Method and apparatus for artificial intelligence based sub-band csi

Abstract

Embodiments herein provide a method and system for utilizing time correlation in AI based sub-band CSI systems. The method includes receiving a RRC configuration along with AI configuration parameters and a CSI-RS from a network apparatus. Further, the method includes estimating a precoder based on the CSI-RS. Further, the method includes decomposing the precoder into a W1, a W2, and a Wf. Further, the method includes encoding the W1 and the Wf to obtain an encoded W1 and Wf matrix using a non-AI encoding approach. Further, the method includes encoding the W2 using an AI model selected based on the AI configuration parameters and the CSI-RS to obtain an encoded W2 matrix. Further, the method includes transmitting the encoded W2 matrix, the encoded W1 matrix, and / or the encoded Wf matrix to the network apparatus based on a timer associated with the AI configuration parameters.

Description

METHOD AND APPARATUS FOR ARTIFICIAL INTELLIGENCE BASED SUB-BAND CSI

[0001] The disclosure is related to the field of artificial intelligence (AI). More particularly, the disclosure is related to a method and system for utilizing time correlation in AI based sub-band channel state information (CSI) systems.

[0002] 5G mobile communication technologies define broad frequency bands such that high transmission rates and new services are possible, and can be implemented not only in "Sub 6GHz" bands such as 3.5GHz, but also in "Above 6GHz" bands referred to as mmWave including 28GHz and 39GHz. In addition, it has been considered to implement 6G mobile communication technologies (referred to as Beyond 5G systems) in terahertz bands (for example, 95GHz to 3THz bands) in order to accomplish transmission rates fifty times faster than 5G mobile communication technologies and ultra-low latencies one-tenth of 5G mobile communication technologies.

[0003] At the beginning of the development of 5G mobile communication technologies, in order to support services and to satisfy performance requirements in connection with enhanced Mobile BroadBand (eMBB), Ultra Reliable Low Latency Communications (URLLC), and massive Machine-Type Communications (mMTC), there has been ongoing standardization regarding beamforming and massive MIMO for mitigating radio-wave path loss and increasing radio-wave transmission distances in mmWave, supporting numerologies (for example, operating multiple subcarrier spacings) for efficiently utilizing mmWave resources and dynamic operation of slot formats, initial access technologies for supporting multi-beam transmission and broadbands, definition and operation of BWP (BandWidth Part), new channel coding methods such as a LDPC (Low Density Parity Check) code for large amount of data transmission and a polar code for highly reliable transmission of control information, L2 pre-processing, and network slicing for providing a dedicated network specialized to a specific service.

[0004] Currently, there are ongoing discussions regarding improvement and performance enhancement of initial 5G mobile communication technologies in view of services to be supported by 5G mobile communication technologies, and there has been physical layer standardization regarding technologies such as V2X (Vehicle-to-everything) for aiding driving determination by autonomous vehicles based on information regarding positions and states of vehicles transmitted by the vehicles and for enhancing user convenience, NR-U (New Radio Unlicensed) aimed at system operations conforming to various regulation-related requirements in unlicensed bands, NR UE Power Saving, Non-Terrestrial Network (NTN) which is UE-satellite direct communication for providing coverage in an area in which communication with terrestrial networks is unavailable, and positioning.

[0005] Moreover, there has been ongoing standardization in air interface architecture / protocol regarding technologies such as Industrial Internet of Things (IIoT) for supporting new services through interworking and convergence with other industries, IAB (Integrated Access and Backhaul) for providing a node for network service area expansion by supporting a wireless backhaul link and an access link in an integrated manner, mobility enhancement including conditional handover and DAPS (Dual Active Protocol Stack) handover, and two-step random access for simplifying random access procedure (2-step RACH for NR). There also has been ongoing standardization in system architecture / service regarding a 5G baseline architecture (for example, service based architecture or service based interface) for combining Network Functions Virtualization (NFV) and Software-Defined Networking (SDN) technologies, and Mobile Edge Computing (MEC) for receiving services based on UE positions.

[0006] As 5G mobile communication systems are commercialized, connected devices that have been exponentially increasing will be connected to communication networks, and it is accordingly expected that enhanced functions and performances of 5G mobile communication systems and integrated operations of connected devices will be necessary. To this end, new research is scheduled in connection with eXtended Reality (XR) for efficiently supporting AR (Augmented Reality), VR (Virtual Reality), MR (Mixed Reality) and the like, 5G performance improvement and complexity reduction by utilizing Artificial Intelligence (AI) and Machine Learning (ML), AI service support, metaverse service support, and drone communication.

[0007] Furthermore, such development of 5G mobile communication systems will serve as a basis for developing not only new waveforms for providing coverage in terahertz bands of 6G mobile communication technologies, multi-antenna transmission technologies such as Full Dimensional MIMO (FD-MIMO), array antennas and large-scale antennas, metamaterial-based lenses and antennas for improving coverage of terahertz band signals, high-dimensional space multiplexing technology using OAM (Orbital Angular Momentum), and RIS (Reconfigurable Intelligent Surface), but also full-duplex technology for increasing frequency efficiency of 6G mobile communication technologies and improving system networks, AI-based communication technology for implementing system optimization by utilizing satellites and AI (Artificial Intelligence) from the design stage and internalizing end-to-end AI support functions, and next-generation distributed computing technology for implementing services at levels of complexity exceeding the limit of UE operation capability by utilizing ultra-high-performance communication and computing resources.

[0008] The principal object of the disclosure is to provide a method and system for utilizing time correlation in AI based sub-band channel state information (CSI) systems.

[0009] Another object of the disclosure is to provide a framework for improving the efficiency of a multiple-input multiple-output (MIMO) setup by improving the CSI compression accuracy by utilizing the time-correlation in sub-band CSI systems.

[0010] Yet another object of the disclosure is to distribute feedback bits across various layers of the UE to maximize the reconstruction accuracy.

[0011] Yet another object of the disclosure is to exploit the time-correlation in AI based sub-band CSI systems and to intelligently distribute the feedback bits across various layers to improve the CSI reconstruction accuracy, thus translating to a performance gain in MIMO systems.

[0012] In an aspect, the objectives are achieved by providing a method performed by a user equipment (UE) in a wireless communication system supporting artificial intelligence (AI) based sub-band channel state information (CSI). The method includes receiving a radio resource control (RRC) configuration along with AI configuration parameters and a CSI reference signal (CSI-RS) from a network apparatus. Further, the method includes estimating a precoder based on the CSI-RS received from the network apparatus. Further, the method includes decomposing the precoder into a wideband spatial component matrix (W1), a wideband coefficient matrix (W2), and a wideband frequency component matrix (Wf). Further, the method includes encoding the W1 and the Wf to obtain an encoded W1 matrix and an encoded Wf matrix. The encoded W1 matrix and the encoded Wf matrix are obtained using a non-AI encoding approach. Further, the method includes encoding the W2 using an AI model selected based on the AI configuration parameters and the CSI-RS received from the network apparatus to obtain an encoded W2 matrix. Further, the method includes transmitting at least one of the encoded W2 matrix, the encoded W1 matrix, or the encoded Wf matrix to the network apparatus based on a timer associated with the AI configuration parameters.

[0013] In another aspect, the objectives are achieved by providing a method performed by a network apparatus in a wireless communication system supporting artificial intelligence (AI) based sub-band channel state information (CSI). The method includes receiving an encoded wideband spatial component matrix (W1) matrix, an encoded wideband coefficient matrix (W2) matrix, and an encoded wideband frequency component matrix (Wf) matrix from a user equipment (UE). Further, the method includes decoding the encoded W1 matrix to obtain a decoded W1 matrix and the encoded Wf matrix to obtain a decoded Wf matrix. The encoded W1 matrix and the encoded Wf matrix are decoded using a non-AI decoding approach. Further, the method includes decoding the encoded W2 matrix using an AI model selected based on AI configuration parameters and a CSI reference signal (CSI-RS) to obtain a decoded W2 matrix. Further, the method includes performing a reconstruction of a precoder based on the decoded W1 matrix, the decoded W2 matrix, and the decoded Wf matrix obtained. Further, the method includes scheduling a data transmission between the UE and the network apparatus upon reconstruction of the precoder.

[0014] In another aspect, the objectives are achieved by providing a user equipment (UE) in a wireless communication system supporting artificial intelligence (AI) based sub-band channel state information (CSI). The UE includes a first processor, a first memory coupled to the first processor, and a first time correlation controller communicatively coupled to the first processor and the first memory. The first time correlation controller receives a radio resource control (RRC) configuration along with AI configuration parameters and a CSI reference signal (CSI-RS) from a network apparatus. Further, the first time correlation controller estimates a precoder based on the CSI-RS received from the network apparatus. Further, the first time correlation controller decomposes the precoder into a wideband spatial component matrix (W1), a wideband coefficient matrix (W2), and a wideband frequency component matrix (Wf). Further, the first time correlation controller encodes the W1 and the Wf to obtain an encoded W1 matrix and an encoded Wf matrix. The encoded W1 matrix and the encoded Wf matrix are obtained using a non-AI encoding approach. Further, the first time correlation controller encodes the W2 using an AI model selected based on the AI configuration parameters and the CSI-RS received from the network apparatus to obtain an encoded W2 matrix. Further, the first time correlation controller transmits at least one of the encoded W2 matrix, the encoded W1 matrix, or the encoded Wf matrix to the network apparatus based on a timer associated with the AI configuration parameters.

[0015] In another aspect, the objectives are achieved by providing a network apparatus in a wireless communication system supporting artificial intelligence (AI) based sub-band channel state information (CSI). The network apparatus includes a second processor, a second memory coupled to the second processor, and a second time correlation controller communicatively coupled to the second processor and the second memory. The second time correlation controller receives an encoded W1 matrix, an encoded W2 matrix, and an encoded Wf matrix from a user equipment (UE). Further, the second time correlation controller decodes the encoded W1 matrix to obtain a decoded W1 matrix and the encoded Wf matrix to obtain a decoded Wf matrix. The encoded W1 matrix and the encoded Wf matrix are decoded using a non-AI decoding approach. Further, the second time correlation controller decodes the encoded W2 matrix using an AI model selected based on AI configuration parameters and a CSI reference signal (CSI-RS) to obtain a decoded W2 matrix. Further, the second time correlation controller performs a reconstruction of a precoder based on the decoded W1 matrix, the decoded W2 matrix, and the decoded Wf matrix obtained. Further, the second time correlation controller schedules data transmissions between the UE and the network apparatus upon reconstruction of the precoder.

[0016] These and other aspects of the disclosure will be better appreciated and understood when considered in conjunction with the following description and the accompanying drawings. It should be understood, however, that the following descriptions, while indicating preferred embodiments and numerous specific details thereof, are given by way of illustration and not of limitation. Many changes and modifications be made within the scope of the embodiments herein.

[0017] These and other features, aspects, and advantages of the disclosure are illustrated in the accompanying drawings, throughout which like reference letters indicate corresponding parts in the various figures. The disclosure will be better understood from the following description with reference to the drawings, in which:

[0018] Fig. 1 is a sequence diagram that illustrates CSI-RS based inner-loop link adaptation according to the prior art.

[0019] Fig. 2 is a schematic diagram that illustrates spatial domain compression according to the prior art.

[0020] Fig. 3 is a block diagram that illustrates AI based W compression according to the prior art.

[0021] Fig. 4 is a sequence diagram that illustrates AI based W compression according to the prior art.

[0022] Fig. 5 is a schematic diagram that illustrates a schematic of the UE implemented to carry out the disclosed subject matter according to an embodiment as disclosed herein.

[0023] Fig. 6 is a schematic diagram that illustrates a schematic of the network apparatus implemented to carry out the disclosed subject matter according to an embodiment as disclosed herein.

[0024] Fig. 7 is a block diagram that illustrates AI based sub-band compression according to an embodiment as disclosed herein.

[0025] Figs. 8A and 8B are sequence diagrams that illustrates time correlation utilization in AI based sub-band CSI systems according to an embodiment as disclosed herein.

[0026] Fig. 9 is a sequence diagram that illustrates AI basedW2compression according to an embodiment as disclosed herein.

[0027] Fig. 10 is a sequence diagram that illustrates AI basedW2compression according to an embodiment as disclosed herein.

[0028] Fig. 11 is a sequence diagram that illustrates a flow for reportingW1,W2, andWfat different periodicities according to an embodiment as disclosed herein.

[0029] Fig. 12 is a sequence diagram that illustrates reportingW1,W2, andWfat different priorities according to an embodiment as disclosed herein.

[0030] Fig. 13 is a sequence diagram that illustratesW2compression when models as stored at the network apparatus according to an embodiment as disclosed herein.

[0031] Fig. 14 is a sequence diagram that illustratesW2compression when models as stored at the UE according to an embodiment as disclosed herein.

[0032] Figs. 15A and 15B are flow diagrams that illustrates a method for utilizing time correlation in AI based sub-band CSI systems by the UE according to an embodiment as disclosed herein.

[0033] Fig. 16 is a flow diagram that illustrates that illustrates a method for utilizing time correlation in AI based sub-band CSI systems by the network apparatus according to an embodiment as disclosed herein.

[0034] Hereinafter, embodiments of the disclosure will be described in detail with reference to the accompanying drawings.

[0035] In describing the embodiments, while numerous details are set forth for the purpose of illustration, it is understood that some aspects of the disclosure may be practiced with less than all of these details. Numerous variations and alternatives to the details provided herein are possible and are considered within the scope of the disclosure. In some instances, descriptions related to technical contents well-known in the art may be omitted so as to not obscure an understanding of the disclosure, and such omitted descriptions are understood to be within the scope of the disclosure.

[0036] For the same reason, in the accompanying drawings, some elements may be exaggerated, omitted, or schematically illustrated. Further, the size of each element does not completely reflect the actual size. In the drawings, identical or corresponding elements are provided with identical reference numerals or different reference numerals.

[0037] The advantages and features of the disclosure and ways to achieve them will be apparent by making reference to embodiments as described herein in detail in conjunction with the accompanying drawings. However, the disclosure is not limited to the embodiments set forth herein, but may be implemented in various different forms. Other features, aspects, and advantages of the subject matter described herein will become apparent from the disclosure. The following embodiments are merely examples to aid in an understanding of the disclosure and should not be construed to narrow the scope or spirit of the subject matter described herein in any way, but on the contrary, the disclosure covers all modifications, equivalents and alternatives falling within the spirit and scope of the subject matter as defined by the appended claims and equivalents thereof. Throughout the specification, the same or like reference numerals designate the same or like elements. Furthermore, terms which will be described herein are terms defined in consideration of the functions in the disclosure, and may be different according to users, intentions of the operators, or customs. Therefore, the definitions of the terms should be made based on the contents throughout the specification.

[0038] Herein, it will be understood that each block of flowchart illustrations, and combinations of blocks in the flowchart illustrations, may be performed based on computer program instructions. These computer program instructions may be loaded collectively onto at least one processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which perform through any one of, or in any combination of, the at least one processor of the computer or other programmable data processing apparatus, create means for performing the functions specified in the flowchart block(s). These computer program instructions may also be stored in a non-transitory computer usable or computer-readable memory that may direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer usable or computer-readable memory produce an article of manufacture including instruction means that perform the function specified in the flowchart block(s). The computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable data processing apparatus to produce a computer executed process such that the instructions that perform on the computer or other programmable data processing apparatus provide steps for executing the functions specified in the flowchart block(s).

[0039] Further, each block may represent a module, segment, or portion of code, which includes one or more executable instructions for executing the specified logical function(s). It should also be noted that in some alternative implementations, the functions noted in the blocks may occur out of the order. For example, two blocks(or functions) shown in succession may in fact be performed substantially concurrently or the blocks may sometimes be performed in the reverse order, depending upon the functionality involved.

[0040] As used in embodiments of the disclosure, a "~unit / module" may refer to a software element or a hardware element, such as a field programmable gate array (FPGA) or an application specific integrated circuit (ASIC), which performs a predetermined function. However, the term including the word "~unit / module" does not always have a meaning limited to software or hardware. The "~unit / module" may be constructed either to be stored in an addressable storage medium or to execute one or more processors. Therefore, the "~unit / module" includes, for example, software elements, object-oriented software elements, components such as class elements and task elements, processes, functions, properties, procedures, sub-routines, segments of a program code, drivers, firmware, micro-codes, circuits, data, database, data structures, tables, arrays, and parameters. The components and functions provided by the "~unit / module" may be either combined into a smaller number of components and a "~unit / module," or divided into additional components and a "~unit / module." Moreover, the components and "~units / modules" may be implemented to reproduce one or more central processing units (CPUs) within a device or a security multimedia card. Further, in the embodiments, the "~unit / module" may include one or more processors.

[0041] The entirety of the one or more computer programs may be stored in a single memory device or the one or more computer programs may be divided with different portions stored in different multiple memory devices.

[0042] Any of the functions or operations described herein can be processed by one processor or a combination of processors. The one processor or the combination of processors is circuitry performing processing and includes circuitry like an application processor (AP, e.g. a CPU), a communication processor (CP, e.g., a modem), a graphics processing unit (GPU), a neural processing unit (NPU) (e.g., an artificial intelligence (AI) chip), a Wi-Fi chip, a Bluetooth® chip, a global positioning system (GPS) chip, a near field communication (NFC) chip, connectivity chips, a sensor controller, a touch controller, a finger-print sensor controller, a display driver integrated circuit (IC), an audio CODEC chip, a universal serial bus (USB) controller, a camera controller, an image processing IC, microprocessors, microcontrollers, digital　signal processors, FPGA, ASIC, a microprocessor unit (MPU), a system on chip (SoC), an IC, or the like. The one processor or the combination of processors executes instructions that can be stored in a memory, such as the　operating system, in order to control the overall operation of the device.　Also, the one processor or the combination of processors is also capable of executing other processes and programs　resident in the memory, such as processes for the disclosure.

[0043] It will be appreciated that various embodiments of the disclosure according to the claims and description in the specification can be realized in the form of hardware, software or a combination of hardware and software.

[0044] Any such software may be stored in non-transitory computer readable storage media. The non-transitory computer readable storage media store one or more computer programs (software modules), the one or more computer programs include computer-executable instructions that, when executed by one or more processors of an electronic device individually or collectively, cause the electronic device to perform a method of the disclosure. Additionally, or alternatively, such software may be a computer program [product] comprising instructions which, when executed by one or more processors of an electronic device individually or collectively, cause the electronic device to perform a method of the disclosure.

[0045] Any such software may be stored in the form of volatile or non-volatile storage such as, for example, a storage device like read only memory (ROM), whether erasable or rewritable or not, or in the form of memory such as, for example, random access memory (RAM), memory chips, device or integrated circuits or on an optically or magnetically readable medium such as, for example, a compact disk (CD), digital versatile disc (DVD), magnetic disk or magnetic tape or the like. It will be appreciated that the storage devices and storage media are various embodiments of non-transitory machine-readable storage that are suitable for storing a computer program or computer programs comprising instructions that, when executed, implement various embodiments of the disclosure. Accordingly, various embodiments of the present disclosure may provide a program comprising code for implementing apparatus or a method as claimed in any one of the claims of this specification and a non-transitory machine-readable storage storing such a program.

[0046] Hereinafter, the determination of priority between A and B in the present disclosure may refer to various actions such as selecting the one having a higher priority based on a predefined priority rule and performing an operation corresponding thereto, or omitting or dropping an operation corresponding to the one having a lower priority.

[0047] Hereinafter, "A or B" as described in the present disclosure may be understood as "A and / or B," which may include A, or B, or both A and B.

[0048] In addition, "at least one of A, B, and C" as described in the present disclosure may be understood to include A, or B, or C, or any combination of A, B, and C.

[0049] In addition, "at least one of A, B, or C" as described in the present disclosure may be understood to include A, or B, or C, or any combination of A, B, and C.

[0050] Furthermore, "A / B" as described in the present disclosure may be understood as "A and / or B," which may include A, or B, or both A and B.

[0051] Furthermore, "A, B" as described in the present disclosure may be understood as "A and / or B," which may include A, or B, or both A and B.

[0052] Furthermore, "A and B" as described in the present disclosure may be understood as "A and / or B," which may include A, or B, or both A and B.

[0053] Furthermore, "if condition A and condition B are satisfied," as described in the present disclosure, may not be limited to a case where both condition A and condition B are satisfied, but may be understood to include a case where either condition A or condition B is individually satisfied, both condition A and condition B are satisfied, or one or more additional conditions are satisfied in combination.

[0054] Furthermore, throughout this disclosure, ordinal terms such as "first," "second," "third," etc., (and similar qualifiers) are used merely to distinguish between different instances, occurrences, configurations, messages, stages, elements or aspects of elements, operations, or information as described herein. Unless the context clearly dictates otherwise, the use of such ordinal terms does not itself require that the elements, operations, or information distinguished by these terms be structurally different, numerically distinct, or substantively dissimilar. For example, a "first signal" and a "second signal" may refer to instances of the same signal transmitted at different times or containing the same core information despite minor variations, or they may refer to signals with different content or characteristics, depending on the specific context. Similarly, a "first value" and a "second value" may represent the same magnitude but measured or applied in different circumstances, or they may represent different magnitudes. The interpretation should be guided by the specific technical context, function, and relationship described in the relevant portion of the specification and claims.

[0055] Furthermore, the terms "first ~", "second ~", etc., as described in the present disclosure with respect to various elements (e.g., information, objects, operation, sequences, or the like), should not limit those elements. These terms may only be intended to distinguish one element from another, and may not be intended to indicate a specific order. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element.

[0056] Furthermore, even if "first ~" and "second ~" are described in the present disclosure, it may be understood that element(s) referred to by "first ~" and "second ~" may be the same or different. For example, in case of element(s) being information, first information and second information may both be same information and, in some cases, are separate and different information.

[0057] In addition, the terms "if ~" and "in case that ~" as used in the disclosure or claims may be interpreted to include the meanings of "when (or upon) ~," "in response to ~," "based on ~," or "according to ~," and may be used interchangeably with these expressions. In addition, expressions other than those exemplified herein may also be used, as long as they have substantially the same meaning and do not impair the technical features of the present disclosure. If a method step (e.g. transmit a signal) is performed according to the disclosure of the application in connection with one of the above terms (such as "in case that ~" or the like), it may be interpreted to include the meanings (disclosure) of a prior determination that a feature has a specific state "~" (e.g. a bit length is above X), and then perform the method step in response to said determination.

[0058] For example, the physical layer signaling may be referred to as Layer 1 (L1) signaling and may include downlink control information (DCI). In addition, the higher layer signaling may include a medium access control (MAC) control message, a radio resource control (RRC) signaling message, a non-access stratum (NAS) signaling message, or an application layer message. The RRC signaling message may be referred to as L3 (layer 3) signaling. It should be noted, however, that the higher layer signaling is not limited to the aforementioned examples.

[0059] In addition, the term "not perform" as used in the present disclosure or claims may, in context, be understood to mean that the corresponding step is omitted or skipped. Such a term may be replaced with other terms having the same or substantially equivalent meaning.

[0060] In addition, "transmitting a message including A and B" as described in the present disclosure, may be understood as encompassing both (i) transmitting A and B in a single message, and (ii) transmitting A and B separately via multiple messages (e.g., transmitting a first message including A and a second message including B). This interpretation may also apply to messages that include two or more items (e.g., A, B, C), transmitted either together or separately.

[0061] In addition, "transmitting a message including A and transmitting a message including B" may also be interpreted as transmitting a message including A and B in a single message.

[0062] In the embodiments of the present disclosure described herein, terms or components included in the disclosure may be expressed in singular or plural form depending on the specific embodiments presented. However, such singular or plural expressions are selected appropriately for convenience of description, and the present disclosure is not limited to a singular or plural number of components. A component expressed in the plural form may be implemented as a single component, and a component expressed in the singular form may be implemented as multiple components.

[0063] The drawings or flowcharts described herein illustrate example methods that may be implemented according to the principles of the present disclosure, and various modifications may be made to the methods illustrated in the flowcharts of the present disclosure. For example, although illustrated as a series of steps, various steps in each drawing or flowchart may overlap, occur in parallel, occur in a different order, or be repeated. In other examples, any step may be omitted or replaced with another step.

[0064] The process of the flowchart may be performed by a device. One or more of the steps of the　flowchart can be implemented by one or more processors / computer programs executing　instructions to perform the noted functions.

[0065] The methods and apparatuses proposed in the embodiments of the present disclosure may be disclosed in connection with drawings disclosing flowcharts to illustrate example methods that may be implemented according to the principles of the present disclosure. Such flowcharts may contain different branches and / or sub-branches. It is understood that the principles of the present disclosure do not only contain the combination of all branches / sub-branches disclosed in the embodiment, but the present disclosure also contains at least one isolated branch / isolated sub-branch, in particular to a single branch / single sub-branch.

[0066] The methods and apparatuses proposed in the embodiments of the present disclosure are not limited to each embodiment individually, but may also be applied in combination of all or some of the embodiments proposed in the disclosure. Therefore, the embodiments of the present disclosure may be modified and applied without significantly departing from the scope of the present disclosure, as would be understood by those skilled in the art.

[0067] In this case, even if certain wordings are described differently across embodiments, they may be used interchangeably or in substitution or in combination if their underlying concepts are equivalent. For example, for the same or equivalent concept, even if one embodiment uses the expression "A" and another embodiment uses the expression "B", such expressions may be understood interchangeably, in substitution, or in combination.

[0068] The terms used in the following description to refer to access nodes, network entities, messages, interfaces between network entities, various types of identification information, and the like, are provided merely for the convenience of explanation by way of example. Therefore, the present disclosure is not limited to the terms described herein, and other terms having equivalent technical meanings may also be used. Such terms may also be interchangeable with terms defined in any 3rd generation partnership project　(3GPP) technical specifications (TS) or similar technical specifications, e.g., from the European telecommunications standards institute (ETSI), where appropriate.

[0069] Hereinafter, a base station (BS) is an entity that allocates resources to terminals, and may be at least one of a gNode B, an eNode B, a Node B, a wireless access unit, a BS controller, or a node on a network.

[0070] Furthermore, the base station of the present disclosure may include a split architecture comprising a central unit (CU) and a distributed unit (DU). In this structure, the CU is configured to process the higher layers of the control and user planes, while the DU is configured to process lower-layer radio resource functions. The embodiments of the present disclosure may be equally applicable to 5th generation (5G) base station architectures in which such CU and DU functional splits are implemented.

[0071] A terminal may include a user equipment (UE), a mobile station (MS), a cellular phone, a smartphone, a computer, a tablet, a wearable device, an Internet of Things (IoT) device, or any other device / system capable of performing communication functions.

[0072] In the disclosure, a downlink (DL) refers to a radio link through which a BS transmits a signal to a terminal, and an uplink (UL) refers to a radio link through which a terminal transmits a signal to a BS.

[0073] Furthermore, hereinafter, 5G mobile communication technologies (e.g., 5G new radio (NR)), 6th generation (6G) mobile communication technologies may be described by way of example, but the embodiments of the present disclosure may also be applied to other communication systems having similar technical backgrounds or channel types. For example, newly evolved mobile communication systems developed after 5G and 6G may be included. Furthermore, based on determinations by those skilled in the art, the embodiments of the present disclosure may also be applied to other communication systems (e.g., Wi-Fi systems) through some modifications without significantly departing from the scope of the present disclosure

[0074] In the following description, the terms physical channel and signal may be used interchangeably with data or control signal. For example, the term physical downlink shared channel (PDSCH) refers to a physical channel through which data is transmitted, but the term PDSCH may also be used to refer to the data itself. That is, in the present disclosure, the expression "transmit a physical channel" may be interpreted as being equivalent to the expression "transmit data or a signal via a physical channel."

[0075] Hereinafter, in the context of the present disclosure, higher layer signaling may refer to signaling corresponding to at least one or any combination of the following: master information block (MIB), system information block (SIB) or SIB M (M = 1, 2, ...), RRC, or MAC control element (CE), or a non-access stratum (NAS) signaling message, or an application layer message. The RRC signaling message may be referred to as Layer 3 (L3) signaling.

[0076] In addition, L1 signaling may refer to signaling corresponding to at least one or any combination of signaling techniques using the at least one or any combination of the following physical layer channels or signaling: physical downlink control channel (PDCCH), DCI, UE-specific DCI, group-common DCI, common DCI, scheduling DCI (e.g., DCI used for scheduling downlink or uplink data), non-scheduling DCI (e.g., DCI not used for scheduling downlink or uplink data) physical uplink control channel (PUCCH), or uplink control information (UCI). The L1 signaling message may be referred to as a physical layer signaling.

[0077] Hereinafter, the expression that information is configured by the BS, as used in the present disclosure or claims, may, in context, be understood to mean that the terminal receives the corresponding information from the BS via a physical layer signaling or a higher layer signaling. Such an expression may be replaced with other terms having the same or substantially equivalent meaning.

[0078] Hereinafter, the operational principle of the present disclosure will be described in detail with reference to the accompanying drawings.

[0079] Channel fading refers to the time-varying attenuation of a signal as it propagates through a wireless communication channel. This can occur due to multiplicity of factors like atmospheric conditions, large obstacles (like trees, buildings), UE motion and multipath propagation. Channel fading can significantly degrade the performance of wireless communication systems by: leading to a lower Signal-to-Interference-plus-Noise Ratio (SINR) at the receiver, increasing a Bit Error Rate(BER), reducing data throughput thus leading to dropped connections. In the 5G New Radio (NR), CSI link adaptation utilizes the reported CSI from the user equipment (UE) to dynamically adjust various transmission parameters at the gNodeB (gNB).

[0080] Generalization refers to the ability of an AI model to perform well on data from different sources or environments, even if it was trained on a specific dataset.In AI-based CSI compression, the encoder (deployed at the UE) and decoder (deployed at the BS) are trained separately by different vendors (for example, UE vendor and gNB vendor). When an encoder trained on one vendor's dataset is used with a decoder trained on another vendor's dataset, the performance significantly degrades. This limitation makes it challenging to ensure consistent performance across different vendors, which is critical for interoperability in real-world networks.

[0081] Hence, is desirable to address the above mentioned problems and disadvantages or at least provide a useful alternative.

[0082] Fig. 1 is a sequence diagram that illustrates CSI-RS based inner-loop link adaptation according to the prior art. As shown, the sequence diagram includes a UE (100) and a network apparatus (200) in communication with each other. The network apparatus (200) configures CSI-RS to the UE (100) that is transmitted either periodically or aperiodically (trigger-based). Using the CSI-RS, the UE (100) estimates the channelHfor the downlink, which is of the dimensionN×R×K. Here,Ndenotes the number of Tx ports,Rdenotes the number of receiver ports at the UE (100), andKdenotes the number of sub-bands (SBs) in the frequency domain. For each sub-band(s), the UE (100) computes a precoder for each MIMO layer using Singular Value Decomposition (SVD). The precoder for thelthMIMO layer is computed as of the dimensionN×1. The set of precoders of all sub-bands are denoted as which is of the dimensionN×K. Hence, . The overall CSI is denoted asWencompassing all the L layers each with CSI matrix . This is compressed and reported to the network apparatus (200).

[0083] The inner loop procedure in new radio (NR) refers to the fast adaptation of the modulation and coding scheme (MCS) for downlink transmissions. This is based on the estimated CSI which is based on the CSI-RS received from the network apparatus (200). As part of this, the UE (100) estimates multiple metrics and reports to the network apparatus (200) for effective link adaptation in a MIMO setup. In essence, the inner loop procedure plays a crucial role in maximizing spectral efficiency and maintaining reliable communication in the dynamic wireless environment.

[0084] In existing 3GPP standards, the precoderWis then compressed using a predefined approach and reported to the network apparatus (200). In addition to the precoder, the UE (100) also computes and reports additional metrics such as channel quality indicator (CQI) to indicate the effective modulation and code rate preference rank indicator (RI) that indicates the preferred number of MIMO data streams for scheduling in the downlink. Note that for , range of layer index is . Based on the CSI metrics received from the UE (100), the network apparatus (200) selects the PDSCH configuration such as modulation and coding scheme (MCS), rank indicator (RI), for sharing the data in the downlink, and shares this information to the UE (100) through downlink control information (DCI). The CSI metrics to be reported by the UE (100) contain a large amount of data and hence needs to be compressed before transmitting it to the network apparatus (200). Else, it would lead to wastage of precious bandwidth.

[0085] Fig. 2 is a schematic diagram that illustrates spatial domain compression according to the prior art. In 3GPP Release 15 Type II CSI reporting, the precoder matrix is compressed in spatial domain without any compression in frequency domain using a dual-stage Type II CSI codebook. Each column of is represented as an LC ofLbeams represented by the matrixW1of sizeN×2L, an example of which is shown in Fig. 2. The set ofLbeams are chosen in common across the MIMO layers, sub-bands, and antenna polarization. The combining coefficients for each case is given by the matrix which is of dimension 2L×K. Finally, the beam and coefficient information is quantized and reported to the network apparatus (200).

[0086] 3GPP Release 15 standard compresses the precoder in the spatial dimension, which exploits the correlation across the coefficients. However, the correlation also exists in the frequency dimension, which is not exploited. For example, the number of columns matrix is equal to the number of subbandss. 3GPP Release 16 Type-II CSI exploits this aspect by representing , where the effective dimension of the coefficient matrixW2is reduced to 2L×Mand, anWfis an additional matrix of sizeM×K, respectively. Basically, the matrixWfdenotes the set ofMfrequency domain beams, similar toW1in the spatial domain.

[0087] Fig. 3 is a block diagram that illustrates AI basedWcompression according to the prior art. Fig. 3 depicts the auto-encoder (AE) based AI-CSI compression system model that is prevalently used. It consists of an AE-encoder at the UE (100) to compress the CSI into latent space. A reconstructed version of the precoder is generated by an AE-decoder at the network apparatus (200).

[0088] Fig. 4 is a sequence diagram that illustrates AI basedWcompression according to the prior art. As shown in the sequence diagram, the UE (100) and the network apparatus (200) are in communication with each other. For runtime operation, the AE-encoder can be downloaded from the network apparatus (200) or could be pre-stored at the UE (100). The network apparatus (200) provides the configuration details containing the AI based encoding to be performed at the UE (100) through RRC signalling. After precoder estimation, the UE (100) performs AI-based encoding of the precoder based on the RRC configuration received from the network apparatus (200). The encoded CSI consisting of precoder is then transmitted to the network apparatus (200), which decodes it using AI based method. The decoded precoder information is then used for data scheduling.

[0089] In current AI-based CSI compression methods, the precoder data (W) is reported at every instance, making it difficult to exploit time correlation effectively. Without leveraging time correlation, the system cannot minimize feedback overhead, leading to inefficiencies in data transmission. In the current AI-based CSI compression method, the precoder dataWof dimensionN×Kis compressed, whereKis the number of sub-bands andNis the number of transmit ports.

[0090] In the prior art, the network apparatus (200) sends CSI-RS to the UE (100) based on which the UE (100) estimates the precoder, compresses it and shares it with the network apparatus (200) (e.g., BS) at every reporting instance. In the proposed solution, after precoder computation, the precoder is decomposed to obtain theW1(wide-band),W2(sub-band) andWf(frequency-band) matrices. TheW1andWfmatrices which are less frequently changing are reported less often whereas the more dynamic componentW2is reported at every instance. This enables reduction in the feedback bits overhead at every reporting instance and also improves the reconstruction accuracy of the precoder at the network apparatus (200) as compared to the conventional systems. The time correlation in sub-band CSI systems is exploited by reporting the wideband and frequency band components of precoder at different periodicities. This enables reducing the feedback bits overhead and also improve the CSI reconstruction accuracy using an AI based system.

[0091] In the prior art, the number of feedback bits to be used for transmission of the precoder information from the UE (100) to the network apparatus (200) has a static distribution across the various layers of MIMO. This leads to inefficiency and wastage of feedback bits. In the proposed solution, the feedback bits are duly re-designed across various layers of MIMO setup, which helps in improving the CSI reconstruction accuracy. The proposed solution provides the necessary signal flows and message exchanges between the network apparatus (200) and the UE (100) to enable the AI based sub-band compression.

[0092] Fig. 5 is a block diagram that illustrates a schematic of a UE (100) implemented to carry out the disclosed subject matter according to an embodiment as disclosed herein. Examples of the UE (100) can include, but are not limited to, Consumer Electronics (such as Mobile Phones and Smartphones), Tablets, Wearable Devices, Computing Devices (such as Laptops, Notebooks, Desktops, Workstations, etc.), IoT Devices, Automotive Systems (such as connected cars, Autonomous Vehicles, Vehicle-to-Everything (V2X) communication devices, etc.), Enterprise Devices such as robotics, Specialized Equipment (such as Medical Devices, Public Safety Devices, etc.), Media Devices (such as Gaming Consoles, Streaming Devices, etc.).

[0093] In an embodiment, in Fig. 5, the UE (100) includes a first processor (102), a first memory (104), a first I / O interface (106), and first time correlation controller (108) coupled to the first processor (102) and the first memory (104). The components are explained in further detail below.

[0094] The first processor (102) communicates with the first memory (104), the first I / O interface (106), and the first time correlation controller (108). The first processor (102) is configured to execute instructions stored in the first memory (104) and to perform various processes. The first processor (102) includes one or a plurality of processors, is a general-purpose processor such as the CPU, an application processor (AP), or the like, a graphics-only processing unit such as a graphics processing unit (GPU), a visual processing unit (VPU), and / or an Artificial Intelligence (AI) dedicated processor such as a neural processing unit (NPU).

[0095] The first memory (104) includes storage locations to be addressable through the first processor (102). The first memory (104) stores AI configuration parameters, a CSI-RS, the precoder, the encodedW2matrix, the encodedW1matrix, the encodedWfmatrix, and the like. The first memory (104) is not limited to a volatile memory and / or a non-volatile memory. Further, the first memory (104) includes a plurality of computer-readable storage media. The first memory (104) includes non-volatile storage elements. For example, non-volatile storage elements include magnetic hard disks, optical disks, floppy disks, flash memories, or forms of electrically programmable memories (EPROM) or electrically erasable and programmable (EEPROM) memories.

[0096] The first I / O interface (106) transmits the information between the first memory (104) and external peripheral devices. The peripheral devices are the input-output devices associated with the UE (100). Further, the first time correlation controller (108) communicates with the first I / O interface (106) and the first memory (104). The first time correlation controller (108) is coupled to the first memory (104) and the first processor (102). This coupling allows for efficient data transfer and communication between the components, ensuring that the first time correlation controller (108) can enable dynamic frequency adaptation in an electronic device having multi-core processors for managing multiple TC activities.

[0097] The first time correlation controller (108) is an innovative integrated circuit that is implemented in the UE (100). In an embodiment, the structure of such innovative integrated circuit includes a multi-core architecture that enables utilizing time correlation in AI based sub-band CSI systems. Each core is optimized for specific tasks, such as estimating the precoder, decomposing the precoder to obtain the matrices (W1,W2,Wf), encoding the matrices, and the like. The innovative integrated circuit for the above-mentioned points is made of a combination of analog and digital components designed to enable utilizing time correlation in AI based sub-band CSI systems. The analog components include a low-noise amplifier and a high-precision analog-to-digital converter to ensure accurate signal processing. The digital components consist of a microcontroller unit (MCU) and a digital signal processor (DSP) that work in tandem to enable utilizing time correlation in AI based sub-band CSI systems.

[0098] Fig. 6 is a schematic diagram that illustrates a schematic of the network apparatus (200) implemented to carry out the disclosed subject matter according to an embodiment as disclosed herein. The network apparatus (200) includes various hardware and software components that facilitate communication between user equipment and network infrastructure. Examples of the network apparatus (200) can include, but is not limited to Base Stations (such as macro cells, small cells, femtocells, picocells) for wireless communication, Antennas and RF Units (e.g., MIMO, beamforming) to enhance signal coverage and data throughput, Core Network Equipment (e.g., MMEs, S-GWs, P-GWs in 4G; AMFs, UPFs in 5G) for data routing, mobility, and session control, Network Function Virtualization (NFV) and Software-Defined Networking (SDN) for dynamic resource allocation and scalability, Edge Computing Nodes (e.g., MEC servers) for low-latency processing,　Backhaul and Transport Equipment (e.g., fiber-optic links, microwave relays, Ethernet switches) to connect base stations to the core network, Network Management Systems (NMS) and Operation Support Systems (OSS) for network configuration, fault management, and optimization, Radio Network Controllers (RNCs) in 3G, Distributed Units (DUs), and Centralized Units (CUs) in 5G, Network Slicing Components for virtualized resource allocation, Security elements (e.g., Firewalls, IDS, AAA Servers) for secure communication.

[0099] In an embodiment, in Fig. 6, the network apparatus (200) includes a second processor (202), a second memory (204), a second I / O interface (206), and a second time correlation controller (208) coupled to the second processor (202) and the second memory (204). The components are explained in further detail below.

[0100] The second processor (202) communicates with the second memory (204), the second I / O interface (206), and the second time correlation controller (208). The second processor (202) is configured to execute instructions stored in the second memory (204) and to perform various processes. The second processor (202) includes one or a plurality of processors, is a general-purpose processor such as the CPU, an application processor (AP), or the like, a graphics-only processing unit such as a graphics processing unit (GPU), a visual processing unit (VPU), and / or an Artificial Intelligence (AI) dedicated processor such as a neural processing unit (NPU).

[0101] The second memory (204) includes storage locations to be addressable through the second processor (202). The second memory (204) stores the encoded matrices (encodedW1matrix, encodedW2matrix, encodedWfmatrix), a reconstruction of the precoder, and the like. The second memory (204) is not limited to a volatile memory and / or a non-volatile memory. Further, the second memory (204) includes a plurality of computer-readable storage media. The second memory (204) includes non-volatile storage elements. For example, non-volatile storage elements include magnetic hard disks, optical disks, floppy disks, flash memories, or forms of electrically programmable memories (EPROM) or electrically erasable and programmable (EEPROM) memories.

[0102] The second I / O interface (206) transmits the information between the second memory (204) and external peripheral devices. The peripheral devices are the input-output devices associated with the network apparatus (200). Further, the second time correlation controller (208) communicates with the second I / O interface (206) and the second memory (204). The second time correlation controller (208) is coupled to the second memory (204) and the second processor (202). This coupling allows for efficient data transfer and communication between the components, ensuring that the second time correlation controller (208) can enable utilizing time correlation in AI based sub-band CSI systems.

[0103] The second time correlation controller (208) is an innovative integrated circuit that is implemented in the network apparatus (200). In an embodiment, the structure of such innovative integrated circuit includes a multi-core architecture that enables utilizing time correlation in AI based sub-band CSI systems. Each core is optimized for specific tasks, such as decoding the encoded matrices, performing a reconstruction of the precoder based on the decoded matrices, scheduling data transmissions, and the like. The innovative integrated circuit for the above-mentioned points is made of a combination of analog and digital components designed to enable utilizing time correlation in AI based sub-band CSI systems. The analog components include a low-noise amplifier and a high-precision analog-to-digital converter to ensure accurate signal processing. The digital components consist of a microcontroller unit (MCU) and a digital signal processor (DSP) that work in tandem to enable utilizing time correlation in AI based sub-band CSI systems.

[0104] Fig. 7 is a block diagram that illustrates AI based sub-band compression according to an embodiment as disclosed herein. Fig. 7 shows the proposed solution for the AI-based sub-band (SB) CSI compression performed at the UE (100) using an encoder (100A) and the network apparatus (200) using a decoder (200A).

[0105] At the UE (100):

[0106] In step (A), the network apparatus (200) initially provides the CSI configuration, specifically w.r.t generationW1andWf. This is typically achieved through RRC signalling from the network apparatus (200) to the UE (100). Using this configuration, in step (B), the UE (100) decomposes the estimated precoder into its componentsW1,Wfas the received CSI config in step (A), and the corresponding coefficients matrixW2, of size 2L×M. In general,W1represents the wideband spatial component andWfrepresents the wideband frequency component The generation ofW1andW2could be done by re-using the existing 3GPP schemes or a new scheme could be employed optimized to AI-based SB CSI compression. In step-(C), the spatial matrixW1and frequency domain beam matrixWfis encoded using conventional approach to generate aL1fbit payload. This can be through re-use of existing 3GPP schemes, or through alternatives to enhance the performance of AI-based SB CSI. In step (D), the coefficient matrixW2is compressed using AI-based auto-encoder approach to generate aLaebit payload. The total payload ofL=Lae+L1fis compressed and transmitted through the CSI reporting air interface.

[0107] At the network apparatus (200):

[0108] The network apparatus (200) receives the encodedL=Lae+L1fbits through the CSI reporting procedure. Using theL1fbits, the matricesW1andWfare reconstructed through conventional means as shown in step (E). Further, theLaebits are passed through the AE-decoder to generate the , the quantized version of theW2matrix. The AE is trained offline, to minimize the reconstruction loss between the actual and reconstructed versions of the coefficient matrix,W2and , respectively. This is shown in step (F). Finally, as shown in step (G), the quantized version of the precoder is reconstructed as . The proposed method uses a hybrid approach by employing both the traditional and AI-based approaches to compress parts of the precoders.

[0109] Figs. 8A and 8B are sequence diagrams that illustrate time correlation utilization in AI based sub-band CSI systems according to an embodiment as disclosed herein. As shown in the sequence diagram, the UE (100) and the network apparatus (200) are in communication with each other. The network apparatus (200) shares the RRC configuration along with AI configuration parameters to the UE (100). The AI configuration parameters include at least one of a look-up table (LUT) provided by the network apparatus (200), an AI model indicator (AMI) obtained based on the LUT, a periodicity associated with theW1and theWf(PW1Wf), a periodicity associated with theW2(PW2), or parameter combinations (param_comb). The parameter combinations specify a number of feedback bits allocated across different layers of the UE (100).

[0110] The UE (100) estimates the precoder from the CSI-RS transmitted by the network apparatus (200). The UE (100) calculates theW1,W2andWffrom the precoder. The UE (100) uses the LUT and the param_comb information shared by the network apparatus (200) to distribute feedback bits across layers. TheW1andWfmatrices are encoded using the conventional method. TheW2is encoded using an AI model, selected by the UE (100) on the basis of the AMI.

[0111] The UE (100) then checks for a PW1Wf timer. If the timer has not expired, the UE (100) transmits only the encodedW2matrix to the network apparatus (200). When the timer expires, the network apparatus (200) reports all the encoded matrices (W1,W2andWf) to the network apparatus (200) and resets the timer. Upon receiving the CSI report, the network apparatus (200) selects the appropriate AI model using AMI for decoding. The network apparatus (200) also decodes the most recently receivedW1andWfusing conventional methods and decodesW2using the AI model. The network apparatus (200) then combines the decoded matrices to reconstruct the precoder. On the basis of the reconstructed precoder, the network apparatus (200) schedules the data transmission.

[0112] Fig. 9 is a sequence diagram that illustrates AI basedW2compression according to an embodiment as disclosed herein. As shown in the sequence diagram, the UE (100) and the network apparatus (200) are in communication with each other. Initially, the network apparatus (200) provides the configuration details containing the AI based encoding to be performed at the UE (100), which could be done via RRC signalling. The network apparatus (200) transmits the CSI-RS to the UE (100), using which the UE (100) estimates the CSI including precoder.

[0113] After precoder estimation, the UE (100) performs decomposition to obtain the componentsW1,W2, andWf. The matricesW1andWfare encoded using conventional approach. The coefficient matrixW2is compressed using the AI-based encoder selected based on the AI-CSI configuration received from the network apparatus (200). The encoded versions of the matrices are then reported to the network apparatus (200).

[0114] At the network apparatus (200), the appropriate AI-based decoder is selected, using which theW2is reconstructed. Further,W1andWfare reconstructed through conventional means. Then, the SB CSI precoders are reconstructed using the expressionW=W1W2Wf. This is further used by the network apparatus (200) while scheduling data transmissions to the UE (100).

[0115] In order to achieve different levels of reconstruction accuracy, 3GPP specifies various parameter combinations (param_comb) that specify the number of bits that can be used for feedback. The network apparatus (200) specifies the param_comb to the UE (100) via RRC signalling. As the param_comb is increased, the reconstruction accuracy improves as given in Table 1. For a given param_comb let the total number of bits assigned bePo.

[0116] [Table 1] Reconstruction accuracy vs Feedback bits

[0117]

[0118] Let , and be the number of bits assigned toW1,W2, andWf, respectively, for layerl. Thus, . Consequently, . Thus, it is observed that for reportingW2across all the layers, a fixed number of bits are present given by equation above. The typical distribution of the bits assigned across layers are nearly equal. However, this results in limited performance improvement over the conventional enhanced Type-II (eType-II) codebooks. Hence, in order to maximize the performance improvement, the feedback bits are intelligently assigned to each layer. By intelligently assigning feedback bits to each layer, a significant improvement is observed in the reconstruction accuracy as shown in Table 2.

[0119] [Table 2] Performance conventional vs Proposed

[0120]

[0121] Fig. 10 is a sequence diagram that illustrates AI basedW2compression according to an embodiment as disclosed herein. As shown in the sequence diagram, the UE (100) and the network apparatus (200) are in communication with each other. The network apparatus (200) having a higher computational capacity and memory will be suitable for training and storing the AI models. There are two different ways in which the network apparatus (200) can perform AI training.

[0122] Offline training

[0123] The network apparatus (200) performs an offline training using a predefined dataset. In order to maximize the reconstruction accuracy across various layers, the network apparatus (200) performs hyper-parameter tuning to determine the best possible combination of feedback bits across various layers. The network apparatus (202) can generate the LUT based on the training. The LUT can be used by the network apparatus (200) to generate the AI model indicator (AMI), which is then shared with the UE (100) via RRC signaling. The model selected through AMI implicitly contains the encoded payload size for the corresponding layer.

[0124] Online training

[0125] In order to cater to various deployment scenarios, the network apparatus (200) can perform online training using the localized dataset. The network apparatus (200) can generate the LUT and share the updated AMI with the UE (100) via RRC signaling.

[0126] Fig. 11 is a sequence diagram that illustrates a flow for reportingW1,W2, andWfat different periodicities according to an embodiment as disclosed herein. As shown, the sequence diagram includes the UE (100) and the network apparatus (200) in communication with each other.

[0127] In the proposed pre-processing, the spatial and frequency componentsW1andWfvary relatively slowly as compared to their coefficientsW2. This enables the proposed AI-based SB CSI method to use different periodicity of reporting forW2compared toW1andWf. The network apparatus (200) initially configures different periodicity of reporting for the CSI componentsW1,W2, andWf. After precoder estimation and matrix computation, the matrixW2is reported in each measurement occasion. On the other hand, the matricesW1andWfare reported only once everyNslots. This enables the proposed AI-based compression method to exploit the advantages of time-correlation. For example, the tables below depict the conventional 3GPP Release 16 Type-II codebook performance when separate periodicities are used forW1andWf.

[0128] [Table 3] Release 16 Type-II codebook performance for various periodicities

[0129]

[0130] [Table 4] Performance of AI-basedW2compression with different periodicities forW2thanW1andWf

[0131]

[0132] Table 3 shows the performance of compressingW2matrix using AI-based CSI approach for two cases:W1andWfreported with the same periodicity as that ofW2shown in column (A); andW1andWfreported with one third the periodicity as that ofW2shown in columns (B). Both the cases show gains compared to the 3GPP Release 16 Type-II compression performance. Further, columns (A) and (B) show comparable performance with respect to the gain of the proposed method over the conventional 3GPP Release 16 Type-II compression.

[0133] In an example, assume a scenario whereW2is being reported with a periodicity of 5 milliseconds (ms) andW1andWfare being reported at a periodicity of 20 ms. The network apparatus (200) instructs the UE (100) to use param_comb = 1 via RRC signalling. Conventionally, this means that at every reporting instance, the UE (100) can use 135 bits for feedback out of which 96 bits would be assigned forW2reporting and 39 bits forW1andWfreporting. In the proposed solution, while using different periodicities forW1andWf, we may encounter that in 3 / 4thof the slots onlyW2is being reported and only in 1 / 4thof the slotsW1,W2, andWfare being reported. Thus, in order to handle the AI basedW2compression for such a scenario, two methods are proposed.

[0134] Method 1

[0135] In the slots where onlyW2is being reported, the entire feedback bits available (135 bits) are utilized for AI basedW2compression. In the rest of the slots, where all three are reported, the conventional distribution of feedback bits can be utilized. This helps in achieving a better reconstruction accuracy. However, this increases the number of AI models that need to be stored / downloaded at the UE (100), since different AI models would be needed when using 135 bits or 96 bits for compression.

[0136] Method 2

[0137] In this method, irrespective of the slots, the conventionally determined feedback bits (96 bits) are utilized for AI basedW2compression. The additional bits available for AI basedW2compression are not utilized. Thus, this method is much simpler in terms of implementation as dealing with multiple AI models is not required.

[0138] Fig. 12 is a sequence diagram that illustrates reportingW1,W2, andWfat different priorities according to an embodiment as disclosed herein. As shown in the sequence diagram, the UE (100) and the network apparatus (200) are in communication with each other. The sequence diagram illustrates implementations of method 1 and method 2 are explained above:

[0139] Implementation of Method 1:

[0140] The network apparatus (200) shares the AI configuration and Periodicity_W1_Wf parameter with the UE (100) via the RRC messages. The AI configuration contains AI model indication information, which helps to determine which model to use at what stage. The AMI 1 corresponds to the AI model which uses all the feedback bits for that slot forW2compression. The AMI 2 corresponds to the AI model which uses only a portion of the feedback bits for that slot forW2compression. Based on the Periodicity_W1_Wf parameter, the UE (100) creates a timer, which expires whenever the timer reaches this value.

[0141] The UE (100) then checks if the timer has expired or not. If the timer has not expired, the UE (100) selects the AI model based on AMI 1. The UE (100) also estimatesW1,W2andWffrom the precoder and performs AI basedW2compression. The UE (100) shares the encodedW2with the network apparatus (200), which performs the decoding of the encodedW2. If the timer expires, then the UE (100) selects the AI model based on AMI 2 and encodes theW2using AMI 2. The UE (100) sharesW1andWfalong with the encodedW2to the network apparatus (200) which performs the decoding. The UE (100) then resets the timer and the cycle continues.

[0142] Implementation of Method 2:

[0143] The network apparatus (200) shares the AI configuration and Periodicity_W1_Wf parameter with the UE (100) via the RRC messages. The AI configuration contains AI model indication information, which helps to determine which model to use at what stage. The AMI 2 corresponds to the AI model which uses only a portion of the feedback bits for that slot forW2compression. Based on the Periodicity_W1_Wf parameter the UE (100) creates a timer, which expires whenever the timer reaches this value. The UE (100) then checks if the timer has expired or not.

[0144] If the timer has not expired, the UE (100) selects the AI model based on AMI 2. The UE (100) also estimatesW1,W2, andWffrom the precoder and performs AI basedW2compression. The UE (100) shares the encodedW2with the network apparatus (200), which performs the decoding of the encodedW2. If the timer expires, then the UE (100) selects the AI model based on AMI 2 and encodes theW2using AMI 2. The UE (100) also sharesW1andWfalong with the encodedW2with the network apparatus (202), which performs the decoding. The UE (100) then resets the timer and the cycle continues.

[0145] In order to achieve different grades of reconstruction accuracy, 3GPP specifies various Parameter Combinations(param_comb) that specify the number of bits that can be used for feedback. The network apparatus (200) specifies the param_comb to the UE (100) via RRC signalling. Table 5 depicts the values of feedback bits, M_v and L associated with some of the param_comb. The AI model is trained for each layer and each param_comb. Thus, there would exist multiplicity of AI models depending on the param_comb, layer, L and M_v and feedback_bits values. Thus, in order to uniquely identify each of these models, the AMI parameter is used.

[0146] [Table 5] Values for different param_comb

[0147]

[0148] Fig. 13 is a sequence diagram that illustratesW2compression when models are stored at the network apparatus (200) according to an embodiment as disclosed herein. As shown in the sequence diagram, the UE (100) and the network apparatus (200) are in communication with each other. The network apparatus (200) prepares the AMI based on the LUT. The network apparatus (200) shares the AI model along with the AMI with the UE (100). The UE (100) configures the AI model using the AMI and performs the AI basedW2encoding. The UE (100) shares the encodedW2with the network apparatus (200), which the network apparatus (200) uses to perform the AI based decoding.

[0149] Fig. 14 is a sequence diagram that illustratesW2compression when models are stored at the UE (100) according to an embodiment as disclosed herein. As shown in the sequence diagram, the UE (100) and the network apparatus (200) are in communication with each other. The network apparatus (200) prepares the AMI based on the LUT and shares it with the UE (100) via RRC signaling. The UE (100) selects the appropriate model based on AMI and configures the same. The UE (100) performs the AI basedW2encoding. The UE (100) shares the encodedW2with the network apparatus (200), which the network apparatus (200) uses to perform the AI based decoding.

[0150] The following changes to the RRC configuration are proposed below (marked in Bold):

[0151] csi-ReportConfigToAddModList: 1 item

[0152] Item 0

[0153] CSI-ReportConfig

[0154] reportConfigId: 0

[0155] resourcesForChannelMeasurement: 1

[0156] csi-IM-ResourcesForInterference: 2

[0157] reportConfigType: aperiodic (3)

[0158] reportFreqConfiguration

[0159] cqi-FormatIndicator: widebandCQI (0)

[0160] pmi-FormatIndicator: widebandPMI (0)

[0161] csi-ReportingBand: subbands9 (6)

[0162] subbands9: ff80 [bit length 9, 7 LSB pad bits, 1111 1111 1... .... decimal value 511]

[0163] AIconfiguration: 2 items

[0164] Item 0

[0165] AMI

[0166] PC : 1

[0167] Layer: 1

[0168] L : 2

[0169] Mv: 2

[0170] Feedback_bits : 32

[0171] Item 1

[0172] Periodicity_W1_Wf: 20

[0173] timeRestrictionForChannelMeasurements: notConfigured (1)

[0174] timeRestrictionForInterferenceMeasurements: notConfigured (1)

[0175] codebookConfig

[0176] codebookType: type1 (0)

[0177] type1

[0178] subType: typeI-SinglePanel (0)

[0179] typeI-SinglePanel

[0180] nrOfAntennaPorts: moreThanTwo (1)

[0181] moreThanTwo

[0182] n1-n2: two-one-TypeI-SinglePanel-Restriction (0)

[0183] two-one-TypeI-SinglePanel-Restriction: ff [bit length 8, 1111 1111 decimal value 255]

[0184] typeI-SinglePanel-ri-Restriction: 0f [bit length 8, 0000 1111 decimal value 15]

[0185] codebookMode: 1

[0186] Figs. 15A and 15B are flow diagrams that illustrate a method for utilizing time correlation in AI based sub-band CSI systems by the UE (100) according to an embodiment as disclosed herein. The method includes steps (1502-1526). Each step is explained in further detail below.

[0187] At step (1502), the UE (100) receives a RRC configuration along with AI configuration parameters and a CSI-RS from the network apparatus (200). For instance, the AI configuration parameters include at least one of a LUT provided by the network apparatus (200), an AMI obtained based on the LUT, a periodicity associated with theW1and theWf(PW1Wf), a periodicity associated with theW2(PW2), or parameter combinations (param_comb). The parameter combinations specify a number of feedback bits allocated across different layers of the UE (100).

[0188] At step (1504), the UE (100) estimates a precoder based on the CSI-RS received from the network apparatus (200). Upon receiving the CSI-RS, the UE (100) processes the reference signal to extract relevant channel quality information. Based on this channel quality information, the UE (100) determines or estimates the precoder. The precoder optimizes the transmission or reception performance in accordance with the prevailing channel conditions.

[0189] At step (1506), the UE (100) decomposes the precoder into a wideband spatial component matrix (W1), a wideband coefficient matrix (W2), and a wideband frequency component matrix (Wf). TheW1matrix represents the spatial characteristics of the wideband channel across antenna elements. TheW1matrix captures how signals propagate through different antenna parts. TheW2matrix contains the complex path gain coefficients of the wideband channel. TheWfmatrix represents the frequency-domain structure of the wideband channel.

[0190] At step (1508), the UE (100) encodes theW1matrix and theWfmatrix to obtain an encodedW1matrix and an encodedWfmatrix. The encodedW1matrix and the encodedWfmatrix are obtained using a non-AI encoding approach or conventional techniques.

[0191] At step (1510), the UE (100) encodes theW2using an AI model (AMI 1 or AMI 2) to obtain an encodedW2matrix. The AI model is selected based on the AI configuration parameters and the CSI-RS received from the network apparatus (200).

[0192] At step (1512), the UE (100) distributes feedback bits across different layers of the UE (100) based on the LUT and a param_comb provided within the AI configuration parameters. The feedback bits specify a number of bits available for AI based compression of theW2. The feedback bits are quantized information that can represent a channel quality, spatial properties, and a recommended precoding.

[0193] At step (1514), the UE (100) compresses theW2using an AMI 1 that is provided within the AI configuration parameters to obtain the encodedW2matrix. The AMI 1 corresponds to the AI model that uses all feedback bits within a slot for performing the compression of theW2. Else, at step (1516), the UE (100) compresses theW2using an AMI 2 that is provided within the AI configuration parameters to obtain the encodedW2matrix. The AMI 2 corresponds to the AI model that uses only a portion of the feedback bits within a slot for performing the compression of theW2.

[0194] At step (1518), the UE (100) transmits at least one of the encodedW2matrix, the encodedW1matrix, or the encodedWfmatrix to the network apparatus (200). This transmission is based on a timer associated with the AI configuration parameters.

[0195] At step (1520), the UE (100) generates the timer based on a periodicity associated with theW1and theWf(PW1Wf) provided within the AI configuration parameters. At step (1522), the UE (100) detects whether the timer has expired. The timer expires when a value of the timer is equivalent of the PW1Wf provided within the AI configuration parameters. At step (1524), the UE (100) transmits the encodedW2matrix obtained using the AMI 1 when the timer has not expired. Else, at step (1526), the UE (100) transmits the encodedW2matrix obtained using the AMI 2, the encodedW1matrix, and the encodedWfmatrix to the network apparatus (200), when the timer has expired.

[0196] Fig. 16 is a flow diagram that illustrates a method for utilizing time correlation in AI based sub-band CSI systems by the network apparatus (200) according to an embodiment as disclosed herein. The method includes steps (1602-1622). Each step is explained in further detail below.

[0197] At step (1602), the network apparatus (200) receives an encodedW1matrix, an encodedW2matrix, and an encodedWfmatrix from the UE (100). At step (1604), the network apparatus (200) decodes the encodedW1matrix to obtain a decodedW1matrix and the encodedWfmatrix to obtain a decodedWfmatrix. The encodedW1matrix and the encodedWfmatrix are decoded using a non-AI decoding approach.

[0198] At step (1606), the network apparatus (200) decodes the encodedW2matrix using an AI model (AMI 1 or AMI 2). The AI model is selected based on AI configuration parameters and the CSI-RS to obtain a decodedW2matrix. At step (1608), the network apparatus (200) performs an offline training using a predefined dataset or an online training using a localized dataset. The offline training and the online training are performed to determine how to distribute feedback bits across different layers of the UE (100).

[0199] At step (1610), the network apparatus (200) generates a LUT for distributing the feedback bits across different layers of the UE (100) based on the online training and the offline training performed. The LUT is a pre-computed table that stores the feedback bits in an array of stored values that are indexed.

[0200] At step (1612), the network apparatus (200) generates an AMI 1 and an AMI 2 based on the LUT generated. The AMI 1 corresponds to the AI model that uses all feedback bits within a slot and the AMI 2 corresponds to the AI model that uses a portion of the feedback bits within a slot. At step (1614), the network apparatus (200) decodes the encodedW2matrix using the AMI 1 or the AMI 2 to obtain the decodedW2matrix. For instance, the encodedW2matrix is decoded using AMI 1 when a timer (PW1Wf timer) has not expired. Else, the encodedW2matrix is decoded using AMI 2 when the timer has expired.

[0201] At step (1616), the network apparatus (200) performs a reconstruction of a precoder. This reconstruction is performed based on the decodedW1matrix, the decodedW2matrix, and the decodedWfmatrix obtained. The network apparatus (200) reconstructs the precoder based on the feedback bits provided within the decodedW1matrix, the decodedW2matrix, and the decodedWfmatrix.

[0202] At step (1618), the network apparatus (200) generates a combined matrix (W) based on a product of the decodedW1matrix, the decodedW2matrix, and the decodedWfmatrix. The combined matrix (W) is obtained using the equation:W=W1W2Wf. At step (1620), the network apparatus (200) reconstructs the precoder based on the W obtained.

[0203] At step (1622), the network apparatus (200) schedules data transmissions with the UE (100) upon reconstruction of the precoder. The data transmission can be scheduled based on factors such as rank indicator (RI), channel quality indicator (CQI), and the like. The RI indicates how many spatial layers can be used during the transmission. The CQI indicates the modulation and coding schemes to match the channel quality. Thus, scheduling data transmissions upon reconstruction of the precoder enhances throughput and reliability of the data transmitted between the UE (100) and the network apparatus (200).

[0204] The foregoing description of the specific embodiments will so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and or adapt for various applications such specific embodiments without departing from the generic concept, and, therefore, such adaptations and modifications are intended to be comprehended within the meaning and range of equivalents of the disclosed embodiments. It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the scope of the embodiments as described herein.

Claims

A method performed by a user equipment (UE) in a wireless communication system supporting artificial intelligence (AI) based sub-band channel state information (CSI), the method comprising:receiving a radio resource control (RRC) configuration along with AI configuration parameters and a CSI reference signal (CSI-RS) from a network apparatus;estimating a precoder based on the CSI-RS received from the network apparatus;decomposing the precoder into a wideband spatial component matrix (W1), a wideband coefficient matrix (W2), and a wideband frequency component matrix (Wf);encoding the W1 and the Wf to obtain an encoded W1 and an encoded Wf, wherein the encoded W1 and the encoded Wf are obtained using a non-AI encoding approach;encoding the W2 using an AI model selected based on the AI configuration parameters and the CSI-RS received from the network apparatus to obtain an encoded W2; andtransmitting at least one of the encoded W2, the encoded W1, or the encoded Wf to the network apparatus based on a timer associated with the AI configuration parameters.The method of claim 1, wherein the AI configuration parameters comprise at least one of a look-up table (LUT) provided by the network apparatus, an AI model indicator (AMI) obtained based on the LUT, a periodicity associated with the W1 and the Wf (PW1Wf), a periodicity associated with the W2 (PW2), or parameter combinations (param_comb) that specify a number of feedback bits allocated across different layers of the UE.The method of claim 1, wherein encoding the W2 comprises:distributing feedback bits across different layers of the UE based on a look-up table (LUT) and a parameter combinations (param_comb) provided within the AI configuration parameters, wherein the feedback bits specify a number of bits available for AI based compression of the W2; andperforming one of:　compressing the W2 using an AI model indicator 1 (AMI 1) that is provided within the AI configuration parameters to obtain the encoded W2, wherein the AMI 1 corresponds to an AI model that uses all feedback bits within a slot for performing the compression of the W2; and　compressing, by the UE (100), the W2 using an AMI 2 that is provided within the AI configuration parameters to obtain the encoded W2, wherein the AMI 2 corresponds to an AI model that uses only a portion of the feedback bits within a slot for performing the compression of the W2.The method of claim 3, wherein transmitting the at least one of the encoded W2, the encoded W1, or the encoded Wf to the network apparatus based on the timer associated with the AI configuration parameters comprises:generating the timer based on a periodicity associated with the W1 and the Wf (PW1Wf) provided within the AI configuration parameters;detecting whether the timer expires, wherein the timer expires when a value of the timer is equivalent of the PW1Wf provided within the AI configuration parameters; andperforming one of:　transmitting the encoded W2 obtained using the AMI 1 when the timer does not expire; and　transmitting the encoded W2 obtained using the AMI 2, the encoded W1, and the encoded Wf to the network apparatus, when the timer expires.A method performed by a network apparatus in a wireless communication system supporting artificial intelligence (AI) based sub-band channel state information (CSI), the method comprising:receiving an encoded wideband spatial component matrix (W1), an encoded wideband coefficient matrix (W2), and an encoded wideband frequency component matrix (Wf) from a user equipment (UE);decoding the encoded W1 to obtain a decoded W1 and the encoded Wf to obtain a decoded Wf, wherein the encoded W1 and the encoded Wf are decoded using a non-AI decoding approach;decoding the encoded W2 using an AI model selected based on AI configuration parameters and a CSI reference signal (CSI-RS) to obtain a decoded W2;performing a reconstruction of a precoder based on the decoded W1, the decoded W2, and the decoded Wf; andscheduling a data transmission between the UE and the network apparatus based on the reconstructed precoder.The method of claim 5, wherein decoding the encoded W2 comprises:performing an offline training using a predefined dataset or an online training using a localized dataset, wherein the offline training and the online training are performed to determine how to distribute feedback bits across different layers of the UE;generating a look-up table (LUT) for distributing the feedback bits across the different layers of the UE based on the online training and the offline training performed;generating an AI model indicator (AMI) 1 and an AMI 2 based on the LUT generated, wherein the AMI 1 corresponds to an AI model that uses all feedback bits within a slot and the AMI 2 corresponds to an AI model that uses a portion of the feedback bits within a slot; anddecoding the encoded W2 using the AMI 1 or the AMI 2 to obtain the decoded W2.The method of claim 6, wherein decoding the encoded W2 using the AMI 1 or the AMI 2 comprises:decoding the encoded W2 to obtain the decoded W2 using the AMI 2 when a timer associated with the UE expires; ordecoding the encoded W2 to obtain the decoded W2 using the AMI 1 when the timer does not expire.The method of claim 7, further comprising:generating a radio resource control (RRC) configuration including the AI configuration parameters, wherein the AI configuration parameters comprise at least one of the LUT, the AMI 1, or the AMI 2 obtained along with distribution of the feedback bits across the different layers of the UE; andtransmitting the RRC configuration including the AI configuration parameters to the UE.The method of claim 5, wherein performing the reconstruction of the precoder comprises:generating a combined matrix (W) based on a product of the decoded W1, the decoded W2, and the decoded Wf; andreconstructing the precoder based on the combined matrix (W).A user equipment (UE) in a wireless communication system supporting artificial intelligence (AI) based sub-band channel state information (CSI), the UE comprising:a processor;a memory coupled to the processor; anda controller communicatively coupled to the processor and the memory, wherein the controller:receives a radio resource control (RRC) configuration along with AI configuration parameters and a CSI reference signal (CSI-RS) from a network apparatus;estimates a precoder based on the CSI-RS received from the network apparatus;decomposes the precoder into a wideband spatial component matrix (W1), a wideband coefficient matrix (W2), and a wideband frequency component matrix (Wf);encodes the W1 and the Wf to obtain an encoded W1 and an encoded Wf, wherein the encoded W1 and the encoded Wf are obtained using a non-AI encoding approach;encodes the W2 using an AI model selected based on the AI configuration parameters and the CSI-RS received from the network apparatus to obtain an encoded W2; andtransmits at least one of the encoded W2, the encoded W1, or the encoded Wf to the network apparatus based on a timer associated with the AI configuration parameters.The UE of claim 10, wherein encoding the W2 comprises:distributing feedback bits across different layers of the UE based on a look-up table (LUT) and a parameter combinations (param_comb) provided within the AI configuration parameters, wherein the feedback bits specify a number of bits available for AI based compression of the W2; andperforming one of:　compressing the W2 using an AI model indicator 1 (AMI 1) that is provided within the AI configuration parameters to obtain the encoded W2, wherein the AMI 1 corresponds to an AI model that uses all feedback bits within a slot for performing the compression of the W2; and　compressing the W2 using an AMI 2 that is provided within the AI configuration parameters to obtain the encoded W2, wherein the AMI 2 corresponds to an AI model that uses only a portion of the feedback bits within a slot for performing the compression of the W2.The UE of claim 11, wherein transmitting the at least one of the encoded W2, the encoded W1, or the encoded Wf to the network apparatus based on the timer associated with the AI configuration parameters comprises:generating the timer based on a periodicity associated with the W1 and the Wf (PW1Wf) provided within the AI configuration parameters;detecting whether the timer expires, wherein the timer expires when a value of the timer is equivalent of the PW1Wf provided within the AI configuration parameters; andperforming one of:　transmitting the encoded W2 obtained using the AMI 1 when the timer does not expire; and　transmitting the encoded W2 obtained using the AMI 2, the encoded W1, and the encoded Wf to the network apparatus, when the timer expires.A network apparatus in a wireless communication system supporting artificial intelligence (AI) based sub-band channel state information (CSI), the network apparatus comprising:a processor;a memory coupled to the processor; anda controller communicatively coupled to the processor and the memory, wherein the controller:receives an encoded W1, an encoded W2, and an encoded Wf from a user equipment (UE);decodes the encoded W1 to obtain a decoded W1 and the encoded Wf to obtain a decoded Wf, wherein the encoded W1 and the encoded Wf are decoded using a non-AI decoding approach;decodes the encoded W2 using an AI model selected based on AI configuration parameters and a CSI reference signal (CSI-RS) to obtain a decoded W2;performs a reconstruction of a precoder based on the decoded W1, the decoded W2, and the decoded Wf obtained; andschedules a data transmission between the UE and the network apparatus based on the reconstructed precoder.The network apparatus of claim 13, wherein decoding the encoded W2 comprises:performing an offline training using a predefined dataset or an online training using a localized dataset, wherein the offline training and the online training are performed to determine how to distribute feedback bits across different layers of the UE;generating a look-up table (LUT) for distributing the feedback bits across the different layers of the UE based on the online training and the offline training performed;generating an AI model indicator (AMI) 1 and an AMI 2 based on the LUT generated, wherein the AMI 1 corresponds to an AI model that uses all feedback bits within a slot and the AMI 2 corresponds to an AI model that uses a portion of the feedback bits within a slot; anddecoding the encoded W2 using the AMI 1 or the AMI 2 to obtain the decoded W2.The network apparatus of claim 14, wherein decoding the encoded W2 using the AMI 1 or the AMI 2 comprises:decoding the encoded W2 to obtain the decoded W2 using the AMI 2 when a timer associated with the UE expires; ordecoding the encoded W2 to obtain the decoded W2 using the AMI 1 when the timer does not expire.

Images