to cause the computer, other programmable apparatus, or other device to perform a series of operational steps to produce a computer-implemented process, such that the instructions executing on the computer, other programmable apparatus, or other device perform the functions / acts specified in one or more blocks of the flowcharts and / or block diagrams.

[0359] The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer-readable media according to various embodiments. In this regard, each block in a flowchart or block diagram may represent a portion of a microservice, module, segment, or instruction set, which includes one or more executable instructions for implementing the specified logical function(s). The methods, computer systems, and computer-readable media may include additional, fewer, different, or differently arranged blocks than those shown in the figures. In some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may actually be executed concurrently or substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. It should also be noted that each block of the block diagrams and / or flowchart diagrams, and combinations of blocks in the block diagrams and / or flowchart diagrams, may be implemented by a special-purpose hardware-based system that performs the specified functions or actions or executes a combination of special-purpose hardware and computer instructions.

[0360] It will be apparent that the systems and / or methods described herein may be implemented in different forms of hardware, firmware, or a combination of hardware and software. The actual dedicated control hardware or software code used to implement these systems and / or methods is not limiting of the implementation. Thus, the operation and behavior of the systems and / or methods are described herein without reference to specific software code. It should be understood that software and hardware can be designed to implement the systems and / or methods based on the description herein.

[0361] Various further respective aspects and features of embodiments of the present disclosure can be defined by the following clauses: Item [1] A system for implementing an advanced sleep mode in an open radio access network (O-RAN), the system comprising: a near-real-time radio access network (RAN) intelligent controller (nRT-RIC); and a service management and orchestration (SMO) framework, the SMO framework comprising a non-real-time radio access network (RAN) intelligent controller (NRT-RIC), the system: collecting measurement data for training an artificial intelligence / machine learning (AI / ML) model by the SMO framework; training an AI / ML model by the NRT-RIC based on the collected measurement data and deploying the AI / ML model to the nRT-RIC; activating the trained AI / ML model in the nRT-RIC by the SMO framework; and transmitting the AI / ML model to an open radio unit (O-RAN) via an E2 node by the NRT-RIC. monitoring energy optimization data for AI / ML model inference from an O-RU (Operating-Responsive Unit); activating, by the SMO framework, at least one Advanced Sleep Mode (ASM) in the nRT-RIC; collecting, by the nRT-RIC, data to temporarily shut down one or more O-RU components from the O-RU via the E2 node based on the activation of the at least one ASM; evaluating, by the nRT-RIC, the collected data to temporarily shut down one or more O-RU components based on the activated AI / ML model and the at least one ASM; requesting, by the nRT-RIC, via the E2 node to the O-RU to start the at least one ASM based on the evaluation; and implementing, by the O-RU, the at least one ASM based on the ASM start request. Item [2] The system is configured to collect the data to temporarily stop the one or more O-RU components based on the activation of the at least one ASM by: sending a data collection request for energy saving to the E2 node via the E2 interface by the nRT-RIC; receiving the data collection request for energy saving from the nRT-RIC by the E2 node; collecting the data to temporarily stop the one or more O-RU components from the O-RU via the open FH M-Plane interface by the E2 node; and sending the collected data to the nRT-RIC via the E2 interface by the E2 node to temporarily stop the one or more O-RU components. Item [3] The system is configured to evaluate the collected data to temporarily stop the one or more O-RU components by: receiving the collected data from the E2 node by the nRT-RIC to temporarily stop the one or more O-RU components; applying AI / ML model inference by the nRT-RIC based on the collected data to temporarily stop the one or more O-RU components; generating at least one E2 control command by the nRT-RIC to request the E2 node to start the at least one ASM that complies with the capabilities of the O-RU based on the AI / ML model inference; and sending the at least one E2 control command to the E2 node by the nRT-RIC. The system described in Item [1 or 2] is configured to evaluate the collected data to temporarily stop the one or more O-RU components by: receiving the collected data from the E2 node by the nRT-RIC to temporarily stop the one or more O-RU components; applying AI / ML model inference by the nRT-RIC based on the collected data to temporarily stop the one or more O-RU components; Item [4] The system is configured to evaluate the collected data to temporarily stop the one or more O-RU components by: receiving the collected data from the E2 node by the nRT-RIC to temporarily stop the one or more O-RU components; applying AI / ML model inference by the nRT-RIC based on the collected data to temporarily stop the one or more O-RU components; generating at least one E2 policy command for energy saving (ES) by the nRT-RIC based on the AI / ML model inference to guide the E2 node in compliance with the O-RU's capabilities to initiate the at least one ASM; and transmitting the at least one E2 policy command to the E2 node by the nRT-RIC. The system described in Item [1 or 2] is configured to evaluate the collected data to temporarily stop the one or more O-RU components by: receiving the collected data from the E2 node by the nRT-RIC to temporarily stop the one or more O-RU components; applying AI / ML model inference by the nRT-RIC based on the collected data to temporarily stop the one or more O-RU components; Item [5] The system is further configured to: receive feedback including a performance analysis of the AI / ML model by the NRT-RIC via the SMO framework based on the implementing; analyze the performance of the AI / ML model in the nRT-NIC by the NRT-RIC; determine, based on the performance of the AI / ML model, that at least one predetermined performance goal has not been achieved by the NRT-RIC; and initiate, by the NRT-RIC, a fallback mechanism associated with the at least one predetermined performance goal. The system described in any one of items [1 to 4]. Item [6] The system is configured to implement the at least one ASM by scheduling at least one symbol to minimize the number of symbols in the time domain by the E2 node or the O-RU based on the capability of the O-RU, and the E2 node or the O-RU is configured to perform the scheduling by converging the at least one symbol from the time domain to the frequency domain to extend a suspension period of the one or more O-RU components, wherein the converged at least one symbol is at least one of a physical downlink channel (PDSCH) symbol and a physical downlink control channel (PDCCH) symbol. Item [7] The system is configured to implement the at least one ASM by: scheduling at least one slot by the E2 node or the O-RU to minimize the number of slots in the time domain based on the capability of the O-RU, and the E2 node or the O-RU is configured to perform the scheduling by converging the at least one slot from the time domain to the frequency domain to extend the outage period of the one or more O-RU components, wherein the converged at least one slot is at least one of a synchronization signal block - master information block (SSB-MIB) slot, a system information block type 1 (SIB1) slot, an SI slot, a paging frame slot, and a physical downlink channel (PDSCH) slot. Item [8] A method for implementing an advanced sleep mode in an open radio access network (O-RAN), the method comprising: collecting measurement data for training an artificial intelligence / machine learning (AI / ML) model by a service management and orchestration (SMO) framework; training an AI / ML model by a non-real-time radio access network (RAN) intelligent controller (NRT-RIC) based on the collected measurement data and deploying the AI / ML model to a near-real-time radio access network (RAN) intelligent controller (nRT-RIC); activating the trained AI / ML model in the nRT-RIC by the SMO framework; and receiving energy for AI / ML model inference from an open radio unit (O-RU) via an E2 node by the NRT-RIC. monitoring optimization data; activating, by the SMO framework, at least one advanced sleep mode (ASM) in the nRT-RIC; collecting, by the nRT-RIC, data to temporarily suspend one or more O-RU components from the O-RU via the E2 node based on the activation of the at least one ASM; evaluating, by the nRT-RIC, the collected data to temporarily suspend one or more O-RU components based on the activated AI / ML model and the at least one ASM; requesting, by the nRT-RIC, via the E2 node to the O-RU to activate the at least one ASM based on the evaluation; and implementing, by the O-RU, the at least one ASM based on the ASM activation request. Item [9] The method described in Item [8], wherein collecting the data to temporarily stop the one or more O-RU components based on the activation of the at least one ASM includes: sending, by the nRT-RIC, a data collection request for energy saving to the E2 node via the E2 interface; receiving, by the E2 node, the data collection request for energy saving from the nRT-RIC; collecting, by the E2 node, the data to temporarily stop the one or more O-RU components from the O-RU via the open FH M-Plane interface; and sending, by the E2 node, the collected data to the nRT-RIC via the E2 interface to temporarily stop the one or more O-RU components. Item

[10] The method described in Item [8 or 9], wherein evaluating the collected data to temporarily stop the one or more O-RU components includes: receiving, by the nRT-RIC, the collected data to temporarily stop the one or more O-RU components from the E2 node; applying, by the nRT-RIC, AI / ML model inference based on the collected data to temporarily stop the one or more O-RU components; generating, by the nRT-RIC, at least one E2 control command to request the E2 node to start at least one ASM that complies with the capabilities of the O-RU based on the AI / ML model inference; and transmitting, by the nRT-RIC, at least one of the E2 control commands to the E2 node. Item

[11] The method described in Item [8 or 9], wherein evaluating the collected data to temporarily stop the one or more O-RU components includes: receiving, by the nRT-RIC, the collected data to temporarily stop the one or more O-RU components from the E2 node; applying, by the nRT-RIC, AI / ML model inference based on the collected data to temporarily stop the one or more O-RU components; generating, by the nRT-RIC, at least one E2 policy command for energy saving (ES) that guides the E2 node in accordance with the capabilities of the O-RU to initiate the at least one ASM based on the AI / ML model inference; and transmitting, by the nRT-RIC, the at least one E2 policy command to the E2 node. Item

[12] The method of any one of items [8 to 11] further includes: receiving feedback by the NRT-RIC via the SMO framework based on the implementing, the feedback including a performance analysis of the AI / ML model; analyzing the performance of the AI / ML model in the nRT-NIC by the NRT-RIC; determining by the NRT-RIC that at least one predetermined performance goal has not been achieved based on the performance of the AI / ML model; and initiating a fallback mechanism associated with the at least one predetermined performance goal by the NRT-RIC. Item

[13] The method of any one of items [8 to 12], wherein implementing the at least one ASM includes scheduling at least one symbol by the E2 node or the O-RU based on the capability of the O-RU to minimize the number of symbols in the time domain, and the scheduling includes converging the at least one symbol from the time domain to the frequency domain to extend a shutdown period of the one or more O-RU components, wherein the converged at least one symbol is at least one of a physical downlink channel (PDSCH) symbol and a physical downlink control channel (PDCCH) symbol. Item

[14] The method of any one of items [8 to 12], wherein implementing the at least one ASM by the E2 node and the O-RU includes scheduling at least one slot by the E2 node or the O-RU to minimize the number of slots in the time domain based on the capabilities of the O-RU, and the scheduling includes converging the at least one slot from the time domain to the frequency domain to extend the outage period of the one or more O-RU components, wherein the converged at least one slot is at least one of a synchronization signal block - master information block (SSB-MIB) slot, a system information block type 1 (SIB1) slot, an SI slot, a paging frame slot, and a physical downlink channel (PDSCH) slot. Item

[15] A non-transitory computer-readable recording medium having stored thereon instructions executable by at least one processor, the at least one processor being configured to execute a method for implementing an advanced sleep mode in an open radio access network (O-RAN), the method including: collecting measurement data for training an artificial intelligence / machine learning (AI / ML) model by a service management and orchestration (SMO) framework; training an AI / ML model by a non-real-time radio access network (RAN) intelligent controller (NRT-RIC) based on the collected measurement data and deploying the AI / ML model to a near-real-time radio access network (RAN) intelligent controller (nRT-RIC); activating the trained AI / ML model in the nRT-RIC by the SMO framework; and transmitting the trained AI / ML model to an open radio access network (O-RAN) via an E2 node by the NRT-RIC. monitoring energy optimization data for AI / ML model inference from an O-RU (Operating-Responsive Unit); activating, by the SMO framework, at least one advanced sleep mode (ASM) in the nRT-RIC; collecting, by the nRT-RIC, data to temporarily shut down one or more O-RU components from the O-RU via the E2 node based on the activation of the at least one ASM; evaluating, by the nRT-RIC, the collected data to temporarily shut down one or more O-RU components based on the activated AI / ML model and the at least one ASM; requesting, by the nRT-RIC, via the E2 node to the O-RU to start the at least one ASM based on the evaluation; and implementing, by the O-RU, the at least one ASM based on the ASM start request. Item

[16] The non-transitory computer-readable recording medium described in Item

[15] , wherein collecting the data to temporarily stop the one or more O-RU components based on the activation of the at least one ASM includes: sending, by the nRT-RIC, a data collection request for energy saving to the E2 node via an E2 interface; receiving, by the E2 node, the data collection request for energy saving from the nRT-RIC; collecting, by the E2 node, the data to temporarily stop the one or more O-RU components from the O-RU via the open FH M-Plane interface; and sending, by the E2 node, the collected data to the nRT-RIC via the E2 interface to temporarily stop the one or more O-RU components. Item

[17] The non-transitory computer-readable storage medium described in Item [15 or 16], wherein evaluating the collected data to temporarily stop the one or more O-RU components includes: receiving, by the nRT-RIC, the collected data to temporarily stop the one or more O-RU components from the E2 node; applying AI / ML model inference, by the nRT-RIC, based on the collected data to temporarily stop the one or more O-RU components; generating, by the nRT-RIC, at least one E2 control command to request the E2 node that complies with the O-RU's capabilities to start the at least one ASM based on the AI / ML model inference; generating, by the nRT-RIC, at least one E2 policy command for energy saving (ES) that guides the E2 node that complies with the O-RU's capabilities to start the at least one ASM based on the AI / ML model inference; and transmitting, by the nRT-RIC, at least one of the E2 control command and / or the E2 policy command to the E2 node. Item

[18] The method further includes: receiving feedback by the NRT-RIC via the SMO framework based on the implementing, the feedback including a performance analysis of the AI / ML model; analyzing the performance of the AI / ML model in the nRT-NIC by the NRT-RIC; determining, by the NRT-RIC, that at least one predetermined performance goal has not been achieved based on the performance of the AI / ML model; and initiating, by the NRT-RIC, a fallback mechanism associated with the at least one predetermined performance goal. A non-transitory computer-readable recording medium described in Item [15 or 17]. Item

[19] The non-transitory computer-readable storage medium of any one of items [15 to 18], wherein implementing the at least one ASM includes: scheduling at least one symbol by the E2 node or the O-RU based on the capabilities of the O-RU to minimize the number of symbols in the time domain; and the scheduling includes converging the at least one symbol from the time domain to the frequency domain to extend a shutdown period of the one or more O-RU components, wherein the converged at least one symbol is at least one of a physical downlink channel (PDSCH) symbol and a physical downlink control channel (PDCCH) symbol. Item

[20] The non-transitory computer-readable storage medium of any one of items [15 to 18], wherein implementing the at least one ASM includes scheduling at least one slot by the E2 node or the O-RU to minimize the number of slots in the time domain based on the capabilities of the O-RU, and the scheduling includes converging the at least one slot from the time domain to the frequency domain to extend a stop period of the one or more O-RU components, wherein the converged at least one slot is at least one of a synchronization signal block - master information block (SSB-MIB) slot, a system information block type 1 (SIB1) slot, an SI slot, a paging frame slot, and a physical downlink channel (PDSCH) slot. < / o1> < / fh> < / fh> < / fh> < / o1> < / fh> < / fh> < / fh>

Claims

1. 1. A system for implementing an advanced sleep mode in an open radio access network (O-RAN), the system comprising: Near real-time Radio Access Network (RAN) Intelligent Controller (nRT-RIC); and a service management and orchestration (SMO) framework; The SMO framework comprises a non-real-time radio access network (RAN) intelligent controller (NRT-RIC); The system comprises: collecting measurement data for training an artificial intelligence / machine learning (AI / ML) model via the SMO framework; training an AI / ML model by the nRT-RIC based on the collected measurement data and deploying the AI / ML model to the nRT-RIC; Invoking the trained AI / ML model in the nRT-RIC via the SMO framework; monitoring, by the NRT-RIC, energy optimization data for AI / ML model inference from an open radio unit (O-RU) via an E2 node; Invoking at least one advanced sleep mode (ASM) in the nRT-RIC via the SMO framework; collecting data to temporarily deactivate one or more O-RU components from the O-RU via the E2 node by the nRT-RIC based on the activation of the at least one ASM; evaluating the collected data to temporarily shut down one or more O-RU components by the nRT-RIC based on the activated AI / ML model and the at least one ASM; requesting, by the nRT-RIC, via the E2 node, the O-RU to initiate the at least one ASM based on the evaluating; and implementing, by the O-RU, the at least one ASM based on the ASM initiation request; The system is configured to run

2. The system comprises: sending, by the nRT-RIC, a data collection request for energy saving to the E2 node via an E2 interface; receiving, by the E2 node, the data collection request for energy conservation from the nRT-RIC; collecting, by the E2 node, from the O-RU via the open FH M-Plane interface, the data for temporarily suspending the one or more O-RU components; and transmitting, by the E2 node, via the E2 interface to the nRT-RIC, the collected data for temporarily suspending the one or more O-RU components; and collecting the data to temporarily shut down the one or more O-RU components based on the activation of the at least one ASM. The system of claim 1 .

3. The system comprises: receiving, by the nRT-RIC, the collected data for temporarily suspending the one or more O-RU components from the E2 node; applying, by the nRT-RIC, AI / ML model inference based on the collected data to temporarily shut down the one or more O-RU components; generating, by the nRT-RIC, at least one E2 control command to request the E2 node to initiate the at least one ASM that conforms to the O-RU's capabilities based on the AI / ML model inference; and transmitting, by the nRT-RIC, the at least one of the E2 control commands to the E2 node; and evaluating the collected data to temporarily shut down the one or more O-RU components by The system of claim 1 .

4. The system comprises: receiving, by the nRT-RIC, the collected data for temporarily suspending the one or more O-RU components from the E2 node; applying, by the nRT-RIC, AI / ML model inference based on the collected data to temporarily shut down the one or more O-RU components; generating, by the nRT-RIC, at least one E2 policy command for energy saving (ES) that guides the E2 node conforming to the capability of the O-RU to initiate the at least one ASM based on the AI / ML model inference; and sending, by the nRT-RIC, the at least one E2 policy command to the E2 node; and evaluating the collected data to temporarily shut down the one or more O-RU components by The system of claim 1 .

5. The system further comprises: receiving feedback by the NRT-RIC via the SMO framework based on the implementing, the feedback including a performance analysis of the AI / ML model; analyzing, by the NRT-RIC, the performance of the AI / ML model on the nRT-NIC; determining, by the NRT-RIC, that at least one predetermined performance goal has not been achieved based on the performance of the AI / ML model; and initiating, by said NRT-RIC, a fallback mechanism associated with said at least one predetermined performance goal; configured to run The system of claim 1 .

6. The system comprises: configured to implement, by the E2 node or the O-RU based on the capabilities of the O-RU, the at least one ASM by scheduling at least one symbol to minimize a number of symbols in a time domain; The E2 node or the O-RU, Converging the at least one symbol from a time domain to a frequency domain to extend an outage period of the one or more O-RU components, wherein the converged at least one symbol is at least one of a physical downlink channel (PDSCH) symbol and a physical downlink control channel (PDCCH) symbol; and performing the scheduling by The system of claim 1 .

7. The system comprises: configured to implement the at least one ASM by scheduling, by the E2 node or the O-RU, at least one slot to minimize a number of slots in a time domain based on the capability of the O-RU; The E2 node or the O-RU, Converging the at least one slot from a time domain to a frequency domain to extend an outage period of the one or more O-RU components, wherein the converged at least one slot is at least one of a Synchronization Signal Block-Master Information Block (SSB-MIB) slot, a System Information Block Type 1 (SIB1) slot, an SI slot, a paging frame slot, and a Physical Downlink Channel (PDSCH) slot; and performing the scheduling by The system of claim 1 .

8. 1. A method for implementing an advanced sleep mode in an open radio access network (O-RAN), the method comprising: Collecting measurement data for training artificial intelligence / machine learning (AI / ML) models through a service management and orchestration (SMO) framework; training an AI / ML model by a non-real-time radio access network (RAN) intelligent controller (NRT-RIC) based on the collected measurement data, and deploying the AI / ML model to a near-real-time radio access network (RAN) intelligent controller (nRT-RIC); Invoking the trained AI / ML model in the nRT-RIC via the SMO framework; monitoring, by the NRT-RIC, energy optimization data for AI / ML model inference from an open radio unit (O-RU) via an E2 node; Invoking at least one advanced sleep mode (ASM) in the nRT-RIC via the SMO framework; collecting data to temporarily deactivate one or more O-RU components from the O-RU via the E2 node by the nRT-RIC based on the activation of the at least one ASM; evaluating, by the nRT-RIC, the collected data to temporarily shut down one or more O-RU components based on the activated AI / ML model and the at least one ASM; requesting, by the nRT-RIC, via the E2 node, the O-RU to initiate the at least one ASM based on the evaluating; and implementing, by the O-RU, the at least one ASM based on the ASM initiation request; A method comprising:

9. Collecting the data to temporarily suspend the one or more O-RU components based on the activation of the at least one ASM includes: sending, by the nRT-RIC, a data collection request for energy saving to the E2 node via an E2 interface; receiving, by the E2 node, the data collection request for energy conservation from the nRT-RIC; collecting, by the E2 node, from the O-RU via the open FH M-Plane interface, the data for temporarily suspending one or more O-RU components; and transmitting, by the E2 node, the collected data to the nRT-RIC via the E2 interface to temporarily disable the one or more O-RU components; The method of claim 8.

10. Evaluating the collected data to temporarily disable the one or more O-RU components includes: receiving, by the nRT-RIC, the collected data for temporarily suspending the one or more O-RU components from the E2 node; applying, by the nRT-RIC, AI / ML model inference based on the collected data to temporarily shut down the one or more O-RU components; generating, by the nRT-RIC, at least one E2 control command to request the E2 node to initiate the at least one ASM that conforms to the O-RU's capabilities based on the AI / ML model inference; and transmitting, by the nRT-RIC, the at least one of the E2 control commands to the E2 node; The method of claim 8.

11. Evaluating the collected data to temporarily disable the one or more O-RU components includes: receiving, by the nRT-RIC, the collected data for temporarily suspending the one or more O-RU components from the E2 node; applying, by the nRT-RIC, AI / ML model inference based on the collected data to temporarily shut down the one or more O-RU components; generating, by the nRT-RIC, at least one E2 policy command for energy saving (ES) that guides the E2 node conforming to the capability of the O-RU to initiate the at least one ASM based on the AI / ML model inference; and transmitting, by the nRT-RIC, the at least one E2 policy command to the E2 node; The method of claim 8.

12. The method further comprises: receiving feedback by the NRT-RIC via the SMO framework based on the implementing, the feedback including a performance analysis of the AI / ML model; analyzing, by the NRT-RIC, the performance of the AI / ML model on the nRT-NIC; determining, by the NRT-RIC, that at least one predetermined performance goal has not been achieved based on the performance of the AI / ML model; and initiating, by the NRT-RIC, a fallback mechanism associated with the at least one predetermined performance goal; The method of claim 8.

13. Implementing the at least one ASM includes: scheduling, by the E2 node or the O-RU, at least one symbol to minimize a number of symbols in a time domain based on the capability of the O-RU; The scheduling comprises: Converging the at least one symbol from a time domain to a frequency domain to extend an outage period of the one or more O-RU components, wherein the converged at least one symbol is at least one of a physical downlink channel (PDSCH) symbol and a physical downlink control channel (PDCCH) symbol. The method of claim 8.

14. Implementing the at least one ASM by the E2 node and the O-RU includes: scheduling at least one slot by the E2 node or the O-RU based on the capability of the O-RU to minimize the number of slots in a time domain; The scheduling comprises: Converging the at least one slot from a time domain to a frequency domain to extend an outage period of the one or more O-RU components, wherein the converged at least one slot is at least one of a Synchronization Signal Block-Master Information Block (SSB-MIB) slot, a System Information Block Type 1 (SIB1) slot, an SI slot, a paging frame slot, and a Physical Downlink Channel (PDSCH) slot. The method of claim 8.

15. 1. A non-transitory computer-readable storage medium having stored thereon instructions executable by at least one processor, the at least one processor being configured to perform a method for implementing an advanced sleep mode in an open radio access network (O-RAN); The method comprises: Collecting measurement data for training artificial intelligence / machine learning (AI / ML) models through a service management and orchestration (SMO) framework; training an AI / ML model by a non-real-time radio access network (RAN) intelligent controller (NRT-RIC) based on the collected measurement data, and deploying the AI / ML model to a near-real-time radio access network (RAN) intelligent controller (nRT-RIC); Invoking the trained AI / ML model in the nRT-RIC via the SMO framework; monitoring, by the NRT-RIC, energy optimization data for AI / ML model inference from an open radio unit (O-RU) via an E2 node; Invoking at least one advanced sleep mode (ASM) in the nRT-RIC via the SMO framework; collecting data to temporarily deactivate one or more O-RU components from the O-RU via the E2 node by the nRT-RIC based on the activation of the at least one ASM; evaluating the collected data to temporarily shut down one or more O-RU components by the nRT-RIC based on the activated AI / ML model and the at least one ASM; requesting, by the nRT-RIC, via the E2 node, the O-RU to initiate the at least one ASM based on the evaluating; and implementing, by the O-RU, the at least one ASM based on the ASM initiation request; A non-transitory computer-readable recording medium comprising:

16. Collecting the data to temporarily suspend the one or more O-RU components based on the activation of the at least one ASM includes: sending, by the nRT-RIC, a data collection request for energy saving to the E2 node via an E2 interface; receiving, by the E2 node, the data collection request for energy conservation from an nRT-RIC; collecting, by the E2 node, from the O-RU via the open FH M-Plane interface, the data for temporarily suspending the one or more O-RU components; transmitting, by the E2 node, the collected data to the nRT-RIC via the E2 interface to temporarily disable the one or more O-RU components; 16. The non-transitory computer-readable storage medium of claim 15.

17. Evaluating the collected data to temporarily disable the one or more O-RU components includes: receiving, by the nRT-RIC, the collected data for temporarily suspending the one or more O-RU components from the E2 node; applying, by the nRT-RIC, AI / ML model inference based on the collected data to temporarily shut down the one or more O-RU components; generating, by the nRT-RIC based on the AI / ML model inference, at least one of E2 control commands to request initiation of the at least one ASM to the E2 node that conforms to the capabilities of the O-RU; generating, by the nRT-RIC, at least one E2 policy command for energy saving (ES) that guides the E2 node conforming to the capability of the O-RU to initiate the at least one ASM based on the AI / ML model inference; and transmitting, by the nRT-RIC, at least one of the E2 control command and / or the E2 policy command to the E2 node; 16. The non-transitory computer-readable storage medium of claim 15.

18. The method further comprises: receiving feedback by the NRT-RIC via the SMO framework based on the implementing, the feedback including a performance analysis of the AI / ML model; analyzing, by the NRT-RIC, the performance of the AI / ML model on the nRT-NIC; determining, by the NRT-RIC, that at least one predetermined performance goal has not been achieved based on the performance of the AI / ML model; and initiating, by the NRT-RIC, a fallback mechanism associated with the at least one predetermined performance goal; 16. The non-transitory computer-readable storage medium of claim 15.

19. Implementing the at least one ASM includes: scheduling, by the E2 node or the O-RU, at least one symbol to minimize a number of symbols in a time domain based on the capability of the O-RU; The scheduling comprises: Converging the at least one symbol from a time domain to a frequency domain to extend an outage period of the one or more O-RU components, wherein the converged at least one symbol is at least one of a physical downlink channel (PDSCH) symbol and a physical downlink control channel (PDCCH) symbol.

16. The non-transitory computer-readable storage medium of claim 15.

20. Implementing the at least one ASM includes: scheduling at least one slot by the E2 node or the O-RU based on the capability of the O-RU to minimize the number of slots in a time domain; The scheduling comprises: Converging the at least one slot from a time domain to a frequency domain to extend an outage period of the one or more O-RU components, wherein the converged at least one slot is at least one of a Synchronization Signal Block-Master Information Block (SSB-MIB) slot, a System Information Block Type 1 (SIB1) slot, an SI slot, a paging frame slot, and a Physical Downlink Channel (PDSCH) slot.

16. The non-transitory computer-readable storage medium of claim 15.

Images