Data collection method and process for AI-based CSI prediction

Abstract

Apparatus, systems, and methods for AI-based CSI feedback with CSI prediction, including systems, methods, and mechanisms for encoding, by a user equipment (UE), a data collection request for training a CSI prediction AI model, and sending the data collection request to a next generation Node B (gNB). The UE may decode a data collection request response received from the gNB. The UE may perform a UE measurement of a CSI reference signal (CSI-RS) based on a data collection request response from the gNB for training a CSI prediction AI model. The UE may encode a data collection stop request for transmission to the gNB. The UE may store UE measurements of the CSI-RS in a memory coupled to one or more processors.

Description

Technical Field

[0001] This invention relates to wireless communication, and more specifically, to data collection apparatus, systems, and methods for predicting channel state information (CSI) based on artificial intelligence (AI), including systems, methods, and mechanisms for a UE to initiate data collection during 5G NR communication to train an AI prediction model.

[0002] Description of related technologies The use of wireless communication systems is growing rapidly. In recent years, wireless devices, such as smartphones and tablets, have become increasingly sophisticated. In addition to supporting telephone calls, many mobile devices now offer access to the internet, email, text messaging, and navigation using the Global Positioning System (GPS), and can operate complex applications that utilize these functions. Furthermore, many different wireless communication technologies and standards exist.

[0003] Long Term Evolution (LTE), also known as Evolved Universal Terrestrial Radio Access Network (E-UTRAN), has become the technology of choice for most wireless network operators worldwide, providing mobile broadband data and high-speed internet access to their user base. LTE was first proposed in 2004 and standardized for the first time in 2008. Since then, with the exponential growth in the use of wireless communication systems, the demand from wireless network operators for higher capacity to support higher density mobile broadband users has also increased. Therefore, research into new radio access technologies began in 2015, and in 2017, the first version of the 3GPP Fifth Generation New Radio (5G NR) was standardized. The fifth-generation mobile network or fifth-generation wireless system is called 3GPP NR (also known as 5G-NR or NR-5G, i.e., 5G New Radio, or simply NR). NR provides higher capacity for higher density mobile broadband users while supporting device-to-device, ultra-reliable and massive machine-type communications, as well as lower latency and lower battery consumption than the LTE standard.

[0004] Compared to LTE, 5G-NR offers higher capacity for higher-density mobile broadband users, while also supporting device-to-device, ultra-reliable, and massive machine-type communications with lower latency and / or lower battery consumption. Furthermore, NR allows for more flexible UE scheduling compared to current LTE. Therefore, ongoing efforts are underway to develop 5G-NR to leverage the potentially higher throughput at higher frequencies.

[0005] One aspect of wireless communication systems (e.g., systems for NR cellular wireless communication) is the transmission and measurement of reference signals, including channel state information reference signals (CSI-RS). Summary of the Invention

[0006] The implementation scheme relates to wireless communication, and more specifically, to data collection apparatus, systems, and methods for AI-based CSI prediction, including systems, methods, and mechanisms for a UE to initiate and collect data during 5G NR communication to train an AI prediction model.

[0007] For example, in some implementations, the UE may send a data collection request to a base station (e.g., a next-generation Node B (gNB)) to train a CSI prediction artificial intelligence (AI) model. The UE may receive a data collection request response from the gNB. The UE may then perform UE measurements of the CSI reference signal (CSI-RS) based on the data collection request response from the gNB to train the CSI prediction AI model.

[0008] The UE may send a data collection stop request to the gNB. The UE may store the CSI-RS measurements in a memory coupled to one or more processors. The configuration of the data collection request may include at least a request to collect input data and output data from the CSI prediction AI model so that the gNB can train the CSI prediction AI model. The configuration of the input data for the CSI prediction AI model includes the total number of samples collected in the time domain used as input to the CSI prediction AI model, the measurement distance between one or more samples in the total number of samples collected in the time domain, and information for classifying the total number of samples. The configuration of the output data for the CSI prediction AI model may include the total number of samples to be predicted and the prediction distance between one or more samples in the total number of samples.

[0009] The techniques described herein can be implemented in and / or used with a variety of different types of devices, including but not limited to base stations, access points, cellular phones, tablets, wearable computing devices, portable media players, automobiles, and various other computing devices.

[0010] The present invention is intended to provide a brief overview of some of the subjects described in this document. Therefore, it should be understood that the above features are merely illustrative and should not be construed as narrowing the scope or substance of the subjects described herein in any way. Other features, aspects, and advantages of the subjects described herein will become apparent from the following detailed description, drawings, and claims. Attached Figure Description

[0011] A better understanding of the subject matter can be obtained by considering the following detailed description of various embodiments in conjunction with the accompanying drawings, in which: Figure 1A Example wireless communication systems according to some implementation schemes are illustrated.

[0012] Figure 1B Examples of base stations and access points communicating with user equipment (UE) devices according to some implementation schemes are illustrated.

[0013] Figure 2 Example block diagrams of base stations according to some implementation schemes are shown.

[0014] Figure 3 Example block diagrams of servers according to some implementation schemes are shown.

[0015] Figure 4 Example block diagrams of a UE according to some implementation schemes are shown.

[0016] Figure 5 Example block diagrams of cellular communication circuits according to some implementation schemes are shown.

[0017] Figure 6A Examples of 5G network architectures according to some implementation schemes are illustrated, which combine both 3GPP (e.g., cellular) and non-3GPP (e.g., noncellular) access to 5GCN.

[0018] Figure 6B Examples of 5G network architectures according to some implementation schemes are illustrated, which combine dual 3GPP (e.g., LTE and 5G NR) access to 5GCN and non-3GPP access.

[0019] Figure 7 Examples of baseband processor architectures for UEs according to some implementation schemes are illustrated.

[0020] Figure 8 Examples of devices according to some implementation schemes are shown.

[0021] Figure 9 Example baseband circuits according to some implementation schemes are illustrated.

[0022] Figure 10 Examples of control plane protocol stacks based on some implementation schemes are shown.

[0023] Figure 11 An example is shown based on CSI feedback from 3GPP Release 17.

[0024] Figure 12 An example is shown based on CSI feedback from 3GPP Release 18.

[0025] Figure 13 Examples of CSI feedback with CSI predictions are illustrated according to some implementation schemes.

[0026] Figure 14A Examples of AI models for CSI prediction based on some implementation schemes are shown.

[0027] Figure 14B Examples of one-dimensional LSTM AI models for CSI prediction using the time domain are illustrated according to some implementation schemes.

[0028] Figure 14C Examples of two-dimensional convolutional neural network (CNN) AI models for CSI prediction using the time and frequency domains are illustrated according to some implementation schemes.

[0029] Figure 14D Examples of three-dimensional convolutional neural network (CNN) AI models for CSI prediction using the time domain, frequency domain, and spatial (antenna) domain are illustrated according to some implementation schemes.

[0030] Figure 15 Example signaling timing diagrams for data collection for AI-based CSI prediction are illustrated according to some implementation schemes.

[0031] Figure 16 Examples of UE data collection requests for AI-based CSI prediction data collection are illustrated according to some implementation schemes.

[0032] Figure 17 An example is illustrated where a base station transmits data collection commands according to some implementation schemes for data collection for AI-based CSI prediction.

[0033] Figure 18 Another example of an AI model for CSI prediction based on some implementation schemes is shown.

[0034] Figure 19 A block diagram illustrating an example of a data collection method for predicting channel state information (CSI) based on artificial intelligence (AI) according to some implementation schemes is shown.

[0035] Although the features described herein may be subject to various modifications and alternatives, specific embodiments thereof are shown by way of example in the accompanying drawings and described in detail herein. However, it should be understood that the drawings and their detailed description are not intended to limit one to the specific forms disclosed, but rather to cover all modifications, equivalents, and alternatives falling within the substance and scope of the subject matter as defined by the appended claims. Detailed Implementation

[0036] The following is a glossary of terms used in this disclosure: Memory media—any of various types of nontransitory memory devices or storage devices. The term "memory media" is intended to include mounting media, such as CD-ROMs, floppy disks, or magnetic tape devices; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory; magnetic media, such as hard disk drives or optical storage devices; registers or other similar types of memory elements, etc. Memory media may also include other types of nontransitory memory or combinations thereof. Furthermore, memory media may reside in a first computer system executing a program, or may reside in a different second computer system connected to the first computer system via a network such as the Internet. In the latter example, the second computer system may provide program instructions to the first computer for execution. The term "memory media" may include two or more memory media residing in different locations in different computer systems connected via, for example, a network. Memory media may store program instructions (e.g., embodied in a computer program) that can be executed by one or more processors.

[0037] Carrier media—memory media as described above, and physical transmission media such as buses, networks, and / or other physical transmission media that transmit signals such as electrical signals, electromagnetic signals, or digital signals.

[0038] Programmable hardware elements—including a variety of hardware devices comprising multiple programmable functional blocks connected via programmable interconnects. Examples include FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), FPOAs (Field-Programmable Object Arrays), and CPLDs (Complex PLDs). Programmable functional blocks can range from fine-grained (combinational logic or lookup tables) to coarse-grained (arithmetic logic units or processor cores). Programmable hardware elements may also be referred to as “configurable logic units.”

[0039] Computer system (or computer) — any of the various types of computing or processing systems, including personal computer systems (PCs), mainframe computer systems, workstations, network appliances, internet-connected appliances, personal digital assistants (PDAs), television systems, grid computing systems, or other devices or combinations thereof. In general, the term "computer system" can be broadly defined to encompass any device (or combination of devices) having at least one processor that executes instructions from a memory medium.

[0040] User equipment (UE) (or “UE device”) — any of various types of computer system devices that are mobile or portable and perform wireless communication. Examples of UE devices include mobile phones or smartphones (e.g., iPhone). ™Based on Android ™ Telephones), portable gaming devices (e.g., Nintendo DS) ™ PlayStation Portable ™ Gameboy Advance ™ iPhone ™ ), laptops, wearable devices (e.g., smartwatches, smart glasses), PDAs, portable internet devices, music players, data storage devices, other handheld devices, unmanned aerial vehicles (UAVs) (e.g., drones), UAV controllers (UACs), etc. Generally speaking, the term "UE" or "UE device" can be broadly defined to encompass any electronic device, computing device, and / or telecommunications device (or combination of devices) that is easily transportable by the user and capable of wireless communication.

[0041] Base station—The term “base station” has the full range of its common meaning and includes at least a wireless communication station that is installed in a fixed location and is used for communication as part of a wireless telephone system or radio system.

[0042] A processing element (or processor) is a component or combination of components capable of performing the functions of a device such as a user equipment or cellular network device. A processing element may include, for example: a processor and associated memory, portions or circuitry of individual processor cores, an entire processor core, a processor array, circuitry such as an ASIC (Application-Specific Integrated Circuit), programmable hardware components such as a Field-Programmable Gate Array (FPGA), and any combination thereof.

[0043] A channel is a medium used to transmit information from a transmitter to a receiver. It should be noted that because the characteristics of the term "channel" can vary depending on the wireless protocol, the term "channel" as used herein can be considered to be used in a standard manner consistent with the type of device to which the term is referenced. In some standards, channel width can be variable (e.g., depending on device capabilities, band conditions, etc.). For example, LTE can support scalable channel bandwidths from 1.4 MHz to 20 MHz. In contrast, WLAN channels can be 22 MHz wide, while Bluetooth channels can be 1 MHz wide. Other protocols and standards may include different definitions of channels. Furthermore, some standards may define and use multiple types of channels, for example, different channels for uplink or downlink and / or different channels for different purposes such as data, control information, etc.

[0044] Frequency band—The term “frequency band” has the full range of its general meaning and includes at least a segment of spectrum (e.g., radio frequency spectrum) in which a channel is used or reserved for the same purpose.

[0045] Wi-Fi—The term “Wi-Fi” (or WiFi) has the full range of its usual meaning and includes at least wireless communication networks or RATs, which are provided by and through wireless LAN (WLAN) access points to provide connectivity to the Internet. Most modern Wi-Fi networks (or WLAN networks) are based on the IEEE 802.11 standard and are marketed under the name “Wi-Fi.” Wi-Fi (WLAN) networks are different from cellular networks.

[0046] "3GPP access" refers to access technologies (e.g., radio access technologies) specified by the Third Generation Partnership Project (3GPP) standards. These access technologies include, but are not limited to, GSM / GPRS, LTE, LTE-A, and / or 5G NR. Generally speaking, 3GPP access refers to various types of cellular access technologies.

[0047] Non-3GPP access refers to any access technology (e.g., radio access technologies) not specified by 3GPP standards. These accesses include, but are not limited to, WiMAX, CDMA2000, Wi-Fi, WLAN, and / or fixed networks. Non-3GPP access can be categorized into two types: "trusted" and "untrusted." Trusted non-3GPP access can interact directly with the Evolved Packet Core (EPC) and / or 5G Core (5GC), while untrusted non-3GPP access interoperates with the EPC / 5GC via network entities such as Evolved Packet Data Gateways and / or 5G NR Gateways. Generally speaking, non-3GPP access refers to various types of non-cellular access technologies.

[0048] Automatic—means that an action or operation is performed by a computer system (e.g., software executed by the computer system) or device (e.g., circuits, programmable hardware elements, ASICs, etc.) without requiring direct specification or execution of the action or operation by user input. Therefore, the term "automatically" is the opposite of an operation performed or specified manually by a user, where the user provides input to directly perform the operation. An automatic process may be initiated by user-provided input, but the subsequent actions performed "automatically" are not specified by the user; that is, they are not performed "manually," where the user specifies each action to be performed. For example, a user filling out a form by selecting each field and providing input specifying information (e.g., by typing information, selecting a checkbox, radio selection, etc.) is considered manually filling out the form, even if the computer system can update the form in response to the user's actions. The form can be automatically filled out by a computer system, where the computer system (e.g., software executed on the computer system) analyzes the fields of the form and fills out the form without any user input specifying answers for the fields. As indicated above, the user may invoke the automatic filling of the form but does not participate in the actual filling of the form (e.g., the user does not manually specify answers for the fields, but they are completed automatically). This manual provides various examples of operations that can be performed automatically in response to actions taken by the user.

[0049] Approximately—means a value close to the correct or precise value. For example, approximately could mean a value within 1% to 10% of the precise (or expected) value. However, it should be noted that the actual threshold (or tolerance) can be application-dependent. For example, in some implementations, “approximately” may mean within 0.1% of some specified or expected value, while in various other implementations, the threshold may be, for example, 2%, 3%, 5%, etc., depending on the expectations or requirements of a particular application.

[0050] Concurrency refers to the parallel execution or implementation of tasks, processes, or programs in a manner that at least partially overlaps. For example, concurrency can be achieved using “strong” or strict parallelism, where tasks are executed in parallel (at least partially) on corresponding computing elements; or using “weak parallelism,” where tasks are executed in an interleaved manner (e.g., by time multiplexing of execution threads).

[0051] Various components can be described as being "configured" to perform one or more tasks. In this context, "configured" is a broad expression generally meaning "having a structure" that performs one or more tasks during operation. Therefore, a component can be configured to perform a task even when it is not currently performing one (e.g., a set of electrical conductors can be configured to electrically connect one module to another, even when the two modules are not connected). In some contexts, "configured" can be a broad expression generally meaning "having a circuit" that performs one or more tasks during operation. Therefore, a component can be configured to perform a task even when it is not currently powered on. Generally, the circuit forming the structure corresponding to "configured" can include hardware circuitry.

[0052] For ease of description, various components may be described as performing one or more tasks. Such descriptions should be interpreted as including the phrase "configured to". Statements describing a component as configured to perform one or more tasks are explicitly intended not to invoke the interpretation of 35 USC § 112(f) for that component.

[0053] Figure 1A and Figure 1B Communication system Figure 1A A simplified exemplary wireless communication system according to some implementation schemes is shown. It should be noted that... Figure 1A The system described herein is merely one example of a possible system, and the features of this disclosure can be implemented in any of a variety of systems as needed.

[0054] As shown in the figure, the example wireless communication system includes a base station 102A, which communicates with one or more user equipments 106A, 106B to 106N via a transmission medium. In this document, the user equipment may be referred to as "User Equipment" (UE). Therefore, user equipment 106 is referred to as UE or UE device.

[0055] Base station (BS) 102A may be a transceiver base station (BTS) or a cell site (“cellular base station”), and may include hardware that enables wireless communication with UE 106A to UE 106N.

[0056] The communication area (or coverage area) of a base station may be referred to as a "cell". Base station 102A and UE 106 can be configured to communicate via a transmission medium using any of a variety of Radio Access Technologies (RATs), also known as wireless communication technologies or telecommunications standards, such as GSM, UMTS (as associated with, for example, WCDMA or TD-SCDMA air interfaces), LTE, LTE-A Advanced, 5G New Radio (5G NR), HSPA, 3GPP2 CDMA2000 (e.g., 1xRTT, 1xEV-DO, HRPD, eHRPD), etc. Note that if base station 102A is implemented in an LTE (E-UTRAN) context, its alternative location may be referred to as an "eNodeB" or "eNB". Note that if base station 102A is implemented in a 5G NR context, its alternative location may be referred to as a "gNodeB" or "gNB".

[0057] As shown in the figure, base station 102A can also be configured to communicate with network 100 (e.g., in various possibilities, the core network of a cellular service provider, telecommunications networks such as the Public Switched Telephone Network (PSTN), and / or the Internet). Therefore, base station 102A facilitates communication between user equipments and / or between user equipments and network 100. Specifically, cellular base station 102A can provide UE 106 with various telecommunications capabilities, such as voice, SMS, and / or data services.

[0058] Base station 102A and other similar base stations (such as base stations 102B, ..., 102N) operating according to the same or different cellular communication standards can therefore be provided as a network of cells that can provide continuous or nearly continuous overlapping services to UE 106A-N and similar devices over a geographical area via one or more cellular communication standards.

[0059] Therefore, although base station 102A can act as such Figure 1A The example illustrates the "serving cells" of UEs 106A to 106N, but each UE 106 may also be able to receive signals (and possibly within its communication range) from one or more other cells (which may be provided by base stations 102B to 102N and / or any other base stations), which may be referred to as "neighboring cells." Such cells may also facilitate communication between user equipments and / or between user equipments and network 100. These cells may include "macro" cells, "micro" cells, "pecimen" cells, and / or any other cells of various other granularities providing a service area size. For example, Figure 1A The illustrated base stations 102A to 102B may be macro cells, while base station 102N may be a micro cell. Other configurations are also possible.

[0060] In some implementations, base station 102A may be a next-generation base station, such as a 5G New Radio (5G NR) base station or a “gNB”. In some implementations, the gNB may be connected to a legacy evolved packet core (EPC) network and / or to an NR core (NRC) network. Furthermore, a gNB cell may include one or more transition and receive points (TRPs). Additionally, a UE capable of operating according to 5G NR may be connected to one or more TRPs within one or more gNBs.

[0061] It should be noted that UE 106 may be able to communicate using multiple wireless communication standards. For example, UE 106 may be configured to communicate using wireless networking (e.g., Wi-Fi) and / or peer-to-peer wireless communication protocols (e.g., Bluetooth, Wi-Fi peer-to-peer, etc.) other than at least one cellular communication protocol (e.g., GSM, UMTS (associated with, for example, WCDMA or TD-SCDMA air interfaces), LTE, LTE-A, 5G NR, HSPA, 3GPP2 CDMA2000 (e.g., 1xRTT, 1xEV-DO, HRPD, eHRPD, etc.)). If desired, UE 106 may also be configured, or alternatively, to communicate using one or more Global Navigation Satellite Systems (GNSS, such as GPS or GLONASS), one or more mobile television broadcasting standards (e.g., ATSC-M / H or DVB-H) and / or any other wireless communication protocol. Other combinations of wireless communication standards (including more than two wireless communication standards) are also possible.

[0062] Figure 1B User equipment 106 (e.g., one of devices 106A to 106N) communicating with base station 102 and access point 112 according to some embodiments is illustrated. UE 106 can be a device with cellular and non-cellular communication capabilities (e.g., Bluetooth, Wi-Fi, etc.), such as a mobile phone, handheld device, computer or tablet, or virtually any type of wireless device.

[0063] UE 106 may include a processor configured to execute program instructions stored in memory. UE 106 may execute any method implementation of the method embodiments described herein by executing such stored instructions. Alternatively or additionally, UE 106 may include programmable hardware elements, such as a field-programmable gate array (FPGA) configured to execute any method implementation of the method embodiments described herein or any portion thereof.

[0064] UE 106 may include one or more antennas for communicating using one or more wireless communication protocols or technologies. In some embodiments, UE 106 may be configured to communicate using, for example, CDMA2000 (1xRTT / 1xEV-DO / HRPD / eHRPD), LTE / Advanced LTE, or 5G NR and / or GSM, LTE, Advanced LTE, or 5G NR using a single shared radio component. The shared radio component may be coupled to a single antenna or to multiple antennas (e.g., for MIMO) for performing wireless communication. Generally, the radio component may include any combination of baseband processor, analog RF signal processing circuitry (e.g., including filters, mixers, oscillators, amplifiers, etc.) or digital processing circuitry (e.g., for digital modulation and other digital processing). Similarly, the radio component may use the aforementioned hardware to implement one or more receive chains and transmit chains. For example, UE 106 may share one or more portions of the receive chain and / or transmit chain among multiple wireless communication technologies (such as those discussed above).

[0065] In some implementations, UE 106 may include independent transmit and / or receive chains (e.g., including independent antennas and other radio components) for each wireless communication protocol configured to communicate therewith. As another possibility, UE 106 may include one or more radio components shared among multiple wireless communication protocols, as well as one or more radio components used uniquely by a single wireless communication protocol. For example, UE 106 may include shared radio components for communication using either LTE (E-UTRAN) or 5G NR (or LTE or 1xRTT, or LTE or GSM), and independent radio components for communication using each of Wi-Fi and Bluetooth. Other configurations are also possible.

[0066] Figure 2 Block diagram of a base station Figure 2 Example block diagrams of base station 102 according to some implementation schemes are shown. It should be noted that... Figure 3 The base station shown is merely one example of a possible base station. As illustrated, base station 102 may include processor 204, which executes program instructions for base station 102. Processor 204 may also be coupled to memory management unit (MMU) 240, which may be configured to receive addresses from processor 204 and translate those addresses into locations in memory (e.g., memory 260 and read-only memory (ROM) 250), or to other circuitry or devices.

[0067] Base station 102 may include at least one network port 270. Network port 270 may be configured to couple to a telephone network and provide access to multiple devices, such as UE device 106, as described above in Figure 1 and... Figure 2 Access to the telephone network described in the text.

[0068] Network port 270 (or an additional network port) may also be configured, or alternatively, to be coupled to a cellular network, such as the core network of a cellular service provider. The core network may provide mobility-related services and / or other services to multiple devices, such as UE device 106. In some cases, network port 270 may be coupled to a telephone network via the core network, and / or the core network may provide a telephone network (e.g., between other UE devices served by a cellular service provider).

[0069] In some implementations, base station 102 may be a next-generation base station, such as a 5G New Radio (5G NR) base station, or a “gNB”. In such implementations, base station 102 may be connected to a legacy evolved packet core (EPC) network and / or to an NR core (NRC) network. Furthermore, base station 102 may be considered a 5G NR cell and may include one or more transition and receive points (TRPs). Additionally, UEs capable of operating according to 5G NR may be connected to one or more TRPs within one or more gNBs.

[0070] Base station 102 may include at least one antenna 234, and may include multiple antennas. At least one antenna 234 may be configured to operate as a wireless transceiver and may also be configured to communicate with UE device 106 via radio component 230. Antenna 234 communicates with radio component 230 via communication link 232. Communication link 232 may be a receive link, a transmit link, or both. Radio component 230 may be configured to communicate via various wireless communication standards, including but not limited to 5G NR, LTE, LTE-A, GSM, UMTS, CDMA2000, Wi-Fi, etc.

[0071] Base station 102 can be configured to perform wireless communication using multiple wireless communication standards. In some instances, base station 102 may include multiple radio components that enable base station 102 to communicate according to multiple wireless communication technologies. For example, as one possibility, base station 102 may include an LTE radio component for performing communication according to LTE and a 5G NR radio component for performing communication according to 5G NR. In this case, base station 102 may be able to operate as both an LTE base station and a 5G NR base station. As another possibility, base station 102 may include a multimode radio component capable of performing communication according to any of multiple wireless communication technologies (e.g., 5G NR and Wi-Fi, LTE and Wi-Fi, LTE and UMTS, LTE and CDMA2000, UMTS and GSM, etc.).

[0072] As further described herein, BS 102 may include hardware and software components for implementing or supporting specific implementations of the features described herein. The processor 204 of base station 102 may be configured, for example, to implement or support some or all of the methods described herein by executing program instructions stored on a memory medium (e.g., a non-transitory computer-readable memory medium). Alternatively, processor 204 may be configured as a programmable hardware element (such as an FPGA (Field-Programmable Gate Array)), or as an ASIC (Application-Specific Integrated Circuit), or a combination thereof. Alternatively (or further), in conjunction with one or more of other components 230, 232, 234, 240, 250, 260, 270, the processor 204 of BS 102 may be configured to implement or support some or all of the features described herein.

[0073] Furthermore, as described herein, processor 204 may comprise one or more processing elements. In other words, one or more processing elements may be included in processor 204. Therefore, processor 204 may include one or more integrated circuits (ICs) configured to perform the functions of processor 204. Additionally, each integrated circuit may include circuitry (e.g., a first circuit, a second circuit, etc.) configured to perform the functions of processor 204.

[0074] Furthermore, as described herein, radio component 230 may comprise one or more processing elements. In other words, radio component 230 may include one or more processing elements. Therefore, radio component 230 may include one or more integrated circuits (ICs) configured to perform the functions of radio component 230. Additionally, each integrated circuit may include circuitry (e.g., a first circuit, a second circuit, etc.) configured to perform the functions of radio component 230.

[0075] Figure 3 Server block diagram Figure 3 Example block diagrams of server 104 according to some implementation schemes are shown. Note that... Figure 3 The server shown is merely one example of a possible server. As illustrated, server 104 may include processor 344 capable of executing program instructions for server 104. Processor 344 may also be coupled to memory management unit (MMU) 374, which may be configured to receive addresses from processor 344 and translate those addresses into locations in memory (e.g., memory 364 and read-only memory (ROM) 354) or into other circuitry or devices.

[0076] Server 104 can be configured to provide access to network functions to multiple devices, such as base station 102 and UE device 106, for example, as further described herein.

[0077] In some implementations, server 104 may be part of a radio access network, such as a 5G New Radio (5G NR) access network. In some implementations, server 104 may be connected to a legacy evolved packet core (EPC) network and / or to an NR core (NRC) network.

[0078] As further described herein, server 104 may include hardware and software components for implementing or supporting the implementation of the features described herein. Processor 344 of server 104 may be configured, for example, to implement or support some or all of the methods described herein by executing program instructions stored on a memory medium (e.g., a non-transitory computer-readable storage medium). Alternatively, processor 344 may be configured as a programmable hardware element (such as an FPGA (Field-Programmable Gate Array)), or as an ASIC (Application-Specific Integrated Circuit), or a combination thereof. Alternatively (or in addition), in combination with one or more of other components 354, 364, and / or 374, processor 344 of server 104 may be configured to implement or support some or all of the features described herein.

[0079] Furthermore, as described herein, processor 344 may comprise one or more processing elements. In other words, one or more processing elements may be included in processor 344. Therefore, processor 344 may include one or more integrated circuits (ICs) configured to perform the functions of processor 344. Additionally, each integrated circuit may include circuitry (e.g., a first circuit, a second circuit, etc.) configured to perform the functions of processor 344.

[0080] Figure 4 : UE block diagram Figure 4A simplified block diagram of a communication device 106 according to some implementation schemes is shown. Note that... Figure 4 The block diagram of the communication device is merely one example of possible communication devices. According to the implementation, the communication device 106 may be a user equipment (UE) device, a mobile device or mobile station, a wireless device or wireless station, a desktop computer or computing device, a mobile computing device (e.g., a laptop computer, notebook computer, or portable computing device), a tablet computer, an unmanned aerial vehicle (UAV), a UAV controller (UAC), and / or a combination of devices, as well as other devices. As shown, the communication device 106 may include a set of components 400 configured to perform core functions. For example, this set of components may be implemented as a system-on-a-chip (SOC), which may include portions for various purposes. Alternatively, the set of components 400 may be implemented as separate components or groups of components for various purposes. The set of components 400 may be (e.g., communicatively; directly or indirectly) coupled to various other circuitry of the communication device 106.

[0081] For example, communication device 106 may include various types of memory (e.g., including NAND flash memory 410), input / output interfaces (such as connector I / F 420 (e.g., for connection to a computer system; docking station; charging station; input devices such as microphone, camera, keyboard; output devices such as speaker; etc.)), a display 460 that can be integrated with or external to the communication device 106, cellular communication circuitry 430 such as for 5G NR, LTE, GSM, etc., and short- to medium-range wireless communication circuitry 429 (e.g., Bluetooth). ™ (and WLAN circuitry). In some embodiments, communication device 106 may include wired communication circuitry (not shown), such as, for example, a network interface card for Ethernet.

[0082] Cellular communication circuitry 430 may be coupled (e.g., communicatively grounded; directly or indirectly) to one or more antennas, such as antennas 435 and 436 shown. Short-to-medium-range wireless communication circuitry 429 may also be coupled (e.g., communicatively grounded; directly or indirectly) to one or more antennas, such as antennas 437 and 438 shown. Alternatively, short-to-medium-range wireless communication circuitry 429 may be coupled (e.g., communicatively grounded; directly or indirectly) to antennas 435 and 436, and may also be coupled to antennas 437 and 438, or alternatively (e.g., communicatively grounded; directly or indirectly). Short-to-medium-range wireless communication circuitry 429 and / or cellular communication circuitry 430 may include multiple receive chains and / or multiple transmit chains for receiving and / or transmitting multiple spatial streams, such as in a multiple-input multiple-output (MIMO) configuration.

[0083] In some embodiments, as further described below, the cellular communication circuit 430 may include dedicated receive chains (including and / or (e.g., communicatively; directly or indirectly) coupled to dedicated processors and / or radio components) for multiple RATs (e.g., a first receive chain for LTE and a second receive chain for 5G NR). Furthermore, in some embodiments, the cellular communication circuit 430 may include a single transmit chain that can be switched between radio components dedicated to a particular RAT. For example, a first radio component may be dedicated to a first RAT, such as LTE, and may communicate with a dedicated receive chain and a transmit chain shared with additional radio components, such as a second radio component that may be dedicated to a second RAT (e.g., 5G NR) and may communicate with a dedicated receive chain and a shared transmit chain.

[0084] The communication device 106 may also include one or more user interface elements and / or be configured to be used with one or more user interface elements. The user interface elements may include any of a variety of elements, such as a display 460 (which may be a touch screen display), a keyboard (which may be a separate keyboard or may be implemented as part of a touch screen display), a mouse, a microphone and / or a speaker, one or more cameras, one or more buttons, and / or any of a variety of other elements capable of providing information to the user and / or receiving or interpreting user input.

[0085] The communication device 106 may also include one or more smart cards 445 with SIM (Subscriber Identity Module) functionality, such as one or more UICC (Universal Integrated Circuit Card) 445. It should be noted that the term "SIM" or "SIM entity" is intended to include any of various types of SIM implementations or SIM functions, such as one or more UICC cards 445, one or more eUICCs, one or more eSIMs, removable or embedded, etc. In some embodiments, the UE 106 may include at least two SIMs. Each SIM may execute one or more SIM applications and / or otherwise implement SIM functionality. Thus, each SIM may be a single smart card that can be embedded, for example, soldered to a circuit board in the UE 106, or each SIM 410 may be implemented as a removable smart card. Therefore, a SIM may be one or more removable smart cards (such as UICC cards, sometimes referred to as "SIM cards"), and / or SIM 410 may be one or more embedded cards (such as embedded UICCs (eUICCs), sometimes referred to as "eSIMs" or "eSIM cards"). In some implementations (such as when the SIM includes an eUICC), one or more SIMs within the SIM can implement embedded SIM (eSIM) functionality; in such implementations, a single SIM within the SIM can execute multiple SIM applications. Each SIM may include components such as a processor and / or memory; instructions for performing SIM / eSIM functionality may be stored in memory and executed by the processor. In some implementations, UE 106 may include, as needed, a combination of removable smart cards and fixed / non-removable smart cards (such as one or more eUICC cards implementing eSIM functionality). For example, UE 106 may include two embedded SIMs, two removable SIMs, or a combination of one embedded SIM and one removable SIM. Various other SIM configurations are also envisioned.

[0086] As described above, in some implementations, UE 106 may include two or more SIMs. Including two or more SIMs in UE 106 allows UE 106 to support two different phone numbers and allows UE 106 to communicate on two or more corresponding networks. For example, the first SIM may support a first RAT such as LTE, and the second SIM 106 may support a second RAT such as 5G NR. Other specific implementations and RATs are also possible. In some implementations, when UE 106 includes two SIMs, UE 106 may support Dual SIM Dual Standby (DSDA) functionality. DSDA functionality allows UE 106 to connect to two networks simultaneously (and use two different RATs), or allows two connections supported by two different SIMs using the same or different RATs to be maintained simultaneously on the same or different networks. DSDA functionality also allows UE 106 to receive voice calls or data traffic simultaneously on either phone number. In some implementations, voice calls may be packet-switched communications. In other words, voice calls can be received using LTE-based Voice (VoLTE) technology and / or NR-based Voice (VoNR) technology. In some implementations, UE 106 may support Dual SIM Dual Standby (DSDS) functionality. DSDS functionality allows either of the two SIMs in UE 106 to remain in standby while awaiting a voice call and / or data connection. In DSDS, when a call / data connection is established on one SIM, the other SIM is no longer active. In some implementations, DSDx functionality (DSDA or DSDS functionality) can be implemented using a single SIM (e.g., eUICC) that performs multiple SIM applications for different carriers and / or RATs.

[0087] As shown in the figure, the SOC 400 may include a processor 402 and display circuitry 404. The processor executes program instructions for the communication device 106, and the display circuitry performs graphics processing and provides display signals to the display 460. The processor 402 may also be coupled to a memory management unit (MMU) 440, which may be configured to receive addresses from the processor 402 and translate those addresses into locations in memory (e.g., memory 406, read-only memory (ROM) 450, NAND flash memory 410), and / or coupled to other circuitry or devices (such as display circuitry 404, short-to-medium range wireless communication circuitry 429, cellular communication circuitry 430, connector I / F 420, and / or display 460). The MMU 440 may be configured to perform memory protection and page table translation or setup. In some embodiments, the MMU 440 may be included as part of the processor 402.

[0088] As noted above, communication device 106 may be configured to communicate using wireless and / or wired communication circuits. Communication device 106 may be configured to perform methods for AI-based CSI feedback with CSI prediction, including systems, methods, and mechanisms for indicative prediction of CSI reports for UEs, such as in 5G NR systems and outside, network configuration for CSI feedback, UE PMI report format, and AI model lifecycle management, as further described herein.

[0089] As described herein, communication device 106 may include hardware and software components for implementing the features described above to communicate a scheduling profile for power saving to the network. For example, by executing program instructions stored on a memory medium (e.g., a non-transitory computer-readable storage medium), processor 402 of communication device 106 may be configured to implement some or all of the features described herein. Alternatively (or further), processor 402 may be configured as a programmable hardware element (such as a FPGA (Field-Programmable Gate Array)) or as an ASIC (Application-Specific Integrated Circuit). Alternatively (or further), in conjunction with one or more of other components 400, 404, 406, 410, 420, 429, 430, 440, 445, 450, 460, processor 402 of communication device 106 may be configured to implement some or all of the features described herein.

[0090] Furthermore, as described herein, processor 402 may include one or more processing elements. Therefore, processor 402 may include one or more integrated circuits (ICs) configured to perform the functions of processor 402. Additionally, each integrated circuit may include circuitry (e.g., a first circuit, a second circuit, etc.) configured to perform the functions of processor 402.

[0091] Furthermore, as described herein, the cellular communication circuit 430 and the short-to-medium-range wireless communication circuit 429 may each include one or more processing elements. In other words, one or more processing elements may be included in the cellular communication circuit 430, and similarly, one or more processing elements may be included in the short-to-medium-range wireless communication circuit 429. Therefore, the cellular communication circuit 430 may include one or more integrated circuits (ICs) configured to perform the functions of the cellular communication circuit 430. Furthermore, each integrated circuit may include circuitry (e.g., a first circuit, a second circuit, etc.) configured to perform the functions of the cellular communication circuit 430. Similarly, the short-to-medium-range wireless communication circuit 429 may include one or more ICs configured to perform the functions of the short-to-medium-range wireless communication circuit 429. Furthermore, each integrated circuit may include circuitry (e.g., a first circuit, a second circuit, etc.) configured to perform the functions of the short-to-medium-range wireless communication circuit 429.

[0092] Figure 5 Block diagram of cellular communication circuit Figure 5 Simplified block diagrams of cellular communication circuits according to some implementation schemes are shown. Note that... Figure 5 The block diagram of the cellular communication circuit is merely one example of a possible cellular communication circuit. According to the implementation, the cellular communication circuit 530 (which may be the cellular communication circuit 430) may be included in a communication device such as the communication device 106 described above. As noted above, among other devices, the communication device 106 may be a user equipment (UE) device, a mobile device or mobile station, a wireless device or wireless station, a desktop computer or computing device, a mobile computing device (e.g., a laptop computer, notebook computer, or portable computing device), a tablet computer, and / or a combination of these devices.

[0093] Cellular communication circuit 530 may (e.g., communicatively; directly or indirectly) be coupled to one or more antennas, such as ( Figure 4 Antennas 435a to 435b and 436 are shown in the diagram. In some embodiments, the cellular communication circuitry 530 may include dedicated receive chains for various RATs (including and / or (e.g., communicatively grounded; directly or indirectly) coupled to dedicated processors and / or radio components) (e.g., a first receive chain for LTE and a second receive chain for 5G NR). For example, as... Figure 5 As shown, the cellular communication circuit 530 may include a modem 510 and a modem 520. The modem 510 may be configured for communication according to a first RAT (e.g., such as LTE or LTE-A), and the modem 520 may be configured for communication according to a second RAT (e.g., such as 5G NR).

[0094] As shown, modem 510 may include one or more processors 512 and memory 516 communicating with processors 512. Modem 510 may communicate with radio frequency (RF) front end 530. RF front end 530 may include circuitry for transmitting and receiving radio signals. For example, RF front end 530 may include receiver circuitry (RX) 532 and transmitter circuitry (TX) 534. In some embodiments, receiver circuitry 532 may communicate with downlink (DL) front end 550, which may include circuitry for receiving radio signals via antenna 335a.

[0095] Similarly, modem 520 may include one or more processors 522 and memory 526 communicating with processor 522. Modem 520 may communicate with RF front end 540. RF front end 540 may include circuitry for transmitting and receiving radio signals. For example, RF front end 540 may include receiving circuitry 542 and transmitting circuitry 544. In some embodiments, receiving circuitry 542 may communicate with DL front end 560, which may include circuitry for receiving radio signals via antenna 335b.

[0096] In some implementations, switch 570 may couple transmitting circuitry 534 to uplink (UL) front-end 572. Additionally, switch 570 may couple transmitting circuitry 544 to UL front-end 572. UL front-end 572 may include circuitry for transmitting radio signals via antenna 336. Therefore, when cellular communication circuitry 530 receives an instruction to transmit according to a first RAT (e.g., supported by modem 510), switch 570 may be switched to a first state allowing modem 510 to transmit signals according to the first RAT (e.g., via a transmission chain including transmitting circuitry 534 and UL front-end 572). Similarly, when cellular communication circuitry 530 receives an instruction to transmit according to a second RAT (e.g., supported by modem 520), switch 570 may be switched to a second state allowing modem 520 to transmit signals according to the second RAT (e.g., via a transmission chain including transmitting circuitry 544 and UL front-end 572).

[0097] In some implementations, cellular communication circuit 530 may be configured to perform methods for AI-based CSI feedback with CSI prediction, including systems, methods, and mechanisms for UE to indicate predicted CSI reports, network configuration for CSI feedback, UE PMI report format, and AI model lifecycle management, as further described herein.

[0098] As described herein, modem 510 may include hardware and software components for implementing the features described above or for UL data used in time-division multiplexing NSA NR operation, as well as various other techniques described herein. For example, processor 512 may be configured to implement some or all of the features described herein by executing program instructions stored on a memory medium (e.g., a non-transitory computer-readable storage medium). Alternatively (or additionally), processor 512 may be configured as a programmable hardware element (such as a FPGA (Field-Programmable Gate Array)) or as an ASIC (Application-Specific Integrated Circuit). Alternatively (or additionally), processor 512 may be configured to implement some or all of the features described herein by combining one or more of other components 530, 532, 534, 550, 570, 572, 335, and 336.

[0099] Furthermore, as described herein, processor 512 may include one or more processing elements. Therefore, processor 512 may include one or more integrated circuits (ICs) configured to perform the functions of processor 512. Additionally, each integrated circuit may include circuitry (e.g., a first circuit, a second circuit, etc.) configured to perform the functions of processor 512.

[0100] As described herein, modem 520 may include hardware and software components for implementing the aforementioned features for AI-based CSI feedback with CSI prediction, including systems, methods, and mechanisms for indicative prediction of CSI reports by the UE, network configuration for CSI feedback, UE PMI report format, and AI model lifecycle management, as well as various other techniques described herein. For example, processor 522 may be configured to implement some or all of the features described herein by executing program instructions stored on a memory medium (e.g., a non-transitory computer-readable storage medium). Alternatively (or in addition), processor 522 may be configured as a programmable hardware element such as a FPGA (Field-Programmable Gate Array) or as an ASIC (Application-Specific Integrated Circuit). Alternatively (or in addition), processor 522 may be configured to implement some or all of the features described herein by combining one or more of other components 540, 542, 544, 550, 570, 572, 335, and 336.

[0101] Furthermore, as described herein, processor 522 may include one or more processing elements. Therefore, processor 522 may include one or more integrated circuits (ICs) configured to perform the functions of processor 522. Additionally, each integrated circuit may include circuitry (e.g., a first circuit, a second circuit, etc.) configured to perform the functions of processor 522.

[0102] Figure 6A , Figure 6B and Figure 7 5G Core Network Architecture – Interoperability with Wi-Fi In some implementations, access to the 5G core network (CN) can be made via (or through) cellular connections / interfaces (e.g., via 3GPP communication architectures / protocols) and non-cellular connections / interfaces (e.g., non-3GPP access architectures / protocols such as Wi-Fi connections). Figure 6A An example of a 5G network architecture according to some implementation schemes is illustrated, which combines both 3GPP (e.g., cellular) and non-3GPP (e.g., non-cellular) access to the 5G CN. As shown, a user equipment device (e.g., such as UE 106) can access the 5G CN through both a radio access network (RAN, such as gNB 604, which can be base station 102) and an access point (such as AP 612). AP 612 may include a connection to the Internet 600 and a connection to a non-3GPP interoperability function (N3IWF) 603 network entity. N3IWF may include a connection to the core access and mobility management function (AMF) 605 of the 5G CN. AMF 605 may include an instance of 5G mobility management (5G MM) function associated with UE 106. In addition, the RAN (e.g., gNB 604) may also have a connection to AMF 605. Therefore, the 5G CN can support unified authentication on both connections and allow simultaneous registration for UE 106 access via both gNB 604 and AP 612. As shown in the figure, AMF 605 may include one or more functional entities associated with the 5G CN (e.g., Network Slice Selection Function (NSSF) 620, Short Message Service Function (SMSF) 622, Application Function (AF) 624, Unified Data Management (UDM) 626, Policy Control Function (PCF) 628, and / or Authentication Server Function (AUSF) 630). It should be noted that these functional entities can also be supported through the 5G CN's Session Management Functions (SMF) 606a and SMF 606b. AMF 605 can connect to (or communicate with) SMF 606a. Additionally, gNB 604 can communicate with (or connect to) User Plane Function (UPF) 608a, which can also communicate with SMF 606a. Similarly, N3IWF 603 can communicate with UPF 608b, which in turn can communicate with SMF 606b. Both UPFs can communicate with data networks (e.g., DN 610a and 610b) and / or the Internet 600 and the Internet Protocol (IP) Multimedia Subsystem / IP Multimedia Core Network Subsystem (IMS) Core Network 610.

[0103] Figure 6BAn example of a 5G network architecture according to some implementation schemes is illustrated, which combines both dual 3GPP (e.g., LTE and 5G NR) and non-3GPP access to the 5GCN. As shown, a user equipment device (e.g., such as UE 106) can access the 5G CN via both a radio access network (RAN, such as gNB 604 or eNB 602, which can be base station 102) and an access point (such as AP 612). AP 612 may include a connection to the Internet 600 and a connection to the N3IWF 603 network entity. N3IWF may include a connection to the AMF 605 of the 5G CN. AMF 605 may include an instance of 5G MM functionality associated with UE 106. In addition, the RAN (e.g., gNB 604) may also have a connection to AMF 605. Therefore, the 5G CN can support unified authentication on both connections and allow simultaneous registration of UE 106 accessing via both gNB 604 and AP 612. Additionally, the 5G CN can support dual registration of the UE on both a legacy network (e.g., LTE via eNB 602) and a 5G network (e.g., via gNB 604). As shown in the figure, eNB 602 can have connections to both Mobility Management Entity (MME) 642 and Service Gateway (SGW) 644. MME 642 can have connections to both SGW 644 and AMF 605. Furthermore, SGW 644 can have connections to both SMF 606a and UPF 608a. As shown in the figure, AMF 605 can include one or more functional entities associated with the 5G CN (e.g., NSSF 620, SMSF 622, AF 624, UDM 626, PCF 628, and / or AUSF 630). Note that UDM 626 can also include Home Subscriber Server (HSS) functionality, and PCF can also include Policy and Charging Rules (PCRF) functionality. It should also be noted that these functional entities can also be supported by the 5G CN's SMF 606a and SMF 606b. The AMF605 can connect to (or communicate with) the SMF 606a. Furthermore, the gNB 604 can communicate with (or connect to) the UPF 608a, which in turn can communicate with the SMF 606a. Similarly, the N3IWF 603 can communicate with the UPF 608b, which can also communicate with the SMF606b. Both UPFs can communicate with data networks (e.g., DN 610a and 610b) and / or the Internet 600 and the IMS core network 610.

[0104] It should be noted that, in various implementations, one or more of the network entities described above may be configured to perform methods for AI-based CSI feedback with CSI prediction, including systems, methods, and mechanisms for UEs to indicate predicted CSI reports, network configurations for CSI feedback, UE PMI report formats, and AI model lifecycle management, such as those further described herein.

[0105] Figure 7 Examples of baseband processor architectures for UEs (e.g., such as UE 106) according to some implementation schemes are illustrated. Figure 7 The baseband processor architecture 700 described herein can be implemented on one or more radio components (e.g., radio components 429 and / or 430) or modems (e.g., modems 510 and / or 520) as described above. As shown, the non-access stratum (NAS) 710 may include a 5G NAS 720 and a traditional NAS 750. The traditional NAS 750 may include a communication connection with a traditional access stratum (AS) 770. The 5G NAS 720 may include communication connections with a 5G AS 740, a non-3GPP AS 730, and a Wi-Fi AS 732. The 5G NAS 720 may include functional entities associated with both access strata. Therefore, the 5G NAS 720 may include multiple 5G MM entities 726 and 728 and 5G session management (SM) entities 722 and 724. The traditional NAS 750 may include functional entities such as Short Message Service (SMS) entity 752, Evolved Packet System (EPS) Session Management (ESM) entity 754, Session Management (SM) entity 756, EPS Mobility Management (EMM) entity 758, and Mobility Management (MM) / GPRS Mobility Management (GMM) entity 760. Additionally, the traditional AS 770 may include functional entities such as LTE AS 772, UMTS AS 774, and / or GSM / GPRS AS 776.

[0106] Therefore, the baseband processor architecture 700 allows for a common 5G-NAS for both 5G cellular and non-cellular (e.g., non-3GPP access) networks. The baseband processor architecture 700 can communicate with one or more UICC 745s. Note that, as shown in the figure, the 5GMM can maintain separate connection management and registration management state machines for each connection. Additionally, a device (e.g., UE 106) can register to a single PLMN (e.g., a 5G CN) using both 5G cellular and non-cellular access. Furthermore, a device can be in a connected state in one access and an idle state in another, or vice versa. Finally, for both accesses, there may be common 5G-MM procedures (e.g., registration, deregistration, identification, authentication, etc.).

[0107] It should be noted that, in various implementations, one or more of the aforementioned functional entities of the 5G NAS and / or 5G AS may be configured to perform methods for AI-based CSI feedback with CSI prediction, including systems, methods, and mechanisms for UE to indicate predicted CSI reports, network configuration for CSI feedback, UE PMI report format, and AI model lifecycle management, such as those further described herein.

[0108] Figure 8 Example components of device 800 according to some embodiments are illustrated. In some embodiments, device 800 may include at least application circuitry 802, baseband circuitry 804, radio frequency (RF) circuitry 806, front-end module (FEM) circuitry 808, one or more antennas 810, and power management circuitry (PMC) 812 coupled together as shown. Components of the illustrated device 800 may be included in a UE or RAN node. In some embodiments, device 800 may include fewer components (e.g., the RAN node may not utilize application circuitry 802, but instead include a processor / controller to process IP data received from the EPC). In some embodiments, device 800 may include additional components such as, for example, memory / storage devices, displays, cameras, sensors, or input / output (I / O) interfaces. In other embodiments, the components described below may be included in more than one device (e.g., the circuitry may be individually included in more than one device for a cloud RAN (C-RAN) specific implementation).

[0109] Application circuitry 802 may include one or more application processors. For example, application circuitry 802 may include circuitry such as, but not limited to, one or more single-core or multi-core processors. The one or more processors may include any combination of general-purpose processors and special-purpose processors (e.g., graphics processors, application processors, etc.). These processors may be coupled to or may include memory / storage devices and may be configured to execute instructions stored in the memory / storage device to enable various applications or operating systems to run on device 800. In some embodiments, the processor of application circuitry 802 may process IP data packets received from the EPC.

[0110] Baseband circuitry 804 may include circuitry such as, but not limited to, one or more single-core or multi-core processors. Baseband circuitry 804 may include one or more baseband processors or control logic components to process baseband signals received from the receive signal path of RF circuitry 806 and generate baseband signals for the transmit signal path of RF circuitry 806. Baseband processing circuitry 804 may interact with application circuitry 802 to generate and process baseband signals and control the operation of RF circuitry 806. For example, in some embodiments, baseband circuitry 804 may include a third-generation (3G) baseband processor 804A, a fourth-generation (4G) baseband processor 804B, a fifth-generation (5G) baseband processor 804C, or one or more other existing, under development, or future generations of baseband processors 804D (e.g., second-generation (2G), sixth-generation (6G), etc.). Baseband circuitry 804 (e.g., one or more of baseband processors 804A to 804D) may handle various radio control functions to implement communication with one or more radio networks via RF circuitry 806. In other embodiments, some or all of the functions of the baseband processors 804A to 804D may be included in modules stored in memory 804G and executed via a central processing unit (CPU) 804E. Radio control functions may include, but are not limited to, signal modulation / demodulation, encoding / decoding, and radio frequency shifting. In some embodiments, the modulation / demodulation circuitry of the baseband circuitry 804 may include Fast Fourier Transform (FFT), pre-decoding, or constellation mapping / demapping functions. In some embodiments, the encoding / decoding circuitry of the baseband circuitry 804 may include convolution, tail-biting convolution, turbo, Viterbi, or low-density parity-check (LDPC) encoder / decoder functions. Implementations of the modulation / demodulation and encoder / decoder functions are not limited to these examples, and other suitable functions may be included in other embodiments.

[0111] In some embodiments, the baseband circuitry 804 may include one or more audio digital signal processors (“DSPs”) 804F. The audio DSP 804F may include elements for compression / decompression and echo cancellation, and in other embodiments may include other suitable processing elements. In some embodiments, components of the baseband circuitry may be suitably combined in a single chip, a single chipset, or disposed on the same circuit board. In some embodiments, some or all of the components of the baseband circuitry 804 and the application circuitry 802 may be implemented together, for example, on a system-on-a-chip (SOC).

[0112] In some implementations, baseband circuit 804 can provide communication compatible with one or more radio technologies. For example, in some implementations, baseband circuit 804 can support communication with the Evolved Universal Terrestrial Radio Access Network (EUTRAN) or other Wireless Metropolitan Area Networks (WMAN), Wireless Local Area Networks (WLAN), or Wireless Personal Area Networks (WPAN). Implementations in which baseband circuit 804 is configured to support radio communication with more than one radio protocol may be referred to as multimode baseband circuits.

[0113] RF circuit 806 enables communication with a wireless network via a non-solid medium using modulated electromagnetic radiation. In various embodiments, RF circuit 806 may include switches, filters, amplifiers, etc., to facilitate communication with the wireless network. RF circuit 806 may include a receive signal path, which may include circuitry for down-converting the RF signal received from FEM circuit 808 and providing a baseband signal to baseband circuit 804. RF circuit 806 may also include a transmit signal path, which may include circuitry for up-converting the baseband signal provided by baseband circuit 804 and providing an RF output signal to FEM circuit 808 for transmission.

[0114] In some embodiments, the receive signal path of RF circuit 806 may include mixer circuit 806a, amplifier circuit 806b, and filter circuit 806c. In some embodiments, the transmit signal path of RF circuit 806 may include filter circuit 806c and mixer circuit 806a. RF circuit 806 may also include synthesizer circuit 806d for synthesizing frequencies used by mixer circuit 806a in both the receive and transmit signal paths. In some embodiments, mixer circuit 806a in the receive signal path may be configured to down-convert the RF signal received from FEM circuit 808 based on the synthesized frequency provided by synthesizer circuit 806d. Amplifier circuit 806b may be configured to amplify the down-converted signal, and filter circuit 806c may be a low-pass filter (LPF) or band-pass filter (BPF) configured to remove unwanted signals from the down-converted signal to generate an output baseband signal. The output baseband signal may be provided to baseband circuit 804 for further processing. In some implementations, although not required, the output baseband signal may be a zero-frequency baseband signal. In some implementations, the mixer circuit 806a of the receiving signal path may include a passive mixer, but the scope of the implementations is not limited in this respect.

[0115] In some implementations, the mixer circuit 806a of the transmit signal path can be configured to up-convert the input baseband signal based on the synthesis frequency provided by the synthesizer circuit 806d to generate the RF output signal of the FEM circuit 808. The baseband signal can be provided by the baseband circuit 804 and can be filtered by the filter circuit 806c.

[0116] In some embodiments, the mixer circuit 806a for the receive signal path and the mixer circuit 806a for the transmit signal path may include two or more mixers and may be arranged for quadrature downconversion and quadrature upconversion, respectively. In some embodiments, the mixer circuit 806a for the receive signal path and the mixer circuit 806a for the transmit signal path may include two or more mixers and may be arranged for image rejection (e.g., Hartley image rejection). In some embodiments, the mixer circuit 806a for the receive signal path and the mixer circuit 806a for the transmit signal path may be arranged for direct downconversion and direct upconversion, respectively. In some embodiments, the mixer circuit 806a for the receive signal path and the mixer circuit 806a for the transmit signal path may be configured for superheterodyne operation.

[0117] In some embodiments, the output baseband signal and the input baseband signal may be analog baseband signals, but the scope of the embodiments is not limited in this respect. In some alternative embodiments, the output baseband signal and the input baseband signal may be digital baseband signals. In these alternative embodiments, RF circuit 806 may include analog-to-digital converter (ADC) and digital-to-analog converter (DAC) circuitry, and baseband circuit 804 may include a digital baseband interface for communicating with RF circuit 806.

[0118] In some dual-mode implementations, separate radio IC circuits may be provided to process signals for each spectrum, but the scope of the implementation is not limited in this respect.

[0119] In some implementations, synthesizer circuit 806d may be a fractional-N synthesizer or a fractional-N / N+1 synthesizer, but the scope of implementations is not limited in this respect, as other types of frequency synthesizers may also be suitable. For example, synthesizer circuit 806d may be a Δ-∑ synthesizer, a frequency multiplier, or a synthesizer including a phase-locked loop with a frequency divider.

[0120] Synthesizer circuit 806d can be configured to synthesize an output frequency based on the frequency input and the divider control input for use by mixer circuit 806a of RF circuit 806. In some embodiments, synthesizer circuit 806d can be a fractional N / N+1 synthesizer.

[0121] In some implementations, the frequency input may be provided by a voltage-controlled oscillator (VCO), although this is not mandatory. The divider control input may be provided by the baseband circuitry 804 or the application processor 802 according to the desired output frequency. In some implementations, the divider control input (e.g., N) may be determined from a lookup table based on the channel indicated by the application processor 802.

[0122] The synthesizer circuit 806d of the RF circuit 806 may include a frequency divider, a delay-locked loop (DLL), a multiplexer, and a phase accumulator. In some embodiments, the frequency divider may be a dual-mode divider (DMD), and the phase accumulator may be a digital phase accumulator (DPA). In some embodiments, the DMD may be configured to divide the input signal by N or N+1 (e.g., based on carry) to provide a fractional division ratio. In some example embodiments, the DLL may include a cascaded, tunable delay element, a phase detector, a charge pump, and a set of D-type flip-flops. In these embodiments, the delay elements may be configured to divide the VCO cycle into Nd equal phase groups, where Nd is the number of delay elements in the delay line. Thus, the DLL provides negative feedback to help ensure that the total delay through the delay line is one VCO cycle.

[0123] In some embodiments, synthesizer circuitry 806d may be configured to generate a carrier frequency as the output frequency, while in other embodiments, the output frequency may be a multiple of the carrier frequency (e.g., twice the carrier frequency, four times the carrier frequency) and used in conjunction with quadrature generator and frequency divider circuitry to generate multiple signals having multiple different phases relative to each other at the carrier frequency. In some embodiments, the output frequency may be the LO frequency (fLO). In some embodiments, RF circuitry 806 may include an IQ / polarity converter.

[0124] FEM circuit 808 may include a receive signal path, which may include circuitry configured to operate on RF signals received from one or more antennas 810, amplify the received signals, and provide an amplified version of the received signals to RF circuit 806 for further processing. FEM circuit 808 may also include a transmit signal path, which may include circuitry configured to amplify transmit signals provided by RF circuit 806 for transmission by one or more of the one or more antennas 810. In various embodiments, amplification via the transmit signal path or the receive signal path may be performed only in RF circuit 806, only in FEM 808, or in both RF circuit 806 and FEM 808.

[0125] In some embodiments, FEM circuit 808 may include a TX / RX switch for switching between transmit and receive mode operation. The FEM circuit may include a receive signal path and a transmit signal path. The receive signal path of the FEM circuit may include an LNA for amplifying the received RF signal and providing the amplified received RF signal as an output (e.g., provided to RF circuit 806). The transmit signal path of FEM circuit 808 may include a power amplifier (PA) for amplifying (e.g., provided by RF circuit 806) the input RF signal; and one or more filters for generating RF signals for subsequent transmission (e.g., via one or more antennas in one or more antennas 810).

[0126] In some implementations, the PMC 812 can manage the power supplied to the baseband circuitry 804. Specifically, the PMC 812 can control power selection, voltage scaling, battery charging, or DC-DC conversion. The PMC 812 is typically included when the device 800 can be powered by a battery, for example, when the device is included in a UE. The PMC 812 can improve power conversion efficiency while providing the desired specific implementation size and thermal characteristics.

[0127] and Figure 8 The PMC 812 is shown coupled only to the baseband circuit 804. However, in other embodiments, the PMC 812 may be additionally or alternatively coupled to other components, such as, but not limited to, application circuit 802, RF circuit 806, or FEM 808, and perform similar power management operations.

[0128] In some implementations, the PMC 812 may be controlled or otherwise integrated into various power-saving mechanisms of the device 800. For example, if the device 800 is in the RRC_Connected state, where it remains connected to the RAN node as it expects to receive traffic immediately, then after a period of inactivity, the device may enter a state known as Discontinuous Receive Mode (DRX). During this state, the device 800 may be powered down for short intervals, thereby saving power.

[0129] If no data traffic activity occurs during the extended time period, device 800 may transition to the RRC_Idle state, in which the device disconnects from the network and performs no operations such as channel quality feedback or handover. Device 800 enters a very low-power state and performs paging, during which the device periodically wakes up again to listen to the network before powering down again. Device 800 cannot receive data in this state. To receive data, the device must transition back to the RRC_Connected state.

[0130] An additional power-saving mode renders the device unusable for a period exceeding the paging interval (from seconds to hours). During this time, the device is completely unconnected to the network and may be completely powered off. Any data transmitted during this period will incur significant latency, which is assumed to be acceptable.

[0131] The processor of application circuit 802 and the processor of baseband circuit 804 can be used to execute elements of one or more instances of the protocol stack. For example, the processor of baseband circuit 804 can be used individually or in combination to execute layer 3, layer 2, or layer 1 functionality, while the processor of application circuit 804 can utilize data received from these layers (e.g., packet data) and further execute layer 4 functionality (e.g., transmit communication protocol (TCP) and user datagram protocol (UDP) layers). As mentioned herein, layer 3 may include the radio resource control (RRC) layer, which will be described in further detail below. As mentioned herein, layer 2 may include the media access control (MAC) layer, the radio link control (RLC) layer, and the packet data convergence protocol (PDCP) layer, which will be described in further detail below. As mentioned herein, layer 1 may include the physical (PHY) layer of the UE / RAN node, which will be described in further detail below.

[0132] Figure 9 Example interfaces of baseband circuits according to some implementation schemes are illustrated. As discussed above, Figure 8 The baseband circuit 804 may include processors 804A-804E and a memory 804G utilized by the processors. Each of the processors 804A-804E may respectively include a memory interface 904A-904E for sending / receiving data to / from the memory 804G.

[0133] The baseband circuit 804 may also include one or more interfaces, such as a memory interface 912 (e.g., an interface for transferring / receiving data to / from a memory external to the baseband circuit 804), and an application circuit interface 914 (e.g., an interface for transferring / receiving data to / from a memory external to the baseband circuit 804). Figure 8 Application circuit 802 is an interface for transmitting / receiving data), and RF circuit interface 916 (e.g., for sending / receiving data to / from...). Figure 8 The RF circuit 806 is an interface for transmitting / receiving data, and the wireless hardware connection interface 918 is used for transmitting / receiving data to / from near field communication (NFC) components, Bluetooth, etc. ® Components (e.g., Bluetooth) ® Low power consumption, Wi-Fi ® The component and other communication components transmit / receive data interfaces) and power management interface 920 (e.g., an interface for transmitting / receiving power or control signals to / from PMC 812) are communicatively coupled to other circuits / devices.

[0134] Figure 10 This is an example of a control plane protocol stack according to some implementation schemes. In one implementation scheme, control plane 1000 may be a communication protocol stack between one or more UEs (such as, for example, UE 801 (or alternatively, UE 802) and / or one or more RAN nodes 811 (or alternatively, RAN node 812)) and mobility management entity (MME) 821.

[0135] PHY layer 1001 can transmit or receive information used by MAC layer 1002 through one or more air interfaces. PHY layer 1001 can also perform link adaptive or adaptive modulation and decoding (AMC), power control, cell search (e.g., for initial synchronization and handover purposes), and other measurements used by higher layers (such as RRC layer 1005). PHY layer 1001 can further perform: error detection of the transport channel, forward error correction (FEC) decoding / decoding of the transport channel, modulation / demodulation of the physical channel, interleaving, rate matching, mapping to the physical channel, and multiple-input multiple-output (MIMO) antenna processing.

[0136] MAC layer 1002 can perform the following: mapping between logical channels and transport channels, multiplexing MAC service data units (SDUs) from one or more logical channels onto a transport block (TB) to be delivered to the PHY via the transport channel, demultiplexing MAC SDUs from a transport block (TB) delivered from the PHY via the transport channel onto one or more logical channels, multiplexing MAC SDUs onto a TB, scheduling information reporting, error correction via Hybrid Automatic Repeat Request (HARQ), and prioritizing logical channels.

[0137] RLC layer 1003 can operate in multiple modes, including: Transparent Mode (TM), Unacknowledged Mode (UM), and Acknowledged Mode (AM). RLC layer 1003 can perform the transmission of upper-layer protocol data units (PDUs), error correction via Automatic Repeat Request (ARQ) for AM data transmission, and concatenation, segmentation, and reassembly of RLC SDUs for UM and AM data transmission. RLC layer 1003 can also re-segment RLC data PDUs used for AM data transmission, reorder RLC data PDUs used for UM and AM data transmission, detect duplicate data used for UM and AM data transmission, discard RLC SDUs used for UM and AM data transmission, detect protocol errors used for AM data transmission, and perform RLC re-establishment.

[0138] PDCP layer 1004 can perform header compression and decompression of IP data, maintain PDCP sequence numbers (SNs), perform sequential delivery of upper-layer PDUs during lower-layer re-establishment, eliminate duplication of lower-layer SDUs during lower-layer re-establishment for radio bearers mapped on RLC AM, encrypt and decrypt control plane data, perform integrity protection and integrity verification of control plane data, control timer-based data discarding, and perform security operations (e.g., encryption, decryption, integrity protection, integrity verification, etc.).

[0139] The main services and functions of RRC layer 1005 may include broadcasting system information (e.g., included in the Master Information Block (MIB) or System Information Block (SIB) related to the Non-Access Stratum (NAS), broadcasting system information related to the Access Stratum (AS), paging, establishment, maintenance, and release of RRC connections between the UE and E-UTRAN (e.g., RRC connection paging, RRC connection establishment, RRC connection modification, and RRC connection release), establishment, configuration, maintenance, and release of point-to-point radio bearers, security functions including key management, mobility between Radio Access Technologies (RATs), and measurement configuration for UE measurement reporting. The MIB and SIB may include one or more Information Elements (IEs), each of which may include a separate data field or data structure.

[0140] In one example, the UE (e.g., UE 106A to 106N) and the RAN node (e.g., base station 102) 811 can exchange control plane data via a Uu interface (e.g., LTE-Uu interface) through a protocol stack that includes a PHY layer 1001, a MAC layer 1002, an RLC layer 1003, a PDCP layer 1004, and an RRC layer 1005.

[0141] The Non-Access Stratum (NAS) protocol 1006 forms the highest layer of the control plane between the UE (e.g., UE 106A to 106N) 801 and the MME 821. The NAS protocol 1006 supports the mobility and session management procedures of the UE (e.g., UE 106A to 106N) 801 to establish and maintain IP connectivity between the UE (e.g., UE 106A to 106N) 801 and the P-GW.

[0142] The S1 Application Protocol (S1-AP) layer 1015 supports the functions of the S1 interface and includes basic procedures (EPs). An EP is the interaction unit between a RAN node (e.g., base station 102) 811 and a CN. The S1-AP layer services 1015 may include two sets: UE-associated services and non-UE-associated services. These services perform functions including, but not limited to: E-UTRAN Radio Access Bearer (E-RAB) management, UE capability indication, mobility, NAS signaling transmission, RAN Information Management (RIM), and configuration transmission.

[0143] The Flow Control Transmission Protocol (SCTP) layer (also known as the SCTP / IP layer) 1014 may, in part, rely on the IP protocol supported by the IP layer 1013 to ensure reliable delivery of signaling messages between the RAN node (e.g., base station 102) 811 and the MME 821. The L2 layer 1012 and L1 layer 1011 may refer to the communication links (e.g., wired or wireless communication links) used by the RAN node (e.g., base station 102) and the MME for exchanging information.

[0144] RAN nodes (e.g., base station 102) 811 and MME 821 can exchange control plane data via a protocol stack through the S1-MME interface, which includes L1 layer 1011, L2 layer 1012, IP layer 1013, SCTP layer 1014 and S1-AP layer 1015.

[0145] AI-based CSI feedback with CSI prediction In 3GPP standard development, Channel State Information (CSI) feedback has been a crucial theme in almost every 3GPP standard release. Furthermore, nearly all CSI codebook designs focus on feedback based on current CSI reference signal (CSI-RS) measurements, such as... Figure 11 As illustrated, in conventional CSI feedback (e.g., via 3GPP Release 17), the base station (e.g., gNB) periodically / semi-persistently or aperiodically sends CSI-RS to the UE, and the UE periodically provides CSI feedback (CSI-FB). The base station then applies the CSI feedback to generate the pre-decoded matrix index (PMI) until the next CSI feedback is available. Furthermore, assuming the channel remains substantially unchanged over time (e.g., within milliseconds), the base station can use older CSI feedback. Such a scheme works well for situations with low UE mobility. Figure 12An example of a possible 3GPP Release 18 enhancement is illustrated. It should be noted that 3GPP Release 18 is designed to enhance high-speed UEs using CSI prediction with conventional signal processing methods. As shown, the base station (e.g., gNB) can transmit a cluster of CSI-RS, and the UE can use at least one CSI-RS from the cluster to send CSI feedback to the base station based on the predicted CSI. The base station then applies the CSI feedback to generate a PMI (over time) until the next CSI feedback is available. However, in both instances, frequent CSI feedback is used. Therefore, improvements are desired.

[0146] The embodiments described herein provide apparatus, systems, and methods for receiving and measuring CSI-RS to train a model for predicting artificial intelligence (AI) based on channel state information. Massive multiple-input multiple-output (MIMO) systems using large antenna arrays in the gNB can significantly improve system performance. However, the benefits of massive MIMO rely on knowledge of the downlink CSI. In frequency division duplex (FDD) massive MIMO systems, due to the lack of channel reciprocity, the UE feeds back downlink CSI-RS measurements to the BS via the uplink. The feedback overhead can be significant due to the high dimensionality of CSI in massive MIMO systems. Furthermore, a large amount of power is used at the UE to provide feedback. One way to reduce the amount of feedback and lower power consumption at the UE is to use a model for predicting AI based on CSI. The UE can use the output of the CSI prediction AI model instead of CSI-RS. This reduces the number of CSI-RS that need to be sent from the gNB to the UE. However, training the CSI prediction AI model is still required. This can be achieved by using CSI-RS sent from the gNB to the UE. The CSI-RS can be specifically designated for the CSI prediction AI model. Alternatively, CSI-RS communicated for other UE requirements can also be used to train CSI prediction AI models.

[0147] For example, such as Figure 13 As illustrated, a base station (such as base station 102) can, for example, use a periodic pattern to configure CSI-RS transmission at times T0, T1, and T2. Furthermore, at time T2, a UE (such as UE 106) that has already received and measured CSI-RS transmissions at times T0, T1, and T2 can predict the channel at times T2+t1, T2+t2, etc. This prediction can be at least partially based on CSI-RS measurements, such as CSI-RS measurements at times T0, T1, and T2. In at least some instances, this can be achieved via an AI model (e.g., such as...). Figure 14A The illustrated AI model 1410 generates predictions. For example... Figure 14AAs shown, AI model 1410 can use past CSI-RS measurements (e.g., such as at times T0, T1, and T2) as input and can output a predicted channel at future times (e.g., at times T2+t1 and T2+t2) after the most recent (or last) CSI-RS measurement but before the next CSI-RS measurement. Times t1 and t2 can be times smaller than the time interval between T0, T1, T2, T3, and T4, such as... Figure 13 As shown. In some instances, the UE can calculate CSI feedback at times T2, T2+t1, and T2+t2, as well as joint CSI feedback at T2, T2+t1, and T2+t2, where T2+t2 is less than T3. Furthermore, the UE can then provide CSI feedback at a configured time T2' (e.g., the UE can provide a CSI codebook), where T2' is at least a processing time after T2.

[0148] In other instances, predictions can be generated via AI model 1410, which can be a one-dimensional long short-term memory (LSTM) AI model used for CSI prediction in the time domain. For example... Figure 14B As shown, AI model 1410 can be used solely to predict temporal correlations (such as, for example, LSTM). Time-series CSI measurements can be fed into an LSTM layer, which outputs a vector capturing the temporal dependencies. This vector is then fed into a fully connected (FC) layer to generate CSI prediction outputs.

[0149] In other instances, predictions may be generated via AI model 1410, which may be a two-dimensional convolutional neural network (“CNN”) AI model used for CSI prediction using both the time and frequency domains. Figure 14C As shown, AI model 1410 can be used to predict time-domain and frequency-domain correlations (such as, for example, 2D CNN). That is, the 2D CNN captures a batch of input data, where each sample may include a time series of CSI measurements across different frequency subcarriers. This input tensor is passed through a series of 2D convolutional layers (e.g., a neural network), enabling the model to identify patterns in the CSI that extend across both time steps and subcarriers. AI model 1410 can learn and provide CSI predictions across future time steps and subcarriers based on measurements of CSI-RS that are fed into the AI ​​model as training input.

[0150] In other instances, predictions can be generated via AI model 1410, which can be a three-dimensional convolutional neural network (CNN) AI model used for CSI prediction using the time domain, frequency domain, and spatial (antenna) domain. Figure 14DAs shown, AI model 1410 can be used to predict correlations in the time, frequency, and spatial domains (such as 3D CNNs). When the input passes through 3D convolutional layers (e.g., neural networks), AI model 1410 can learn and predict future CSI values ​​across the time, frequency, and antenna (spatial) domains. It should be noted that the model data collection classification information may differ due to design variations. In particular, 3D CNNs may be sensitive to network antenna panel design and virtualization. In such cases, auxiliary information may be needed for classifying the input data.

[0151] Figure 15 An example of signaling 1500 for performing data collection for channel state information (CSI) prediction (such as, for example, data collection for training AI models for CSI prediction on the UE side) is illustrated according to some implementation schemes. Figure 15 The signaling shown can also be used with any of the systems, methods, or devices. In various embodiments, some of the signaling shown may be executed concurrently in a different order than that shown, or may be omitted. Additional signaling may also be executed as needed. As shown, the signaling can be performed according to one of the following example embodiments.

[0152] Signaling may begin when a UE (such as UE 106) sends a data collection request 1510 to a base station (such as base station 102A, ..., 102N). In one implementation, the UE data collection request 1510 may be optional, as the base station may directly trigger data collection. For example, the base station may trigger the UE to perform the data collection request even if no data collection request 1510 is received from the UE. When the data collection request 1510 is optional, supported input / output configurations may be transmitted via UE capability information or UE assistance information (“UAI”).

[0153] In one example, the UE may use one or more processors (such as baseband processors 804A, ..., 804D) to encode a data collection request so that it can be sent from the UE to the base station. The data collection request 1510 may be used to request the network (i.e., the gNB) to transmit CSI-RS to the UE, which can be used for CSI prediction AI model training. To achieve supervised learning, the gNB will send CSI-RS at the requested input location and the requested output location. The UE uses the measured output location as training labels for supervised learning.

[0154] The UE may indicate in the data collection request 1510 for training the CSI prediction AI model whether to use auxiliary information for classifying the data. The auxiliary information for classification may include additional information useful in classifying the collected data into appropriate subsets. For example, the additional information may include information about the type of antenna used by the gNB, whether gNB antenna virtualization is used, the speed of the UE, etc. This auxiliary information can be used to classify CSI-RS measurements to provide more accurate modeling using the CSI prediction AI model.

[0155] In some instances, data collection request 1510 may also include any combination and / or all (at least one and / or one or more) of the following: requesting the UE to collect input data for the CSI prediction AI model and the type of data to be output from the CSI prediction AI model so that the UE can train the CSI prediction AI model and output the data of the CSI prediction AI model. In one example, data collection request 1510 may include information about the output of the CSI prediction AI model. For example, the data collection request may provide information including the total number of samples the CSI prediction AI model is to predict and the predicted distance between one or more samples. When the samples are periodic, the distance between each sample will be substantially the same. The input data of the CSI prediction AI model may also include any, any combination and / or all (at least one and / or one or more) of the following: the total number of samples collected in the time domain that are used as input to the CSI prediction AI model (i.e., CSI-RS measurements at the UE), the measured distance between one or more samples in the total number of samples collected in the time domain, and / or information for classifying the total number of samples.

[0156] In response to data collection request 1510, the UE may receive trigger command 1512 from a base station (e.g., gNB and / or network) to begin data collection for training the CSI prediction AI model. The UE may receive CSI-RS transmission 1514 from the base station (e.g., gNB and / or network). The UE may perform UE measurements transmitted via CSI-RS. The measurements transmitted via CSI-RS can be used to train the CSI prediction AI model. In some examples, the UE measurements of CSI-RS used for training the CSI prediction AI model may be cached at the UE.

[0157] In some examples, the UE may send a request 1516 to the base station to stop the collection of CSI-RS data used to train the CSI prediction AI model. In response, the UE may receive a data collection stop command 1518 from the base station. In one implementation, the UE stop request 1516 may not be used when the data collection request 1510 is optional (e.g., the UE stop request may also be optional depending on whether the UE data collection request 1510 is used).

[0158] In other words, the UE can measure CSI-RS and collect measurement data, and then send a request to the base station to stop collecting. In one aspect, when the CSI prediction AI model is a UE-side AI prediction model, the UE can cache the CSI-RS measurement data for training. In one example, the UE can pass the data to an over-the-top (“OTT”) server (e.g., a server accessed via the internet rather than a mobile network) for offline training of the CSI prediction AI model (e.g., the data transfer of CSI-RS measurements can be performed via a non-3GPP interface).

[0159] In one example, the UE may use an uplink (UL) radio resource control (RRC) message to transmit a stop request message 1516. For example, the UL RRC message may be transmitted within a UE assistance information (“UAI”) message. Alternatively, the UE may use an uplink media access control element (UL MAC CE) to transmit the stop request message 1516 to cease data collection for CSI-RS measurements. The base station may stop transmitting CSI-RS data used to train the CSI-prediction AI model, and the UE may stop performing CSI-RS measurements.

[0160] Therefore, the CSI prediction AI model can be executed / run on the UE side, allowing the UE to perform CSI-RS measurements and collect data. In one example, the UE has two options for handling the collected data: 1) cache the CSI-RS measurement data locally for model training, and / or 2) pass the collected data to an OTT (over-the-top) server via a non-3GPP interface (such as WiFi). This allows data to be stored on a server over the Internet for offline training without relying on the mobile operator's network. Once the UE has collected enough training data, it can signal the base station to stop the data collection process. The UE can transmit a UL RRC message (e.g., a UAI message) to indicate that it wants to stop data collection. Alternatively, the UE can transmit a UL MAC control element (CE) to indicate that it wants to stop data collection. Upon receiving the stop collection message, the base station can stop the data collection process by disabling the CSI-RS transmission configured for data collection.

[0161] In one example, if the base station (e.g., network "NW") trains a CSI prediction AI model, the UE can also transmit CSI-RS measurements 1520 to the base station. CSI-RS measurements can be encoded so that they can be transmitted from the UE to the base station or network using a user plane solution (such as the Physical Uplink Shared Channel (PUSCH)) or a control plane solution (such as the Physical Uplink Control Channel (PUCCH)).

[0162] Figure 16An example of a UE data collection request for collecting data for AI-based CSI prediction, according to some implementation schemes, is illustrated. A UE (such as UE 106) may send a data collection request 1510 to a base station (such as base station 102a, ..., 102n) for training a UE-side CSI AI prediction model. The data collection request 1510 may also include collecting any combination and / or all (at least one and / or one or more) of the following: requesting the collection of input data for the CSI AI prediction model and output data from the CSI AI prediction model to enable the CSI prediction AI model to be trained. The data collection request 1510 may include information about the expected output of the CSI AI prediction model, such as the total number of samples the model will predict and the prediction distance between one or more samples. When the samples are periodic, the prediction distance between each sample will be substantially the same. Furthermore, as previously discussed, the UE data collection request 1510 may include an indication of whether auxiliary information for classifying the data is needed. For example, the design of the CSI prediction AI model may require additional information from the network to classify / label the data during training. For example, models using spatial domain information may need to know antenna virtualization parameters. Spatial domain information can be used to classify CSI-RS measurement data.

[0163] The input data for the CSI prediction AI model may also include any combination and / or all (at least one and / or one or more) of the following: the total number of samples collected in the time domain that are used as input to the CSI prediction AI model, the measured distances between one or more samples from the total number of samples collected in the time domain, and / or information used to classify the total number of samples. When the samples are periodic, the distances between the samples will be substantially the same. More specifically, such as Figure 16 As illustrated, the UE data collection request may include: 1) the number of samples used to train the CSI AI prediction model. The number of samples can be the number of CSI-RS measurements (samples) that the UE's AI model will use to train the model to output a selected number of CSI predictions. For example, the AI ​​model may use n CSI measurements, where n is a positive integer. Therefore, the UE data collection request 1510 may identify n CSI-RS transmissions that can be used by the UE to obtain CSI-RS measurements, thereby training the CSI AI prediction model.

[0164] The UE data collection request may also include the measurement distance between CSI-RS measurement samples (e.g., the nth distance). CSI-RS can be transmitted periodically, such as every 5 milliseconds (ms). For example, CSI measurement samples can be collected in a periodic pattern, such as at times T0, T1, and T2. In this example, the UE receives T0, T1, and T2 at approximately 5 ms intervals. The UE can move at a specific speed, such as 30 km / h (kph). This will cause a certain distance to occur between each CSI-RS measurement interval. In this example, each CSI-RS measurement will occur at approximately 42 mm intervals. If the UE's speed changes significantly, the distance between measurements will change proportionally to the change in the UE's speed.

[0165] The data collection request may also include expected CSI prediction AI model output data. The output data of the CSI prediction AI model may also include any combination and / or all (at least one and / or one or more) of the following: the total number of samples to be predicted and the predicted distance between one or more samples. When the samples are periodic, the distances between samples will be substantially the same. Therefore, a UE (such as UE 106) that has received and measured CSI-RS transmissions at times T0, T1, and T2 can predict the number of samples at t0, t1, and t2 and the predicted distances between samples 1604. Unlike CSI-RS that is periodically received at the UE and measured as the UE moves, the output of the CSI prediction AI model is not speed-dependent. The distance between each sample can be configured to be substantially the same as the distance between CSI-RS measurements used as input to train the CSI prediction AI model. Alternatively, the CSI prediction AI model may output samples at a rate that is faster (i.e., higher) or slower (i.e., lower) than the CSI-RS measurements used as input. For example, CSI-RS transmissions may occur every 5 ms. The channel predictions output by the CSI prediction AI model can occur at the same rate, such as once every 5 ms, or at different rates, such as once every 0.5 ms, 1 ms, 2 ms, 3 ms, 4 ms, 6 ms, 7 ms, 8 ms, 9 ms, 10 ms, etc. This list of input and output rates is not intended to limit the actual CSI-RS transmission time; the CSI prediction AI model channel prediction sample output rate can be set based on system requirements, including but not limited to the UE's speed and the expected accuracy of the CSI prediction AI model channel sample output. This can result in a prediction distance of 1604 times that of the predicted channel state information from the CSI prediction AI model. Figure 16 The measured distance 1602 illustrated in the example is either shorter or longer.

[0166] For example, output information may include: 1) the number of samples to be predicted, which may refer to the number of future CSI values ​​that the AI ​​model will predict at one time based on the input samples, and 2) the prediction distance. The prediction distance may refer to the time distance of the predicted CSI values ​​in the future, which makes it possible to calculate the physical distance based on the speed at which the UE moves relative to the base station.

[0167] In some examples, the UE may transmit a data collection request message within an uplink radio resource control (RRC) message. Specifically, a UE assistance information (UAI) message may be used. That is, an RRC message can be a UAI message. The UAI message will contain the input and output information of the CSI prediction AI model. Alternatively, the UE may transmit the data collection request within an uplink media access control (MAC) element.

[0168] In one example, the UE can determine whether to trigger a data collection request 1510 based on specific conditions. For example, if the environment changes compared to the environment in which the model has been trained, such as when the UE moves from indoors to outdoors, the UE can trigger the request. This will cause a significant change in the CSI-RS measurement. The UE can also trigger the request if its speed changes, as the model may have been trained for a specific speed range. For example, when the UE's speed changes from 30 kph to 60 kph, the UE can trigger data collection request 1510. Furthermore, the UE may trigger data collection request 1510 when it moves into a cell with a different type of antenna, which may result in different CSI-RS measurements. Therefore, the UE can determine when to trigger a data collection request based on one or more of the following: the UE has not yet been trained using the CSI prediction AI model; the UE has moved from outdoors to indoors; the UE has moved from indoors to outdoors; the UE has moved into an environment where the number of obstacles has increased or decreased by a predetermined threshold; the UE's speed has changed by a predetermined amount; and / or the UE has moved into a cell with an antenna type different from the antenna type used by the cell's gNB.

[0169] Figure 17 An example is illustrated of a base station triggering a CSI-RS transmission for data collection for AI-based CSI prediction, according to some implementation schemes. In one example, based on the UE's data collection request and the input / output preferences of the indicated CSI prediction AI model, the base station may transmit a trigger command to begin data collection.

[0170] A base station (such as, for example, one of base stations 102A to 102N) can be configured to send CSI reference signal (CSI-RS) resources to the UE specifically for data collection. The data collection RS is periodically configured with a repetition rate for the CSI-RS used for data collection. Within each time period, the measurement CSI-RS and prediction CSI transmitted from the base station to the UE are configured to align with the channel measurement and prediction requirements of the UE's AI model. This provides a customized CSI-RS resource tailored for efficiently collecting training data for the CSI prediction AI model.

[0171] In one respect, if the UE currently has periodic CSI-RS resources configured for normal CSI measurements, and the periodicity is aligned with the expected input sample interval of the UE's CSI prediction AI model, these existing CSI-RS resources can be reused to improve efficiency.

[0172] For example, if p-CSI-RSs for CSI are already configured for regular / normal CSI measurements and reporting, and the sample distance is aligned with the input / output distance, then the data collection RSs can reuse periodic CSI-RS resources (“p-CSI-RSs”) for acquiring CSI. Therefore, if p-CSI-RS resources are already configured, and their periodicity happens to align with the measurement intervals required for CSI prediction AI model data collection, existing p-CSI-RSs can be used for dual purposes. In this way, the UE can use the same p-CSI-RSs used for normal CSI measurements to report to the network, and also to collect measurements as training data for its CSI prediction AI model. This avoids the need for the base station to configure completely separate CSI-RS resources solely for data collection, and for the UE to measure them, thus improving efficiency and reducing power usage at the UE. To enable p-CSI-RSs to be reused in this way, the periodicity (or measurement interval) of the p-CSI-RS can be aligned with the periodic measurement intervals of the data collection CSI-RSs used as input to the CSI prediction AI model, as indicated in the UE's data collection request 1510.

[0173] Alternatively, the base station can configure a separate CSI-RS resource set specifically for the data collection needs of the CSI prediction AI model. This configuration may include the periodicity or transmission timing, frequency / time, and port location of the CSI-RS, which the UE measures and provides as input to the CSI prediction AI model. In other words, the data collection CSI-RS can be a separately configured CSI-RS resource set based on the transmission periodicity, port, time, and / or frequency location of the CSI-RS.

[0174] For example, if no existing periodic CSI-RS resources (p-CSI-RS) are already configured, or if their periodicity does not match the measurement intervals required for CSI prediction AI model data collection, the base station can configure a separate set of CSI-RS resources dedicated to CSI prediction AI model data collection. These dedicated CSI resource sets for data collection can be configured with periodicity / transmission timing that matches the measurement sample intervals requested by the UE's data collection request 1510. The base station can also specify the port, time, and frequency locations of these dedicated data collection CSI-RSs within the system bandwidth. This allows the network to fully customize the CSI-RS configuration to meet the specific interval and quality requirements of data collection indicated in the UE's data collection request 1510. Therefore, if existing p-CSI-RS cannot be reused, the base station can configure a separate set of CSI-RS resources solely for CSI prediction AI model training data collection, where transmission periodicity and resource allocation are fully controlled to meet the UE's requested interval and quality requirements.

[0175] In one example, a CSI-RS resource set used solely for collecting training data for the CSI prediction AI model can be transmitted periodically, semi-persistently, or even aperiodically, if needed. If the CSI prediction AI model uses additional spatial domain or antenna information to classify the data, the base station can also configure and transmit these auxiliary classification data parameters. In one aspect, the auxiliary information may include the model type of the UE-side CSI prediction AI model, the base station's antenna type, and / or the base station's use of antenna virtualization. The base station can configure this auxiliary information and transmit it along with the CSI-RS to help the UE correctly classify the training data.

[0176] To improve data collection accuracy, CSI-RS resources can be enhanced to achieve higher-precision CSI measurements across multiple OFDM symbols. For example, for more accurate enhanced CSI-RS, CSI-RS can be enhanced to allow the collection of more accurate channel measurements as training data for CSI prediction AI models. One way to achieve higher accuracy is to be able to perform CSI measurements across multiple OFDM symbols, not just a single symbol. Extending CSI-RS measurements across multiple symbols allows for the collection of more samples, thereby reducing noise and more accurately determining channel characteristics. This improves the accuracy of training the CSI prediction AI model, thus improving the accuracy of the CSI prediction AI model's output. That is, instead of transmitting CSI-RS on a single symbol per time slot, CSI-RS can be transmitted across multiple symbols in a time slot or across consecutive time slots. Extending CSI-RS across multiple symbols allows the UE to obtain multiple samples of CSI-RS within a coherent time interval, thereby effectively averaging the measurements and reducing noise / interference. This, in turn, improves the accuracy of channel condition estimation.

[0177] In some instances, AI models can be used for CSI prediction, for example, such as Figure 18 As illustrated, for example, the CSI prediction AI model 1810 can be based on a specific UE implementation and can be trained and managed locally by the UE. Therefore, as shown, the UE can use past CSI-RS measurements as input to the AI ​​model 1810, and the AI ​​model 1810 can output a predicted channel. Furthermore, the predicted channel can then be used as input to the CSI encoder 1812 to provide uplink feedback to the base station (e.g., to the base station's CSI decoder 1820). It should be noted that if the UE can use the received CSI-RS for training to fine-tune the AI ​​model, the UE can request more frequent CSI-RS transmissions from the base station to facilitate faster fine-tuning.

[0178] Figure 19 A block diagram illustrating an example of a data collection method 1900 for predicting channel state information (CSI) based on artificial intelligence (AI) according to some implementation schemes is shown. Figure 19 The method shown can also be used in conjunction with any of the systems, methods, or devices shown in the figure, as well as other devices. In various implementations, some of the method elements shown may be executed concurrently in a different order than shown, or may be omitted. Additional method elements may also be executed as needed. As shown in the figure, the method can operate as follows.

[0179] At point 1910, a User Equipment (UE) (such as UE 106) may send a data collection request to a base station (such as base station 102) for training a CSI prediction AI model. The data collection request may also include any combination and / or all (at least one and / or one or more) of the following: a request to collect input data for the CSI prediction AI model and output data from the CSI prediction AI model so that the CSI prediction AI model can be trained. The data collection request may also identify the expected output data of the CSI prediction AI model, which may include the total number of samples to be predicted and the predicted distance between each sample. The input data of the CSI prediction AI model may also include any combination and / or all (at least one and / or one or more) of the following: the total number of samples collected in the time domain used as input to the CSI prediction AI model, the measured distance between each sample in the total number of samples collected in the time domain, and / or information for classifying the total number of samples. The output data of the CSI prediction AI model may also include any combination and / or all (at least one and / or one or more) of the following: the total number of samples to be predicted and the predicted distance between each sample.

[0180] At point 1920, the UE may receive a data collection request response from the base station. In some instances, the received data collection request response may be a command to perform UE measurements of CSI-RS for training a CSI prediction AI model.

[0181] At 1930, the UE can receive the CSI Reference Signal (CSI-RS) from the base station. At 1940, the UE can perform UE measurements of the CSI-RS based on the data collection request response from the base station for training the CSI prediction AI model.

[0182] At 1950, UE 106 can send a data collection stop request to the base station. At 1960, UE 106 can store UE measurements from CSI-RS in its memory.

[0183] In some instances, the UE can decode data collection stop request responses received from the base station. The UE can also encode UE measurements for transmission to the base station, enabling the base station to train a CSI prediction AI model. In some instances, the UE can cache UE measurements used to train the CSI prediction AI model.

[0184] In addition, in some instances, the UE may include a transceiver configured to receive CSI-RS from the base station for training a CSI prediction AI model, send data collection requests to the UE, and / or send UE measurements performed by the UE to the base station.

[0185] In some instances, the UE may send a data collection request from the UE to the base station via one or more of an uplink radio resource control (RRC) message and / or an uplink media access control (MAC) control element (CE).

[0186] In addition, in some instances, the UE may determine when to trigger a data collection request based on any one, any combination and / or all (at least one and / or one or more) of the following: the UE has not yet been trained using the CSI prediction AI model; the UE has moved from outdoors to indoors; the UE has moved from indoors to outdoors; the UE has moved to an environment where the number of obstacles has increased or decreased by a predetermined threshold; the UE's speed has changed by a predetermined amount; and / or the UE has moved to a cell with a different type of antenna compared to the previous cell.

[0187] In some instances, the UE may send an encoding request for auxiliary information to the base station to classify UE measurements based on the auxiliary information. The auxiliary information may include any one, any combination, and / or all (at least one and / or one or more) of the following: the model type of the UE-side CSI prediction AI model, the antenna type of the base station, and / or the base station's use of antenna virtualization.

[0188] In some instances, the UE may use a UE-side CSI prediction AI model, which may include any one, any combination, and / or all (at least one and / or one or more) of the following: a Long Short-Term Memory (LSTM) deep recurrent neural network model, a two-dimensional convolutional neural network model with input and output tensors including time and frequency dimensions, and / or a three-dimensional convolutional neural network model with input and output tensors including time, frequency, and antenna type dimensions. The CSI prediction AI model may include any one, any combination, and / or all (at least one and / or one or more) of the following: a UE-side prediction AI model and / or a network-side prediction AI model.

[0189] In some instances, the UE can encode UE measurements of CSI-RS used to train a CSI prediction AI model, enabling these measurements to be transmitted to an external server for offline training of the CSI prediction AI model. In some instances, the UE can receive time-domain repeated CSI-RS port transmissions from the base station to improve the accuracy of CSI-RS measurements. In some instances, the UE can receive CSI-RS measurements from multiple OFDM symbols decoded by the base station to further improve the accuracy of CSI-RS measurements.

[0190] As is widely recognized, the use of personally identifiable information should comply with privacy policies and practices that are generally accepted to meet or exceed industry or governmental requirements for protecting user privacy. Specifically, personally identifiable information data should be managed and processed to minimize the risk of unintentional or unauthorized access or use, and the nature of authorized use should be clearly explained to users.

[0191] Embodiments of this disclosure may be implemented in any of a variety of forms. For example, some embodiments may be implemented as a computer-implemented method, a computer-readable storage medium, or a computer system. Other embodiments may be implemented using one or more custom-designed hardware devices such as ASICs. Other embodiments may be implemented using one or more programmable hardware elements such as FPGAs.

[0192] In some embodiments, a non-transitory computer-readable storage medium may be configured to store program instructions and / or data, wherein, if executed by a computer system, the program instructions cause the computer system to perform a method, such as any method embodiment of the method embodiments described herein, or any combination of method embodiments described herein, or any subset or combination of any such subset of any method embodiments described herein.

[0193] In some embodiments, the device (e.g., UE 106) may be configured to include a processor (or a group of processors) comprising one or more baseband processors, one or more application processors, and a memory medium storing program instructions. The processor is configured to read from and execute the program instructions, which are executable to implement any of the various method embodiments described herein (or any combination of method embodiments described herein, or any subset of any method embodiments described herein, or any combination of such subsets). The device may be implemented in any of the various forms.

[0194] By interpreting each message / signal X received by the user equipment (UE) in the downlink as a message / signal X sent by the base station, and interpreting each message / signal Y sent by the UE in the uplink as a message / signal Y received by the base station, any of the methods described herein for operating the UE can serve as the basis for a corresponding method for operating the base station.

[0195] Although the above embodiments have been described in considerable detail, many variations and modifications will become apparent to those skilled in the art once the above disclosure is fully understood. It is intended that the following claims be construed as encompassing all such variations and modifications.

Claims

1. A user equipment (UE) apparatus configured to perform data collection for channel state information (CSI) prediction using an artificial intelligence (AI) model, the apparatus comprising: One or more processors, said one or more processors being configured to: The data collection request for training the CSI prediction AI model is encoded so that the data collection request can be sent to the next-generation node B (gNB). Decode the data collection request response received from the gNB; as well as UE measurements of the CSI Reference Signal (CSI-RS) are performed based on the data collection request response from the gNB to train the CSI prediction AI model; and A memory coupled to the one or more processors and configured to store the UE measurements of the CSI-RS.

2. The apparatus of claim 1, wherein the one or more processors are further configured to: Encode the data collection stop request to send to the gNB; or Decode the data collection stop request response received from the gNB.

3. The apparatus of claim 1, wherein the one or more processors are further configured to encode the UE measurements for transmission to the gNB, such that the gNB is able to train the CSI prediction AI model.

4. The apparatus of claim 1, wherein the one or more processors are further configured to cache the UE measurements to train the CSI prediction AI model.

5. The apparatus of claim 1, further comprising a transceiver configured to: Receive the CSI-RS from the gNB to train the CSI prediction AI model; Send the data collection request to the UE; and Send the UE measurement performed by the UE to the gNB.

6. The apparatus of claim 1, wherein the data collection request includes at least a request to collect input data for the CSI prediction AI model and output data from the CSI prediction AI model so that the gNB can train the CSI prediction AI model.

7. The apparatus according to any one of claims 1 to 6, wherein the input data for the CSI prediction AI model comprises one or more of the following: The total number of samples collected in the time domain and used as input for the CSI prediction AI model; The measured distance between one or more samples from the total number of samples collected in the time domain; and Information used to classify the total number of samples.

8. The apparatus according to any one of claims 1 to 7, wherein the output data of the CSI prediction AI model includes the total number of samples to be predicted and the prediction distance between one or more samples.

9. The apparatus of claim 1, wherein the one or more processors are further configured to encode the data collection request for transmission from the UE to the gNB via one or more of the following: Uplink Radio Resource Control (RRC) message; or Uplink Media Access Control (MAC) control element (CE).

10. The apparatus of claim 9, wherein the one or more processors are further configured to determine when to trigger the data collection request based on one or more of the following: The UE has not yet been trained using the CSI prediction AI model; The UE has been moved from outdoors to indoors; The UE has been moved from indoors to outdoors; The UE has moved to an environment where the number of obstacles has increased or decreased by a predetermined threshold; or The speed of the UE changed by a predetermined amount.

11. The apparatus of claim 1, wherein the one or more processors are further configured to encode a request for auxiliary information to be sent to the gNB to classify the UE measurements based on the auxiliary information.

12. The apparatus of claim 11, wherein the auxiliary information includes one or more of the following: Model type of AI model for UE-side CSI prediction; The antenna type of the gNB; or The gNB uses antenna virtualization.

13. The apparatus of claim 1, wherein the one or more processors are configured to use a UE-side CSI prediction AI model, the UE-side CSI prediction AI model comprising one or more of the following: Long Short-Term Memory (LSTM) deep recurrent neural network model; A two-dimensional convolutional neural network model, wherein the two-dimensional convolutional neural network model has an input tensor and an output tensor including time and frequency dimensions; or A three-dimensional convolutional neural network model, wherein the three-dimensional convolutional neural network model has input tensors and output tensors including time dimension, frequency dimension and antenna type.

14. The apparatus of claim 1, wherein the CSI prediction AI model is a UE-side prediction model.

15. The apparatus of claim 1, wherein the one or more processors are configured to encode the UE measurements of the CSI-RS used for training the CSI prediction AI model such that the UE measurements can be transmitted to an external server for offline training of the CSI prediction AI model.

16. The apparatus of claim 1, wherein the one or more processors are configured to receive time-domain repetitive CSI-RS port transmissions from the gNB to achieve increased CSI-RS measurement accuracy.

17. The apparatus of claim 1, wherein the one or more processors are configured to measure CSI-RS based on a plurality of OFDM symbols received from the gNB, thereby enabling a larger number of CSI-RS measurements to be provided to the CSI prediction AI model, thereby enabling the CSI prediction AI model to be trained with higher accuracy.

18. The apparatus of claim 1, wherein the data collection request response received from the gNB is a command for the UE to perform the UE measurement of the CSI-RS to collect input samples for training the CSI prediction AI model.

19. An apparatus for a user equipment (UE) configured to perform data collection for channel state information (CSI) prediction using an artificial intelligence (AI) model, the apparatus comprising: One or more processors, said one or more processors being configured to: The data collection request for training the CSI prediction AI model is encoded so that the data collection request can be sent to the next-generation node B (gNB). Decode the data collection request response received from the gNB, wherein the data collection request response received from the gNB includes a command for performing UE measurements of the CSI reference signal (CSI-RS) to train the CSI prediction AI model; The UE measurement of CSI-RS is performed based on the data collection request response from the gNB to train the CSI prediction AI model; Cache the UE measurements to train the CSI prediction AI model; The data collection stop request is encoded and sent to the gNB; as well as Decode the data collection stop request response received from the gNB; and A memory coupled to the one or more processors and configured to store the UE measurements of the CSI-RS.

20. The apparatus of claim 19, wherein the one or more processors are further configured to encode the UE measurements for transmission to the gNB to enable network-side training of the CSI prediction AI model.

21. The apparatus of claim 19, further comprising a transceiver configured to receive the CSI-RS from the gNB to train the CSI prediction AI model.

22. The apparatus of claim 19, further comprising a transceiver configured to: Send the data collection request to the UE; and Send the UE measurement performed by the UE to the gNB.

23. The apparatus of claim 19, wherein the data collection request includes at least a request to collect input data and output data of the CSI prediction AI model to enable the gNB to train the CSI prediction AI model, wherein: The input data for the CSI prediction AI model includes one or more of the following: The total number of samples collected in the time domain and used as input for the CSI prediction AI model; The measured distance between one or more samples from the total number of samples collected in the time domain; and Information used to classify the total number of samples; and The output data of the CSI prediction AI model includes the total number of samples to be predicted and the prediction distance between one or more samples.

24. The apparatus of claim 19, wherein the one or more processors are further configured to: The request for assistance information from the gNB is encoded to classify the UE measurements based on the assistance information, wherein the assistance information includes one or more of the following: The model type of the CSI prediction AI model; The antenna type of the gNB; or The gNB uses antenna virtualization.

25. The apparatus of claim 19, wherein the one or more processors are further configured to send the data collection request from the UE to the gNB via one or more of the following: Uplink Radio Resource Control (RRC) message; or Uplink Media Access Control (MAC) control element (CE).

26. The apparatus of claim 25, wherein the one or more processors are further configured to determine when to send the data collection request based on one or more of the following: The UE has not yet been trained using the CSI prediction AI model; The UE has been moved from outdoors to indoors; The UE has been moved from indoors to outdoors; The UE has moved to an environment where the number of obstacles has increased or decreased by a predetermined threshold; or The speed of the UE changed by a predetermined amount.

27. The apparatus of claim 19, wherein the one or more processors are configured to use a CSI prediction AI model, the CSI prediction AI model comprising one or more of the following: Long Short-Term Memory (LSTM) deep recurrent neural network model; A two-dimensional convolutional neural network model, wherein the two-dimensional convolutional neural network model has an input tensor and an output tensor including time and frequency dimensions; or A three-dimensional convolutional neural network model, wherein the three-dimensional convolutional neural network model has input tensors and output tensors including time dimension, frequency dimension and antenna type.

28. The apparatus of claim 19, wherein the CSI prediction AI model is one of the following: UE-side predictive AI model; or Network-side predictive AI model.

29. The apparatus of claim 19, wherein the one or more processors are configured to encode the UE measurements of the CSI-RS used for training the CSI prediction AI model such that the UE measurements can be transmitted to an external server for offline training of the CSI prediction AI model.

30. The apparatus of claim 19, wherein the one or more processors are configured to receive CSI-RS measurements based on a plurality of OFDM symbols decoded from the gNB to achieve a higher number of CSI-RS measurements for training the CSI prediction AI model, thereby achieving a higher accuracy output from the CSI prediction AI model.

31. An apparatus operable to assist in data collection for channel state information (CSI) prediction in a next-generation node B (gNB), the apparatus comprising: One or more processors, said one or more processors being configured to: The data collection request received from the user equipment (UE) is decoded at the gNB to train the UE-side CSI prediction AI model; The data collection request response is encoded for use in sending the data collection request response to the UE; The CSI reference signal (CSI-RS) to be transmitted from the gNB to the UE is encoded at the gNB so that the UE can perform UE measurements of the CSI-RS to train the UE-side CSI prediction AI model. as well as The data collection stop request is decoded at the gNB to stop the transmission of the CSI-RS from the gNB; and A memory configured to store the data contained in the data collection request.

32. The apparatus of claim 31, further comprising a transceiver configured to: The CSI-RS is transmitted to the UE to train the UE-side CSI prediction AI model; and Send the data collection request response to the UE.

33. The apparatus of claim 31, wherein the data collection request includes at least a request to collect input data for the UE-side CSI prediction AI model and output data from the UE-side CSI prediction AI model so that the gNB can train the UE-side CSI prediction AI model.

34. The apparatus according to any one of claims 31 to 33, wherein the input data for the UE-side CSI prediction AI model comprises one or more of the following: The total number of samples collected in the time domain that are used as inputs to the AI ​​model for predicting CSI on the UE side; The measured distance between one or more samples from the total number of samples collected in the time domain; and Information used to classify the total number of samples.

35. The apparatus according to any one of claims 31 to 33, wherein the output data of the UE-side CSI prediction AI model includes the total number of samples to be predicted and the prediction distance between one or more of the samples.

36. The apparatus of claim 31, wherein the one or more processors are further configured to decode the data collection request received from the UE at the gNB via one or more of the following: Uplink Radio Resource Control (RRC) message; or Uplink Media Access Control (MAC) control element (CE).

37. The apparatus of claim 31, wherein the one or more processors are further configured to decode a request for auxiliary information by the UE at the gNB, such that the UE can classify the UE measurements based on the auxiliary information.

38. The apparatus of claim 37, wherein the auxiliary information includes one or more of the following: The model type of the UE-side CSI prediction AI model; The antenna type of the gNB; or The gNB uses antenna virtualization.

39. The apparatus of claim 31, wherein the one or more processors are configured to use a UE-side CSI prediction AI model, the UE-side CSI prediction AI model comprising one or more of the following: Long Short-Term Memory (LSTM) deep recurrent neural network model; A two-dimensional convolutional neural network model, wherein the two-dimensional convolutional neural network model has an input tensor and an output tensor including time and frequency dimensions; or A three-dimensional convolutional neural network model, wherein the three-dimensional convolutional neural network model has input tensors and output tensors including time dimension, frequency dimension and antenna type.

40. The apparatus of claim 31, wherein the data collection request response transmitted to the UE is a command for performing the UE measurement of the CSI-RS to train the UE-side CSI prediction AI model.

41. A method for performing data collection for channel state information (CSI) prediction using an artificial intelligence (AI) model, the method comprising: Send a data collection request to the base station for training the UE-side CSI prediction AI model; Receive a data collection request response from the base station; Receive CSI reference signal (CSI-RS) from the base station; The UE measurement of the CSI-RS is performed based on the data collection request response from the base station to train the CSI prediction AI model; as well as Send a data collection stop request to the base station.

42. An apparatus for a user equipment (UE) configured to perform data collection for channel state information (CSI) prediction using an artificial intelligence (AI) model, the apparatus comprising: One or more processors, said one or more processors being configured to: Decode the data collection request received from the next-generation node B (gNB) for training the CSI prediction AI model; and UE measurements of CSI reference signals (CSI-RS) are performed based on the data collection request received from the gNB to train the CSI prediction AI model; and A memory coupled to the one or more processors and configured to store the UE measurements of the CSI-RS.

43. The apparatus of claim 42, wherein the one or more processors are further configured to decode a data collection stop request response received from the gNB.

44. The apparatus of claim 42, wherein the data collection request includes at least a request to collect input data and output data of the CSI prediction AI model to enable the gNB to train the CSI prediction AI model, wherein: The input data for the CSI prediction AI model includes one or more of the following: The total number of samples collected in the time domain and used as input for the CSI prediction AI model; The measured distance between one or more samples from the total number of samples collected in the time domain; and Information used to classify the total number of samples; and The output data of the CSI prediction AI model includes the total number of samples to be predicted and the prediction distance between one or more samples.

45. The apparatus of claim 44, wherein the one or more processors are further configured to: The auxiliary information received from the gNB is decoded so that the UE can classify the UE measurements based on the auxiliary information, wherein the auxiliary information includes one or more of the following: The model type of the CSI prediction AI model; The antenna type of the gNB; or The gNB uses antenna virtualization.

46. ​​The apparatus of claim 42, wherein the one or more processors are further configured to receive the data collection request from the gNB to perform the data collection request based on one or more of the following: The UE has not yet been trained using the CSI prediction AI model; The UE has been moved from outdoors to indoors; The UE has been moved from indoors to outdoors; The UE has moved to an environment where the number of obstacles has increased or decreased by a predetermined threshold; or The speed of the UE changed by a predetermined amount.

Images