Method and apparatus for hybrid inference for beam management in artificial intelligence / machine learning systems

Abstract

The WTRU is configured with input and output beam patterns and an AI / ML model with trained weights. Each input beam pattern includes a beam index of a set of input beams, and each output beam pattern includes a beam index of a set of output beams. Each AI / ML model corresponds to one output beam pattern. The WTRU measures the signal quality of a subset of the set of input beams for each input beam pattern to identify the best input beam and reports that beam to the base station. The WTRU receives an indication of one of the input beam patterns and one of the output beam patterns from the base station for predicting an optimal output beam based on the report. The WTRU predicts an optimal output beam using an AI / ML model corresponding to the indicated output beam pattern and the trained weights, and reports the predicted optimal beam to the base station.

Description

[0001] Cross-reference to related applications This application claims the benefit of U.S. Provisional Application Serial No. 63 / 531481, filed August 8, 2023; U.S. Provisional Application Serial No. 63 / 531485, filed August 8, 2023; U.S. Provisional Application Serial No. 63 / 531487, filed August 8, 2023; and U.S. Provisional Application Serial No. 63 / 531492, filed August 8, 2023, the contents of which are hereby incorporated herein by reference. Background Technology

[0002] A consensus has been reached on the 3GPP Radio Access Network (RAN) research project regarding Artificial Intelligence (AI) / Machine Learning (ML) for the 5G NR air interface. Beam management has been selected as one of the target use cases for AI / ML in the air interface. This technology can be a significant foundation for improving the performance and complexity of traditional beam management, including beam prediction for reducing overhead and latency in the time and / or spatial domains, improved beam selection accuracy, and more. Summary of the Invention

[0003] The WTRU is configured with input and output beam patterns and an AI / ML model with trained weights. Each input beam pattern includes a beam index of the input beam set, and each output beam pattern includes a beam index of the output beam set. Each AI / ML model corresponds to one output beam pattern. The WTRU measures the signal quality of a subset of the input beam sets for each input beam pattern to identify the optimal input beam and reports this beam to the base station. The WTRU receives indications of one of the input beam patterns and one of the output beam patterns from the base station and uses these indications to predict the optimal output beam. The WTRU uses the AI / ML model corresponding to the indicated output beam pattern and the trained weights to predict the optimal output beam. The WTRU reports the predicted optimal beam to the base station. Attached Figure Description

[0004] A more detailed understanding can be obtained by referring to the following description, which is given by way of example in conjunction with the accompanying drawings, wherein similar reference numerals in the figures denote similar elements, and wherein: Figure 1A This is a system diagram illustrating an example communication system in which one or more of the disclosed embodiments can be implemented; Figure 1B The illustrations based on one embodiment can be seen in Figure 1A A system diagram of an example wireless transmit / receive unit (WTRU) used within a communication system shown; Figure 1C The illustrations based on one embodiment can be seen in Figure 1AThe system diagram shows an example radio access network (RAN) and an example core network (CN) used within the communication system shown. Figure 1D The illustrations based on one embodiment can be seen in Figure 1A The system diagram shows another example RAN and another example CN used in the communication system shown; Figure 2 This is a diagram showing an example of the input beam pattern type (BPT); Figure 3 This is a diagram showing an example of the output BPT; Figure 4 A flowchart of an example method for beam prediction; Figure 5A A system diagram of an example wireless network configured to perform hybrid inference for beam management; Figure 5B Is Figure 5A A diagram of an example beam pattern at a base station; Figure 6 A diagram of an example AI / ML model used for hybrid inference on the base station side; Figure 7 It is used for hybrid inference in Figure 5A A diagram of an example AI / ML model for the WTRU side of the first and second WTRUs; Figure 8A A diagram illustrating an example of a beam index based on a generic default number; and Figure 8B A diagram showing another example of an indication of a beam index based on a generic default number. Detailed Implementation

[0005] Figure 1A This diagram illustrates an example communication system 100 that may implement one or more of the disclosed embodiments. The communication system 100 may be a multiple access system that provides content such as voice, data, video, messaging, and broadcasting to multiple wireless users. The communication system 100 enables multiple wireless users to access such content by sharing system resources, including wireless bandwidth. For example, the communication system 100 may employ one or more channel access methods, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal FDMA (OFDMA), Single Carrier FDMA (SC-FDMA), Zero-Tail Unique Word Discrete Fourier Transform Spread Spectrum OFDM (ZT-UW-DFT-S-OFDM), Unique Word OFDM (UW-OFDM), Resource Block Filtered OFDM, Filter Bank Multicarrier (FBMC), and the like.

[0006] like Figure 1A As shown, the communication system 100 may include wireless transmit / receive units (WTRUs) 102a, 102b, 102c, 102d, a radio access network (RAN) 104, a core network (CN) 106, a public switched telephone network (PSTN) 108, the Internet 110, and other networks 112. However, it will be appreciated that the disclosed embodiments are contemplated to any number of WTRUs, base stations, networks, and / or network elements. Each of the WTRUs 102a, 102b, 102c, 102d may be any type of device configured to operate and / or communicate in a wireless environment. For example, WTRUs 102a, 102b, 102c, and 102d—any of which can be referred to as a Station (STA)—can be configured to transmit and / or receive wireless signals and can include User Equipment (UE), mobile stations, fixed or mobile subscriber units, subscription-based units, pagers, cellular phones, personal digital assistants (PDAs), smartphones, laptops, netbooks, personal computers, wireless sensors, hotspots or Mi-Fi devices, Internet of Things (IoT) devices, watches or other wearable devices, head-mounted displays (HMDs), vehicles, drones, medical devices and applications (e.g., remote surgery), industrial devices and applications (e.g., robots and / or other wireless devices operating in industrial and / or automated processing chain environments), consumer electronics devices, devices operating on commercial and / or industrial wireless networks, and the like. Any of WTRUs 102a, 102b, 102c, and 102d can be interchangeably referred to as a UE.

[0007] The communication system 100 may also include base stations 114a and / or 114b. Each of base stations 114a and 114b may be any type of device configured to wirelessly interface with at least one of WTRUs 102a, 102b, 102c, and 102d to facilitate access to one or more communication networks, such as CN 106, the Internet 110, and / or other networks 112. For example, base stations 114a and 114b may be base transceiver stations (BTS), NodeBs, eNodeBs (eNBs), home NodeBs, home eNodeBs, next-generation NodeBs (e.g., gNodeBs (gNBs)), New Radio (NR) NodeBs, site controllers, access points (APs), wireless routers, and the like. Although base stations 114a and 114b are each depicted as a single element, it will be appreciated that base stations 114a and 114b may include any number of interconnected base station and / or network elements.

[0008] Base station 114a may be part of RAN 104, which may also include other base stations and / or network elements (not shown), such as base station controllers (BSCs), radio network controllers (RNCs), relay nodes, and the like. Base station 114a and / or base station 114b may be configured to transmit and / or receive radio signals on one or more carrier frequencies, which may be referred to as cells (not shown). These frequencies may be in licensed spectrum, unlicensed spectrum, or a combination of licensed and unlicensed spectrum. A cell may provide coverage for radio services to a specific geographic area, which may be relatively fixed or may change over time. The cell may also be divided into cell sectors. For example, the cell associated with base station 114a may be divided into three sectors. Thus, in one embodiment, base station 114a may include three transceivers, i.e., one transceiver per sector of the cell. In one embodiment, base station 114a may employ multiple-input multiple-output (MIMO) technology and may utilize multiple transceivers for each sector of the cell. For example, beamforming may be used to transmit and / or receive signals in a desired spatial direction.

[0009] Base stations 114a and 114b can communicate with one or more of WTRUs 102a, 102b, 102c, and 102d via air interface 116. Air interface 116 can be any suitable wireless communication link (e.g., radio frequency (RF), microwave, centimeter wave, micrometer wave, infrared (IR), ultraviolet (UV), visible light, etc.). Any suitable radio access technology (RAT) can be used to establish air interface 116.

[0010] More specifically, as described above, the communication system 100 can be a multiple access system and can employ one or more channel access schemes, such as CDMA, TDMA, FDMA, OFDMA, SC-FDMA, and the like. For example, base stations 114a and WTRUs 102a, 102b, and 102c in RAN 104 can implement radio technologies such as Universal Mobile Telecommunications System (UMTS) Terrestrial Radio Access (UTRA), which can establish an air interface 116 using Wideband CDMA (WCDMA). WCDMA can include communication protocols such as High-Speed ​​Packet Access (HSPA) and / or Evolved HSPA (HSPA+). HSPA can include High-Speed ​​Downlink (DL) Packet Access (HSDPA) and / or High-Speed ​​Uplink (UL) Packet Access (HSUPA).

[0011] In one embodiment, base station 114a and WTRUs 102a, 102b, 102c can implement radio technologies such as Evolved UMTS Terrestrial Radio Access (E-UTRA), which can establish an air interface 116 using Long Term Evolution (LTE) and / or Advanced LTE (LTE-A) and / or Advanced LTE Pro (LTE-A Pro).

[0012] In one embodiment, base station 114a and WTRUs 102a, 102b, 102c can implement radio technologies, such as NR radio access, which can use NR to establish air interface 116.

[0013] In one embodiment, base station 114a and WTRUs 102a, 102b, and 102c can implement multiple radio access technologies. For example, base station 114a and WTRUs 102a, 102b, and 102c can implement LTE radio access and NR radio access together, for example, using the dual connectivity (DC) principle. Therefore, the air interface utilized by WTRUs 102a, 102b, and 102c can be characterized by multiple types of radio access technologies and / or transmissions sent to / from multiple types of base stations (e.g., eNBs and gNBs).

[0014] In other embodiments, base station 114a and WTRUs 102a, 102b, 102c may implement radio technologies such as IEEE 802.11 (i.e., Wireless Fidelity (WiFi), IEEE 802.16 (i.e., Global Microwave Access Interoperability (WiMAX)), CDMA2000, CDMA2000 1X, CDMA2000 EV-DO, Provisional Standard 2000 (IS-2000), Provisional Standard 95 (IS-95), Provisional Standard 856 (IS-856), Global System for Mobile Communications (GSM), Enhanced Data Rate GSM Evolution (EDGE), GSMEDGE (GERAN), and the like.

[0015] Figure 1ABase station 114b can be, for example, a wireless router, a home Node B, a home eNode B, or an access point, and can utilize any suitable RAT to facilitate wireless connectivity in local areas such as commercial locations, homes, vehicles, campuses, industrial facilities, air corridors (e.g., for drone use), roads, and the like. In one embodiment, base station 114b and WTRUs 102c, 102d can implement radio technologies such as IEEE 802.11 to establish a wireless local area network (WLAN). In one embodiment, base station 114b and WTRUs 102c, 102d can implement radio technologies such as IEEE 802.15 to establish a wireless personal area network (WPAN). In yet another embodiment, base station 114b and WTRUs 102c, 102d can utilize cellular-based RATs (e.g., WCDMA, CDMA2000, GSM, LTE, LTE-A, LTE-APro, NR, etc.) to establish picocells or femtocells. Figure 1A As shown, base station 114b can be directly connected to Internet 110. Therefore, base station 114b does not need to access Internet 110 via CN 106.

[0016] RAN 104 can communicate with CN 106, which can be any type of network configured to provide voice, data, application, and / or Voice over Internet Protocol (VoIP) services to one or more of WTRUs 102a, 102b, 102c, and 102d. Data can have different Quality of Service (QoS) requirements, such as different throughput requirements, latency requirements, fault tolerance requirements, reliability requirements, data throughput requirements, mobility requirements, and the like. CN 106 can provide call control, billing services, location-based services, prepaid calling, internet connectivity, video distribution, and / or perform advanced security functions such as user authentication. Although in Figure 1A As not shown, but will be appreciated, RAN 104 and / or CN 106 can communicate directly or indirectly with other RANs that use the same RAT as or a different RAT than RAN 104. For example, in addition to connecting to RAN 104, which may utilize NR radio technology, CN 106 can also communicate with another RAN (not shown) that uses GSM, UMTS, CDMA2000, WiMAX, E-UTRA, or WiFi radio technology.

[0017] CN 106 can also serve as a gateway for WTRUs 102a, 102b, 102c, and 102d to access PSTN 108, the Internet 110, and / or other networks 112. PSTN 108 may include a circuit-switched telephone network providing Common Old-Style Telephone Service (POTS). The Internet 110 may include a global system of interconnected computer networks and devices using common communication protocols such as Transmission Control Protocol (TCP), User Datagram Protocol (UDP), and / or Internet Protocol (IP) from the TCP / IP Internet Protocol suite. Network 112 may include wired and / or wireless communication networks owned and / or operated by other service providers. For example, network 112 may include another CN connected to one or more RANs, which may use the same RAT as RAN 104 or a different RAT.

[0018] Some or all of the WTRUs 102a, 102b, 102c, and 102d in the communication system 100 may include multi-mode capability (e.g., WTRUs 102a, 102b, 102c, and 102d may include multiple transceivers for communicating with different wireless networks via different wireless links). For example... Figure 1A The WTRU 102c shown can be configured to communicate with base station 114a, which can employ cellular-based radio technology, and with base station 114b, which can employ IEEE 802 radio technology.

[0019] Figure 1B This is a system diagram illustrating the example WTRU 102. (Example: ...) Figure 1B As shown, WTRU 102 may include a processor 118, a transceiver 120, a transmit / receive element 122, a speaker / microphone 124, a keyboard 126, a display / touchpad 128, non-removable memory 130, removable memory 132, a power supply 134, a Global Positioning System (GPS) chipset 136, and / or other peripheral devices 138, etc. It will be appreciated that WTRU 102 may include any sub-combination of the foregoing elements while remaining consistent with the embodiments.

[0020] Processor 118 may be a general-purpose processor, a special-purpose processor, a conventional processor, a digital signal processor (DSP), multiple microprocessors, one or more microprocessors associated with a DSP core, a controller, a microcontroller, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), any other type of integrated circuit (IC), a state machine, and the like. Processor 118 may perform signal encoding, data processing, power control, input / output processing, and / or any other function that enables WTRU 102 to operate in a wireless environment. Processor 118 may be coupled to transceiver 120, and transceiver 120 may be coupled to transmitting / receiving element 122. Although Figure 1B The processor 118 and transceiver 120 are described as separate components, but it will be understood that the processor 118 and transceiver 120 can be integrated together in an electronic package or chip.

[0021] Transmitting / receiving element 122 can be configured to transmit signals to or receive signals from a base station (e.g., base station 114a) via air interface 116. For example, in one embodiment, transmitting / receiving element 122 can be an antenna configured to transmit and / or receive RF signals. In one embodiment, transmitting / receiving element 122 can be, for example, a transmitter / detector configured to transmit and / or receive IR, UV, or visible light signals. In yet another embodiment, transmitting / receiving element 122 can be configured to transmit and / or receive both RF and optical signals. It will be appreciated that transmitting / receiving element 122 can be configured to transmit and / or receive any combination of wireless signals.

[0022] Despite Figure 1B While the transmit / receive element 122 is described as a single element, the WTRU 102 may include any number of transmit / receive elements 122. More specifically, the WTRU 102 may employ MIMO technology. Thus, in one embodiment, the WTRU 102 may include two or more transmit / receive elements 122 (e.g., multiple antennas) for transmitting and receiving wireless signals via the air interface 116.

[0023] Transceiver 120 can be configured to modulate signals to be transmitted by transmitting / receiving element 122 and demodulate signals received by transmitting / receiving element 122. As described above, WTRU 102 can have multimode capability. Therefore, transceiver 120 can include multiple transceivers for example enabling WTRU 102 to communicate via multiple RATs (such as NR and IEEE 802.11).

[0024] The processor 118 of WTRU 102 can be coupled to and receive user input data from: a speaker / microphone 124, a keyboard 126, and / or a display / touchpad 128 (e.g., a liquid crystal display (LCD) unit or an organic light-emitting diode (OLED) display unit). The processor 118 can also output user data to the speaker / microphone 124, keyboard 126, and / or display / touchpad 128. Furthermore, the processor 118 can access information from and store data in any suitable type of memory (such as non-removable memory 130 and / or removable memory 132). Non-removable memory 130 may include random access memory (RAM), read-only memory (ROM), a hard disk, or any other type of memory storage device. Removable memory 132 may include a subscriber identity module (SIM) card, memory stick, secure digital storage (SD) card, and the like. In other embodiments, the processor 118 can access information from and store data in memory that is not physically located on WTRU 102 (such as a server or home computer (not shown)).

[0025] The processor 118 may receive power from the power supply 134 and may be configured to distribute and / or control power to other components in the WTRU 102. The power supply 134 may be any suitable device for powering the WTRU 102. For example, the power supply 134 may include one or more dry cell batteries (e.g., nickel-cadmium (NiCd), nickel-zinc (NiZn), nickel metal hydride (NiMH), lithium-ion (Li-ion), etc.), solar cells, fuel cells, and the like.

[0026] The processor 118 may also be coupled to a GPS chipset 136, which may be configured to provide location information (e.g., longitude and latitude) about the current location of the WTRU 102. In addition to, or instead of, information from the GPS chipset 136, the WTRU 102 may receive location information from base stations (e.g., base stations 114a, 114b) via the air interface 116, and / or determine its location based on the timing of signals received from two or more nearby base stations. It will be appreciated that the WTRU 102 may acquire location information using any suitable location determination method while remaining consistent with the embodiments.

[0027] The processor 118 may be further coupled to other peripheral devices 138, which may include one or more software and / or hardware modules providing additional features, functions, and / or wired or wireless connectivity. For example, peripheral devices 138 may include accelerometers, electronic compasses, satellite transceivers, digital cameras (for photos and / or videos), Universal Serial Bus (USB) ports, vibration devices, television transceivers, hands-free headsets, Bluetooth modules, FM radio units, digital music players, media players, video game player modules, internet browsers, virtual reality and / or augmented reality (VR / AR) devices, activity trackers, and the like. Peripheral devices 138 may include one or more sensors. These sensors may be one or more of the following: gyroscopes, accelerometers, Hall effect sensors, magnetometers, orientation sensors, proximity sensors, temperature sensors, time sensors, geolocation sensors, altimeters, light sensors, touch sensors, magnetometers, barometers, attitude sensors, biosensors, humidity sensors, and the like.

[0028] WTRU 102 may include a full-duplex radio for which transmission and reception of some or all signals (e.g., associated with a specific subframe for both UL (e.g., for transmission) and DL (e.g., for reception)) may be concurrent and / or simultaneous. The full-duplex radio may include an interference management unit to reduce and / or substantially eliminate self-interference via hardware (e.g., a choke) or via signal processing by a processor (e.g., a separate processor (not shown) or via processor 118). In one embodiment, WTRU 102 may include a half-duplex radio for which transmission and reception of some or all signals (e.g., associated with a specific subframe for UL (e.g., for transmission) or DL ​​(e.g., for reception) may be concurrent and / or simultaneous.

[0029] Figure 1C This is a system diagram illustrating RAN 104 and CN 106 according to one embodiment. As described above, RAN 104 can communicate with WTRUs 102a, 102b, and 102c via air interface 116 using E-UTRA radio technology. RAN 104 can also communicate with CN 106.

[0030] RAN 104 may include eNode-Bs 160a, 160b, and 160c, although it will be understood that RAN 104 may include any number of eNode-Bs while remaining consistent with the embodiments. eNode-Bs 160a, 160b, and 160c may each include one or more transceivers for communicating with WTRUs 102a, 102b, and 102c via air interface 116. In one embodiment, eNode-Bs 160a, 160b, and 160c may implement MIMO technology. Therefore, for example, eNode-B 160a may use multiple antennas to transmit radio signals to and / or receive radio signals from WTRU 102a.

[0031] Each of the eNode-B 160a, 160b, and 160c can be associated with a specific cell (not shown) and can be configured to handle radio resource management decisions, handover decisions, user scheduling in the UL and / or DL, and the like. Figure 1C As shown, eNode-B 160a, 160b, and 160c can communicate with each other via the X2 interface.

[0032] Figure 1C The CN 106 shown may include a Mobility Management Entity (MME) 162, a Serving Gateway (SGW) 164, and a Packet Data Network (PDN) Gateway (PGW) 166. While the foregoing elements are depicted as part of CN 106, it will be understood that any of these elements may be owned and / or operated by an entity other than a CN operator.

[0033] The MME 162 can connect to each of the eNode-Bs 160a, 160b, and 160c in RAN 104 via the S1 interface and can be used as a control node. For example, the MME 162 can be responsible for authenticating users of WTRUs 102a, 102b, and 102c, bearer activation / deactivation, selecting a specific serving gateway during the initial attachment of WTRUs 102a, 102b, and 102c, and so on. The MME 162 can provide control plane functions for switching between RAN 104 and other RANs (not shown) employing other radio technologies such as GSM and / or WCDMA.

[0034] The SGW 164 can connect to each of the eNode-Bs 160a, 160b, and 160c in RAN 104 via the S1 interface. The SGW 164 can typically route and forward user data packets to / from WTRUs 102a, 102b, and 102c. The SGW 164 can perform other functions such as anchoring the user plane during inter-eNodeB handover, triggering paging when DL data is available for WTRUs 102a, 102b, and 102c, managing and storing the context of WTRUs 102a, 102b, and 102c, and so on.

[0035] The SGW 164 can be connected to the PGW 166, which can provide WTRU 102a, 102b, 102c with access to packet-switched networks (such as Internet 110) to facilitate communication between WTRU 102a, 102b, 102c and IP-enabled devices.

[0036] CN 106 can facilitate communication with other networks. For example, CN 106 can provide WTRU 102a, 102b, 102c with access to a circuit-switched network such as PSTN 108 to facilitate communication between WTRU 102a, 102b, 102c and conventional landline communication equipment. For example, CN 106 may include or communicate with an IP gateway (e.g., an IP Multimedia Subsystem (IMS) server) serving as an interface between CN 106 and PSTN 108. Furthermore, CN 106 can provide WTRU 102a, 102b, 102c with access to other networks 112, which may include other wired and / or wireless networks owned and / or operated by other service providers.

[0037] Despite WTRU in Figures 1A-1D While described as a wireless terminal, it is envisioned that in some representative embodiments, such a terminal may (e.g., temporarily or permanently) use a wired communication interface with a communication network.

[0038] In a representative embodiment, another network 112 may be a WLAN.

[0039] In Infrastructure Basic Services Set (BSS) mode, a WLAN may have an Access Point (AP) for the BSS and one or more Stations (STAs) associated with the AP. The AP may have access to or interfacing with a Distributed System (DS) or carry services within and / or out of the BSS to another type of wired / wireless network. Traffic originating outside the BSS destined for a STA can reach and be delivered to the STA via the AP. Traffic originating from a STA destined outside the BSS can be sent to the AP for delivery to the appropriate destination. For example, traffic between STAs within the BSS can be sent via the AP, where the source STA can send traffic to the AP, and the AP can deliver traffic to the destination STA. Traffic between STAs within the BSS can be considered and / or referred to as peer-to-peer traffic. Peer-to-peer traffic can be sent between a source STA and a destination STA (e.g., directly between the source STA and the destination STA) using Direct Link Establishment (DLS). In some representative embodiments, the DLS may use 802.11e DLS or 802.11z Tunneled DLS (TDLS). A WLAN using the Standalone BSS (IBSS) mode may not have an access point (AP), and STAs within the IBSS or using the IBSS (e.g., all STAs) can communicate directly with each other. The IBSS communication mode is sometimes referred to as the "self-organizing" communication mode in this document.

[0040] When using 802.11ac infrastructure operating mode or a similar operating mode, the AP can transmit beacons on a fixed channel, such as the primary channel. The primary channel can be of a fixed width (e.g., a wide bandwidth of 20 MHz) or a dynamically configured width. The primary channel can be the operating channel of the BSS and can be used by the STA to establish a connection with the AP. In some representative embodiments, such as in an 802.11 system, Carrier Sense Multiple Access (CSMA / CA) with collision avoidance can be implemented. For CSMA / CA, each STA, including the AP, can listen on the primary channel. If the primary channel is listened to / detected by a particular STA and / or determined to be busy, that particular STA can back off. A single STA (e.g., only one station) can transmit at any given time within a given BSS.

[0041] High-throughput (HT) STAs can communicate using a 40 MHz wide channel, for example, by combining a primary 20 MHz channel with adjacent or non-adjacent 20 MHz channels.

[0042] Very High Throughput (VHT) STAs can support channels with widths of 20 MHz, 40 MHz, 80 MHz, and / or 160 MHz. 40 MHz and / or 80 MHz channels can be formed by combining adjacent 20 MHz channels. A 160 MHz channel can be formed by combining eight adjacent 20 MHz channels, or by combining two non-adjacent 80 MHz channels—this can be referred to as an 80+80 configuration. For the 80+80 configuration, after channel coding, the data passes through a segment resolver, which splits the data into two streams. Each stream can be processed separately using Inverse Fast Fourier Transform (IFFT) and time-domain processing. These streams can be mapped onto two 80 MHz channels, and the data can be transmitted by the transmitting STA. At the receiver of the receiving STA, the above operations for the 80+80 configuration can be reversed, and the combined data can be sent to the Media Access Control (MAC).

[0043] 802.11af and 802.11ah support sub-1 GHz operating modes. Compared to the operating modes used in 802.11n and 802.11ac, the channel operating bandwidth and carrier in 802.11af and 802.11ah are reduced. 802.11af supports 5 MHz, 10 MHz, and 20 MHz bandwidths in the TV Blank (TVWS) spectrum, and 802.11ah supports 1 MHz, 2 MHz, 4 MHz, 8 MHz, and 16 MHz bandwidths using non-TVWS spectrum. According to a representative embodiment, 802.11ah may support metering-type control / machine-type communication (MTC), such as MTC devices in macro coverage areas. MTC devices may have certain capabilities (e.g., limited capabilities), including supporting (e.g., only supporting) certain and / or limited bandwidths. MTC devices may include batteries with a battery life exceeding a threshold (e.g., to maintain a very long battery life).

[0044] WLAN systems that support multiple channels and channel bandwidths (such as 802.11n, 802.11ac, 802.11af, and 802.11ah) include channels that can be designated as primary channels. The bandwidth of the primary channel can be equal to the maximum common operating bandwidth supported by all STAs in the BSS. The bandwidth of the primary channel can be set and / or limited by the STAs that support the minimum bandwidth operating mode among all STAs operating in the BSS. In the 802.11ah example, for STAs that support (e.g., only support) the 1 MHz mode (e.g., MTC type devices), the primary channel can be 1 MHz wide, even if the AP and other STAs in the BSS support 2 MHz, 4 MHz, 8 MHz, 16 MHz, and / or other channel bandwidth operating modes. Carrier Sense and / or Network Assignment Vector (NAV) settings can depend on the status of the primary channel. If the primary channel is busy, for example because an STA (which only supports the 1 MHz operating mode) is transmitting to the AP, all available bands can be considered busy, even if most available bands remain idle.

[0045] In the United States, the available frequency bands for 802.11ah are from 902 MHz to 928 MHz. In South Korea, the available bands are from 917.5 MHz to 923.5 MHz. In Japan, the available bands are from 916.5 MHz to 927.5 MHz. The total available bandwidth for 802.11ah is 6 MHz to 26 MHz, depending on the status code.

[0046] Figure 1D This diagram illustrates a system diagram of RAN 104 and CN 106 according to one embodiment. As described above, RAN 104 can communicate with WTRUs 102a, 102b, and 102c via air interface 116 using NR radio technology. RAN 104 can also communicate with CN 106.

[0047] RAN 104 may include gNBs 180a, 180b, and 180c, although it should be understood that RAN 104 may include any number of gNBs while remaining consistent with the embodiments. gNBs 180a, 180b, and 180c may each include one or more transceivers for communicating with WTRUs 102a, 102b, and 102c via air interface 116. In one embodiment, gNBs 180a, 180b, and 180c may implement MIMO technology. For example, gNBs 180a and 180b may utilize beamforming to transmit signals to and / or receive signals from gNBs 180a, 180b, and 180c. Thus, for example, gNB 180a may use multiple antennas to transmit radio signals to and / or receive radio signals from WTRU 102a. In one embodiment, gNBs 180a, 180b, and 180c can implement carrier aggregation technology. For example, gNB 180a can transmit multiple component carriers (not shown) to WTRU 102a. A subset of these component carriers may be on unlicensed spectrum, while the remaining component carriers may be on licensed spectrum. In one embodiment, gNBs 180a, 180b, and 180c can implement Coordinated Multipoint (CoMP) technology. For example, WTRU 102a can receive coordinated transmissions from gNBs 180a and 180b (and / or gNB 180c).

[0048] WTRUs 102a, 102b, and 102c can communicate with gNBs 180a, 180b, and 180c using transmissions associated with extendable digitization. For example, the OFDM symbol spacing and / or OFDM subcarrier spacing can differ for different transmissions, different cells, and / or different portions of the radio transmission spectrum. WTRUs 102a, 102b, and 102c can communicate with gNBs 180a, 180b, and 180c using various or extendable lengths of subframes or transmission time intervals (TTIs) (e.g., containing different numbers of OFDM symbols and / or varying absolute durations).

[0049] gNBs 180a, 180b, and 180c can be configured to communicate with WTRUs 102a, 102b, and 102c in standalone and / or non-standalone configurations. In standalone configuration, WTRUs 102a, 102b, and 102c can communicate with gNBs 180a, 180b, and 180c without access to other RANs (e.g., eNode-B160a, 160b, and 160c). In standalone configuration, WTRUs 102a, 102b, and 102c can utilize one or more of gNBs 180a, 180b, and 180c as mobility anchors. In standalone configuration, WTRUs 102a, 102b, and 102c can communicate with gNBs 180a, 180b, and 180c using signals in unlicensed frequency bands. In a non-standalone configuration, WTRUs 102a, 102b, and 102c can communicate / connect with gNBs 180a, 180b, and 180c, while also communicating / connecting with another RAN such as eNode-Bs 160a, 160b, and 160c. For example, WTRUs 102a, 102b, and 102c can implement DC principles to communicate substantially simultaneously with one or more gNBs 180a, 180b, and 180c, as well as one or more eNode-Bs 160a, 160b, and 160c. In a non-standalone configuration, eNode-Bs 160a, 160b, and 160c can act as mobility anchors for WTRUs 102a, 102b, and 102c, and gNBs 180a, 180b, and 180c can provide additional coverage and / or throughput for serving WTRUs 102a, 102b, and 102c.

[0050] Each of gNBs 180a, 180b, and 180c can be associated with a specific cell (not shown) and can be configured to handle radio resource management decisions, handover decisions, user scheduling in UL and / or DL, network slicing support, DC, interoperability between NR and E-UTRA, routing of user plane data to User Plane Functions (UPF) 184a and 184b, routing of control plane information to Access and Mobility Management Functions (AMF) 182a and 182b, and the like. Figure 1D As shown, gNB180a, 180b, and 180c can communicate with each other via the Xn interface.

[0051] Figure 1DThe CN 106 shown may include at least one AMF 182a, 182b, at least one UPF 184a, 184b, at least one Session Management Function (SMF) 183a, 183b, and possibly a Data Network (DN) 185a, 185b. Although the foregoing elements are depicted as part of CN 106, it will be understood that any of these elements may be owned and / or operated by an entity other than a CN operator.

[0052] AMF 182a and 182b can connect to one or more of gNBs 180a, 180b, and 180c in RAN 104 via the N2 interface and can be used as control nodes. For example, AMF 182a and 182b can be responsible for authenticating users of WTRU 102a, 102b, and 102c, supporting network slicing (e.g., handling different Protocol Data Unit (PDU) sessions with different requirements), selecting specific SMF 183a and 183b, managing registration areas, terminating Non-Access Stratum (NAS) signaling, mobility management, and so on. AMF 182a and 182b can use network slicing to customize CN support for WTRU 102a, 102b, and 102c based on the service types being used by WTRU 102a, 102b, and 102c. For example, different network slices can be established for different use cases, such as services that rely on Ultra Reliable Low Latency (URLLC) access, services that rely on Enhanced Massive Mobile Broadband (eMBB) access, services for MTC access, and so on. AMF182a and 182b can provide control plane functions for switching between RAN104 and other RANs (not shown) that employ other radio technologies such as LTE, LTE-A, LTE-A Pro, and / or non-3GPP access technologies such as WiFi.

[0053] SMFs 183a and 183b can connect to AMFs 182a and 182b in CN 106 via the N11 interface. SMFs 183a and 183b can also connect to UPFs 184a and 184b in CN 106 via the N4 interface. SMFs 183a and 183b can select and control UPFs 184a and 184b, and configure service routes through UPFs 184a and 184b. SMFs 183a and 183b can perform other functions, such as managing and assigning UE IP addresses, managing PDU sessions, controlling policy enforcement and QoS, providing DL data notifications, and so on. PDU session types can be IP-based, non-IP-based, Ethernet-based, and so on.

[0054] UPF 184a and 184b can be connected via an N3 interface to one or more of gNB 180a, 180b, and 180c in RAN 104. This N3 interface provides WTRU 102a, 102b, and 102c with access to a packet-switched network (such as the Internet 110) to facilitate communication between WTRU 102a, 102b, 102c and IP-enabled devices. UPF 184 and 184b can perform other functions such as routing and forwarding packets, enforcing user plane policies, supporting multi-destination PDU sessions, handling user plane QoS, buffering DL packets, providing mobility anchoring, and so on.

[0055] CN 106 can facilitate communication with other networks. For example, CN 106 may include or communicate with an IP gateway (e.g., an IP Multimedia Subsystem (IMS) server) acting as an interface between CN 106 and PSTN 108. Furthermore, CN 106 can provide WTRUs 102a, 102b, and 102c with access to other networks 112, which may include other wired and / or wireless networks owned and / or operated by other service providers. In one embodiment, WTRUs 102a, 102b, and 102c may be connected to local DNs 185a and 185b via UPF 184a and 184b through the N3 interface to UPF 184a and 184b and the N6 interface between UPF 184a and 184b and DNs 185a and 185b.

[0056] Given Figures 1A-1D and Figures 1A-1D The corresponding descriptions herein regarding one or more of the functions of WTRU 102a-d, base station 114a-b, eNode-B 160a-c, MME 162, SGW164, PGW 166, gNB 180a-c, AMF 182a-b, UPF 184a-b, SMF 183a-b, DN 185a-b, and / or any other device described herein may be performed by one or more emulation devices (not shown). An emulation device may be one or more devices configured to emulate one or more of the functions described herein. For example, an emulation device may be used to test other devices and / or simulate network and / or WTRU functions.

[0057] Simulation devices can be designed to perform one or more tests on other devices in a laboratory environment and / or a carrier network environment. For example, one or more simulation devices can perform one or more functions while being fully or partially implemented and / or deployed as part of a wired and / or wireless communication network to test other devices within the communication network. One or more simulation devices can perform one or more functions while being temporarily implemented / deployed as part of a wired and / or wireless communication network. Simulation devices can be directly coupled to another device and / or use over-the-air wireless communication to perform tests for testing purposes.

[0058] One or more emulation devices may perform one or more functions (including all functions) but are not implemented / deployed as part of a wired and / or wireless communication network. For example, emulation devices may be used in test scenarios in a test laboratory and / or in non-deployed (e.g., testing) wired and / or wireless communication networks to implement the testing of one or more components. One or more emulation devices may be test devices. Emulation devices may transmit and / or receive data using direct RF coupling and / or wireless communication via RF circuitry (e.g., which may include one or more antennas).

[0059] Artificial intelligence can be broadly defined as the behavior exhibited by machines. For example, this behavior can mimic cognitive functions such as sensing, reasoning, adapting, and acting. Machine learning can refer to a type of algorithm based on learning through experience without explicit programming to solve problems. Machine learning can be considered a subset of AI. Different machine learning paradigms can be envisioned based on the nature of the data or feedback available for learning the algorithm. For example, supervised learning methods may involve learning a function that maps inputs to outputs based on labeled training examples, where each training example can be a pair consisting of an input and a corresponding output. For example, unsupervised learning methods may involve detecting patterns in data without pre-existing labels. For example, reinforcement learning methods may involve performing a series of actions in an environment to maximize cumulative rewards. It is possible to apply machine learning algorithms using combinations or interpolations of the above methods. For example, semi-supervised learning methods can use a combination of a small amount of labeled data and a large amount of unlabeled data during training. In this respect, semi-supervised learning lies between unsupervised learning (without labeled training data) and supervised learning (with only labeled training data).

[0060] Deep learning refers to a class of machine learning algorithms that employ artificial neural networks, also known as deep neural networks (DNNs). DNNs are a special type of machine learning model inspired by the human brain, where the input is linearly transformed and passed multiple times through a non-linear activation function. A DNN can consist of multiple layers, each containing a linear transformation and a given non-linear activation function. DNNs can be trained using training data via the backpropagation algorithm. Recently, DNNs have demonstrated state-of-the-art performance in various fields, such as speech, vision, and natural language processing, and in various machine learning settings, such as supervised, unsupervised, and semi-supervised learning settings.

[0061] AI / ML-based methods / processes enable the achievement of behaviors and / or compliance through data-driven learning, without the need for explicit configuration, steps, or sequences of actions. These methods allow machines to learn complex behaviors that might be difficult to specify and / or achieve using traditional methods. AI / ML models, referred to herein as implementations of AI / ML-based methods, can consist of model parameters and model structure. For example, a DNN-based AI / ML model may include model parameters such as weights and biases; and model structure such as the type and size of each layer of a deep neural network (e.g., dense layers and convolutional layers).

[0062] A WTRU can transmit or receive a physical channel or reference signal based on at least one spatial domain filter. A beam, as used herein, can refer to such a spatial domain filter. A WTRU can use the same spatial domain filter used to receive a reference signal (RS) to transmit a physical channel or signal, such as a Channel State Information (CSI-RS) or Synchronization Signal (SS) block. The WTRU transmission can be referred to as the target, and the received RS or SS block can be referred to as the reference or source. In this case, it can be said that the WTRU transmits the target physical channel or signal based on the spatial relationship of such RS or SS blocks.

[0063] The WTRU can transmit the first physical channel or signal using the same spatial domain filter as the spatial domain filter used to transmit the second physical channel or signal. The first and second transmissions can be referred to as the target and reference (or source), respectively. In this case, it can be said that the WTRU transmits the first target physical channel or signal according to its spatial relationship with respect to the second reference physical channel or signal.

[0064] Spatial relationships can be implicit, configured by the RRC, or signaled by the MAC Control Element (CE) or Downlink Control Information (DCI). For example, the WTRU can implicitly transmit the Physical Uplink Shared Channel (PUSCH) and the Demodulation Reference Signal (DM-RS) of the PUSCH based on the same spatial domain filter as the Schedule Request Indicator (SRI) indicated by the DCI or configured by the Radio Resource Control (RRC). In another example, spatial relationships can be configured by the RRC for the SRS Resource Indicator (SRI) or signaled by the MAC CE for the Physical Uplink Control Channel (PUCCH). This spatial relationship can also be referred to as beam indication.

[0065] The WTRU can receive a first target downlink channel or signal based on the same spatial filter or spatial reception parameters as the second reference downlink channel or signal. For example, this association can exist between a physical channel (e.g., a Physical Downlink Control Channel (PDCCH) or a Physical Downlink Shared Channel (PDSCH)) and its corresponding DM-RS. This association may exist at least when the first and second signals are reference signals, and when the WTRU is configured with Quasi-Cooperative Positioning (QCL) assumption type D between the corresponding antenna ports. This association can be configured as a Transmission Configuration Indicator (TCI) state. The association between the CSI-RS or SS block and the DM-RS can be indicated to the WTRU by an index of a set of TCI states configured by the RRC and / or signaled by the MAC CE. This indication can also be called beam indication.

[0066] A WTRU can be configured with one or more sets of RS resources and / or sets of beams (or beam pairs). Each RS resource, beam, or beam pair can be associated with a transmission from a beam having specific beam parameters (e.g., beam direction and beamwidth). The WTRU can be configured with associated beams and / or RS resources and beam parameters.

[0067] In one example, the WTRU may be configured with a first set of RS resources or beams or beam pairs that can cover the entire RS resource space, beam space, or beam pair space. The WTRU may determine or select sets A and B such that the union of sets A and B covers the entire RS resource space, beam space, or beam pair space. In one example, sets A and B may be mutually exclusive. In one example, set B includes RS resources, to which the WTRU may perform measurements to obtain direct measurements of the first set of beams or beam pairs (e.g., a one-to-one mapping between RS resources and beams or beam pairs) and estimated measurements of the second set of beams or beam pairs (e.g., a many-to-one mapping between RS resources and beams or beam pairs). For example, an AI / ML estimation model may be used to determine the estimated measurements.

[0068] The WTRU can be configured with one or more sets of RS resources associated with each beam. For example, the WTRU can be configured with a first beam associated with two sets of RS resources: the first set includes a single RS resource, and the second set includes multiple RS resources. The WTRU can determine the measurements associated with the beam via direct measurements of the RS resources in the first set or via estimates obtained from measurements of the RS resources in the second set.

[0069] The WTRU can determine a set of measurements for RS resources (e.g., set B) such that for each beam for which measurements must be obtained (directly or via estimation), set B contains at least one of two sets of RS resources associated with that beam. Thereafter, set B can be used interchangeably with sets of RS resources, beams, beam pairs, beam RS resources, RS resources, and beam patterns. Thereafter, set A can be used interchangeably with sets of RS resources, beams, beam pairs, beam RS resources, RS resources, and beam patterns.

[0070] To implement AI / ML for beam management, one option is to typically perform AI / ML model training and inference only on the WTRU side or only on the base station side. Each has its advantages and disadvantages. For AI / ML model training and inference on the WTRU side, the WTRU performing the inference can have access to real-time and accurate channel coefficients. However, each WTRU may have an isolated view of the transmission beams from the network (NW), at least because CSI-RS is traditionally configured in a WTRU-specific manner, resulting in a significant increase in data collection time. For AI / ML model training and inference on the base station side, the NW can have access to computational and storage resources, allowing for more generalized and comprehensive training on the base station side, and enabling the NW to perform complex computations for inference. However, the NW performing the inference may not have access to real-time and accurate channel coefficients.

[0071] Another possibility is that the AI / ML model can be trained at the base station and then transferred to the WTRU for inference. However, it is generally agreed in 3GPP RAN that auxiliary information from the NW may be unavailable due to security concerns, making cell-specific or site-specific models logically difficult. Furthermore, any response from the NW can be used to infer network beamform information when the WTRU queries the NW to establish the validity of using the AI / ML model in the current cell.

[0072] The embodiments described herein provide WTRUs, base stations, and corresponding methods that can use one or more WTRUs to efficiently perform inference and / or training of AI / ML models at a base station (e.g., gNB) for optimal beam prediction.

[0073] In one or more embodiments, the WTRU may receive or be configured or pre-configured with one or more general input and / or output beam pattern types (BPTs), which can be used to indicate beam groups (e.g., subsets of all beams) and beam group-based AI / ML models (e.g., via Radio Resource Control (RRC) signaling). The WTRU may also receive or be configured or pre-configured with weights and / or coefficients corresponding to the received BPTs.

[0074] In at least some embodiments, the WTRU can receive a first set of BPTs (e.g., input BPTs) and a second set of BPTs (e.g., output BPTs). For example, the WTRU can receive, be configured, or pre-configured with one or more BPTs, which can be used to indicate one or more subsets and beamgroup-based AI / ML models. The WTRU can receive cell-common BPTs and corresponding training weights, coefficients, beam indices, etc., via System Information Block (SIB), broadcast messages, multicast messages to a group of WTRUs, RRC signaling, MAC-CE, and / or DCI. Alternatively, the WTRU can receive WTRU-specific BPTs and corresponding training weights, coefficients, beam indices, etc., via RRC, MAC-CE, and / or DCI.

[0075] Figure 2 Figure 200 shows an example of an input BPT. The input BPT can be an index to an input beam pattern pool, which the WTRU can use as input to, for example, an AI / ML system. Figure 2In the example shown, the first input BPT 202 includes the beam to be measured (e.g., a subset of set B beams) of four beams 1a, 6a, 11a, and 16a in a 4×4 beam pattern comprising 16 potential beams 1a, 2a, 3a, 4a, 5a, 6a, 7a, 8a, 9a, 10a, 11a, 12a, 13a, 14a, 15a, and 16a. The second input BPT 204 includes the beam to be measured (e.g., a subset of set B beams) of three beams 1b, 5b, and 9b in a 3×3 beam pattern comprising 9 potential beams 1b, 2b, 3b, 4b, 5b, 6b, 7b, 8b, and 9b. The third input BPT 206 includes the beam to be measured (e.g., a subset of the set B beams) of a pattern consisting of four beams 3c, 9c, 11c, and 15c from a 3×5 beam pattern comprising 15 potential beams 1c, 2c, 3c, 4c, 5c, 6c, 7c, 8c, 9c, 10c, 11c, 12c, 13c, 14c, and 15c. The fourth input BPT 208 includes the beam to be measured (e.g., a subset of the set B beams) of a pattern consisting of four beams 4d, 6d, 10d, and 12d from a 3×5 beam pattern comprising 15 potential beams 1d, 2d, 3d, 4d, 5d, 6d, 7d, 8d, 9d, 10d, 11d, 12d, 13d, 14d, and 15d.

[0076] exist Figure 2Examples of input beam patterns identified by corresponding input BPTs are provided. For example, in input BPT 202, the first beam resource set is indicated as a grid-based 4×4 square of beam resources 1a, 2a, 3a, 4a, 5a, 6a, 7a, 8a, 9a, 10a, 11a, 12a, 13a, 14a, 15a, and 16a, and the second beam resource set is shown as a subset of the first beam resource set, 1a, 6a, 11a, and 16a. In another example, in input BPT 204, the first beam resource set is shown as a grid-based 3×3 square of beam resources 1b, 2b, 3b, 4b, 5b, 6b, 7b, 8b, and 9b, and the second beam resource set is shown as a subset of the first beam resource set, 1b, 5b, and 9b. In another example, in input BPT 206, the first beam resource set is represented as a grid-based 3×5 rectangle of beam resources 1c, 2c, 3c, 4c, 5c, 6c, 7c, 8c, 9c, 10c, 11c, 12c, 13c, 14c, and 15c, wherein the second beam resource set is represented as a subset of the first beam resources 3c, 9c, 11c, and 15c. In another example, in input BPT 208, the first beam resource set is represented as a grid-based 3×b rectangle of beam resources 1d, 2d, 3d, 4d, 5d, 6d, 7d, 8d, 9d, 10d, 11d, 12d, 13d, 14d, and 15d, wherein the second beam resource set is represented as a subset of the first beam resource set 4d, 6d, 10d, and 12d.

[0077] For example, the WTRU can be configured with an input BPT, and the configured input BPT can indicate a corresponding beam pattern, which may include first and second beam resource sets. The first beam resource set may cover the entire beam space or a subset of the beam space. The second beam resource set may be a subset of the first beam resources. The WTRU can use the second beam resource set to perform measurements. The second beam resource set may be a subset of one or more configured or pre-configured set B beam resources. That is, the WTRU can receive configuration information of one or more beam resources included in the second beam resource set to receive and measure one or more parameters (e.g., RSRP, Channel Quality Index (CQI), Signal-to-Interference-Noise Ratio (SINR), Physical Downlink Control Channel (PDCCH) Hypothetical Block Error Rate (BLER), etc.).

[0078] Figure 3 This is chart 300, showing an example of an output BPT. The output BPT can be an index of a model pool (e.g., an AI / ML model). Figure 3In the example shown, the first output BPT 302 includes AI / ML model coefficients for the patterns of four beams 2e, 3e, 6e, and 7e in a 4×2 beam pattern of beams 1e, 2e, 3e, 4e, 5e, 6e, 7e, and 8e. The second output BPT 304 includes AI / ML model coefficients for the patterns of four beams 1f, 4f, 5f, and 8f in a 4×2 beam pattern of beams 1f, 2f, 3f, 4f, 5f, 6f, 7f, and 8f. The third output BPT 306 includes AI / ML model coefficients for the patterns of four beams 1g, 2g, 3g, and 4g in a 2×2 beam pattern of beams 1g, 2g, 3g, and 4g. The fourth output BPT 308 includes AI / ML model coefficients for the patterns of two beams 2h and 3h in a 2×2 beam pattern of beams 1h, 2h, 3h, and 4h.

[0079] exist Figure 3 Examples of output beam patterns identified by corresponding output BPTs are provided. For instance, in output BPT 302, the first beam resource set is represented as a grid-based 4×2 rectangle of beam resources 1e, 2e, 3e, 4e, 5e, 6e, 7e, and 8e, where the second beam resource set is represented as a subset of the first beam resource set, 2e, 3e, 6e, and 7e. In another example, in output BPT 304, the first beam resource set is represented as a grid-based 4×2 rectangle of beam resources 1f, 2f, 3f, 4f, 5f, 6f, 7f, and 8f, where the second beam resource set is represented as a subset of the first beam resource set, 1f, 4f, 5f, and 8f. In another example, in output BPT306, the first beam resource set is indicated as a grid-based 2×2 square of beam resources 1g, 2g, 3g, and 4g, wherein the second beam resource set is shown as a subset of the first beam resource set, 1g, 2g, 3g, and 4g. In another example, in output BPT 308, the first beam resource set is indicated as a grid-based 2×2 square of beam resources 1h, 2h, 3h, and 4h, wherein the second beam resource set is shown as a subset of the first beam resource set, 2h and 3h.

[0080] For example, the WTRU can be configured with an output BPT, wherein the configured output BPT can indicate a corresponding beam pattern, which can include first and second beam resource sets. The first beam resource set can cover the entire beam space or a subset of the beam space. The second beam resource set can be a subset of the first beam resource set.

[0081] As another example, WTRU can use a configured BPT as an indicator to select from one or more configured or pre-configured AI / ML models and corresponding trained weights and / or coefficients for inference based on one or more input beam resources and / or input BPTs. In this way, WTRU can perform inference and beam prediction using a configured or pre-configured AI / ML model corresponding to the configured output BPT. By using the AI / ML model corresponding to the configured output BPT, WTRU can determine that the inferred output may be one of the beam resources from a second set of beam resources.

[0082] A WTRU can use, receive, and / or configure one or more AI / ML models (e.g., for beam prediction), where each AI / ML model can be associated with a pair of input and output BPTs. The AI / ML models can be generalized models that can be used to perform inference at different WTRUs with different scenes, different input beam indices, different output beam indices, etc.

[0083] The WTRU can store received and / or configured AI / ML models. For example, the WTRU can maintain a lookup table (AI / ML model cache) in its memory based on input-output BPT pairs containing recently used AI / ML models. When new AI / ML models are received, they can be added to this lookup table. When more memory is needed at the WTRU, the oldest AI / ML model can be removed to free up memory. Thus, when a new pair of input and output BPTs is received from the base station, the WTRU can first try to look up the associated AI / ML model from the cache. If it is available, the WTRU may move it to the top of the list of recently used AI / ML models. If it is not available, the WTRU can send a request to the base station to receive the AI / ML model.

[0084] The WTRU can determine and / or select a first configured or pre-configured AI / ML model for inference based on the association between the first AI / ML model and the output BPT of the first indication. The WTRU can determine and / or select a second configured or pre-configured AI / ML model for inference based on the association between the second AI / ML model and the output BPT of the second indication.

[0085] The WTRU can receive, be configured, or pre-configured with trained weights and / or coefficients for each AI / ML model. The WTRU can receive, be configured, or pre-configured with different types of general input and output BPTs and their corresponding associations with the AI / ML model. For example, during initial cell access and / or after cell connection, and in connected mode, the WTRU can receive configurations regarding the AI / ML model, input BPTs, and / or output BPTs. The WTRU can receive these configurations via cell-specific and / or WTRU-specific signaling.

[0086] The WTRU can receive (e.g., via RRC, MAC-CE, DCI from a base station) an indication to use a configured or pre-configured AI / ML model (e.g., for beam prediction), wherein the indication may include one or more pieces of information about BPTs and corresponding beam indices (e.g., beam positions of the beam grid). The indication may include one or more of input BPTs, output BPTs, input BPT beam indices, and / or output BPT beam indices. For example, the WTRU can receive an indication of input BPTs (e.g., input BPT 1, input BPT 2). The WTRU can use the indicated input BPTs to determine a subset of the set B required for beam measurements. As another example, the WTRU can receive an indication of output BPTs (e.g., output BPT1, output BPT 2). The WTRU can use the indicated output BPTs to determine the AI / ML model for inference (e.g., for beam prediction). As yet another example, the WTRU can receive an index number corresponding to a beam resource, which the WTRU can use as an input to the AIML model for inference. As another example, the WTRU can receive beam indices corresponding to the beam resources available to the WTRU, as well as the indicated output BPT. For instance, the WTRU can determine one or more of the indicated output BPT beam indices as the output of an AI / ML model, and as the result of inferences based on the associated AI / ML model.

[0087] Base stations (such as gNBs) can make inferences based on information received from the base station or information otherwise configured in the training models on the WTRU and gNB sides. This approach can be referred to as the hybrid inference method in this paper.

[0088] Figure 4A flowchart 400 illustrates an example method for beam prediction. This method can be implemented in a WTRU, as described above, which can be configured with input beam patterns, output beam patterns, artificial intelligence / machine learning (AI / ML) models, and trained weights for each AI / ML model. Each input beam pattern may include a beam index corresponding to a set of input beams. Each output beam pattern may include a beam index corresponding to a set of output beams. Each AI / ML model may correspond to a specific output beam. Additionally, in at least some embodiments, the WTRU may also be configured to receive, measure, and report one or more beams (e.g., beams from a first set B (e.g., CSI-RS resources)) based on a configured set A of beams. The WTRU can measure and report up to N beams from the received beams (e.g., the highest measured reference signal received power (RSRP)).

[0089] exist Figure 4 In the example shown, the WTRU can measure the signal quality of a subset of beams for each input beam pattern to identify at least one optimal input beam with the highest signal quality (402). The WTRU can report at least one optimal input beam to the base station (404). The WTRU can receive an indication of one of the input beam patterns and an indication of one of the output beam patterns from the base station to predict the optimal output beam based on the report (406). The WTRU can predict the optimal output beam (408). The WTRU can predict the optimal output beam using an AI / ML model corresponding to one of the indicated output beam patterns and trained weights corresponding to the AI / ML model. The WTRU can report the predicted optimal beam to the base station (410).

[0090] When WTRU is in the manner described above (e.g., refer to...) Figure 4 When performing inference, it can be called hybrid inference because some inference steps occur at the base station and some at the WTRU. For example, as mentioned above, the WTRU can perform hybrid inference based on the AI / ML system and for beam prediction. For example, a first inference step can be performed at the base station based on measurements received from the WTRU on a first set of set B beams. The base station can send the output results from the first inference step to the WTRU, which includes one or more input BPTs, output BPTs, beam indices, etc. The WTRU can use the received input BPTs to determine a second set of set B beams. The WTRU can use the received output BPTs to determine the AI / ML model. The WTRU can perform a second inference step based on the determined AI / ML model and the second set of set B beams, and predict the optimal beam accordingly.

[0091] In a hybrid inference embodiment, the base station can perform inference based on information received from the WTRU and the base station-side trained model. For example, the base station can predict the optimal output beam, determine the output beam pattern type (BPT), and determine the corresponding beam index. The predicted optimal beam resides in the determined output BPT. As another example, the base station can predict and / or determine the optimal input BPT associated with the predicted optimal beam and the determined output beam pattern. The base station can determine the output BPT and corresponding information for a given set of scenarios (e.g., characterized by auxiliary information received from the WTRU). In this way, the WTRU can perform inference using the determined input and output BPTs and the corresponding beam index, and predict the optimal beam based on WTRU-side channel state information (CSI) measurements.

[0092] Figure 5A This is a system diagram of an example wireless network 500 configured to perform hybrid inference for beam management. Figure 5A In the example shown, the wireless network includes a first WTRU 514 and a second WTRU 516, both of which communicate with a base station 518. In some embodiments, the base station may be a gNB. More specifically, the first WTRU 514 and the second WTRU 516 are configured to communicate with a radio access network (RAN) via base station 518. Figure 5A In the example shown, the first WTRU 514 and the second WTRU 516 exchange messages 502, 504, 506 and 508, 510, 512, respectively, as described in more detail below with reference to the example hybrid inference procedure.

[0093] Figure 5B Is Figure 5A A diagram of an example beam pattern 550 at base station 518. Figure 5B In the example shown, 64 beams are available at base station 550 and are arranged in input and output beam patterns. For example, Figure 5B The input beam patterns 554 and 556 are shown. Input beam pattern 554 includes beams 1, 2, 3, 4, 17, 18, 19, 20, 33, 34, 35, 36, 49, 50, 51, and 52, and input beam pattern 556 includes beams 26, 27, 28, 42, 43, 44, 58, 59, and 60. As another example, output beam patterns 552 and 558 are shown in... Figure 5B As shown in the figure. Output beam pattern 552 includes beams 2, 3, 18, 19, 34, 35, 50, and 51, and output beam pattern 558 includes beams 43, 44, 59, and 60. Each beam has a corresponding unique beam index (not shown).

[0094] As described above, each WTRU 514 and 516 is configured with multiple input beam patterns and multiple output beam patterns, and a subset of beams in each input beam pattern is measured to identify the optimal beam (although...). Figure 5B Only two input beam patterns and two output beam patterns are shown in the image. Figure 5B In the example shown, the measured subset of beams (or set B beams) includes beams 1, 5, 9, 13, 18, 22, 26, 30, 35, 39, 43, 47, 52, 56, 60, and 64.

[0095] Return to reference Figure 5A The first WTRU 514 can measure the set B beam and send a report 502 indicating the optimal measurement beam to the base station 518. Figure 5A and Figure 5B In the example shown, WTRU 514 measures set B beams, identifies beam 18 as the optimal beam, and sends a report 502 to base station 518 indicating that beam 18 is the optimal measurement beam. Similarly, the second WTRU 516 can measure set B beams and send a report 508 to base station 518 indicating the optimal measurement beam. Both WTRUs can also report one or more side information / auxiliary information (e.g., WTRU location, WTRU orientation, WTRU azimuth, etc.).

[0096] The first WTRU 514 can receive message 504 from the base station 518. Message 504 may include an indication to infer using the output beam pattern 552 and beam indices (2, 19, 34, and 51), and an indication to infer using the input beam pattern 554 with beam indices (1, 18, 35, and 52). Similarly, the second WTRU 516 can receive message 510 from the base station 518. Message 510 may include an indication to infer using the output beam pattern 558 and beam indices (44 and 59), and an indication to infer using the input beam pattern 556 with beam indices (26, 43, and 60). As described in more detail above, the first and second WTRUs 514 and 516 may also receive one or more configurations from messages 504 and 510. This approach may be superior to conventional methods because it may not be necessary to transmit beam shape models or detailed information, which could help the base station determine the optimal model.

[0097] First and second WTRUs 514 and 516 can perform inference based on received input and output BPTs, the configuration or pre-configured trained weights and / or coefficients of the corresponding input and / or output BPTs, and measurement parameters of beams from the second set of B beams, and predict the optimal beam accordingly. First WTRU 514 can send message 506 to base station 518, which may include an indication of the predicted optimal beam for the first WTRU. Similarly, WTRU 516 can send message 512 to base station 518, which may include an indication of the predicted optimal beam for the second WTRU. Figure 5A and Figure 5B In the example shown, the first WTRU 514 predicts beam 19 as the predicted optimal beam and indicates it to the base station 514 in message 506. The second WTRU 516 predicts beam 44 as the predicted optimal beam and indicates it to the base station 514 in message 512. In some embodiments, based on inference performed at base station 518, base station 518 may determine that the beam measured and reported as the optimal beam by the WTRU is actually the optimal beam of the WTRU. In this case, the WTRU may receive an indication that no inference is required on the WTRU side (e.g., BPT 0) and that the measured beam is the optimal beam. The WTRU can use the reported optimal beam to receive DL signaling or transmit UL signaling.

[0098] In some embodiments, the WTRU may receive one or more hybrid inference configurations. These configurations may be used to receive, measure, and report one or more beams, and may include, for example, AI / ML model-related information, set A beams (e.g., for base station and / or WTRU beam prediction), set B beams, and reporting configurations.

[0099] AI / ML model-related information may include at least one of the following: multiple beams (e.g., for set A and / or set B), multiple beam patterns, beam pattern types (e.g., input and / or output), AI / ML models, AI / ML model types, AI / ML model / weight sizes, and / or trained AI / ML models / weights. As described above, the WTRU may receive one or more trained AI / ML models / weights, or may determine trained AI / ML models / weights for beam prediction (e.g., based on base station configuration and / or indications).

[0100] The B-beam set can be used for WTRU measurements, and in some embodiments, it may be accompanied by one or more CSI-RS resources and one or more parameters. Such parameters may include, for example, CSI-RS / SSB ID, CSI-RS / SSB periodicity and slot offset (e.g., for periodic and semi-persistent CSI-RS resources), a CSI-RS / SSB resource map defining the number of CSI-RS ports, density, CDM type, OFDM symbol, subcarrier occupancy, the bandwidth portion to which the configured CSI-RS / SSB is allocated, references to TCI states including one or more QCL source RSs and corresponding one or more QCL types, CSI-RS / SSB transmission periodicity for periodic and semi-persistent CSI reporting, CSI-RS / SSB transmission slot offset for periodic, semi-persistent, and aperiodic CSI reporting, a list of CSI-RS / SSB transmission slot offsets for semi-persistent and aperiodic CSI reporting, and / or time constraints for channel and interference measurements.

[0101] One or more reporting configurations may be or include one or more of the following: operating mode, CSI reporting periodicity for periodic and semi-permanent CSI reporting, CSI reporting slot offset for periodic, semi-permanent, and non-permanent CSI reporting, a list of CSI reporting slot offsets for semi-permanent and non-permanent CSI reporting, reporting band configuration (wideband / subband CQI, PMI, etc.), thresholds and calculation modes for reporting quantities (CRI, CQI, RSRP, SINR, LI, RI, etc.), codebook configuration, group-based beam reporting, CQI table, subband size, and / or non-PMI port indication. In some embodiments, the WTRU may receive a hybrid inferred operating mode configuration, and in some embodiments, the hybrid inferred operating mode may be used when CRI is configured as one of the reporting quantities.

[0102] In some embodiments, the WTRU can measure beam-related information and report it to the base station. This measurement can be based on a configured set B beam (e.g., by measuring the SSB / CSI-RS resources configured for the set B beam). Based on this measurement, the WTRU can indicate up to N beams (e.g., via CRI) from the measured beam (e.g., the beam with the highest measured RSRP). In some embodiments, the WTRU can report one or more auxiliary information, which may be or include one or more of, for example, WTRU location, WTRU orientation, or WTRU azimuth.

[0103] In some embodiments, the base station may perform first beam prediction based on the received information and the base station-side training model. For example, the base station may predict one or more best beams and / or best beam patterns for WTRU operation. The base station may determine one or more beam pattern types and / or best beam indices based on the predicted one or more best beams and / or the predicted best beam pattern. For example, if the predicted best beam is included in the beam pattern type and / or beam pattern, the base station may determine the best beam pattern and / or best beam pattern type. For example, the gNB may determine the best input and best output BPT.

[0104] In some embodiments, the base station may determine one or more beam pattern types and / or corresponding information for a given set of scenarios. For example, the base station may determine one or more beam pattern types (e.g., input and output BPT), beam patterns, and corresponding information for each piece of assistance information received from the WTRU. Based on this determination, the WTRU may receive an indication of one or more of the determined beam pattern types (e.g., input and output BPT), beam patterns, and corresponding information. The indication may be a range for each piece of assistance information and / or a set of assistance information.

[0105] For example, if the position of the WTRU is between the first and second position thresholds (e.g., X11 < WTRU position < X12), if the direction of the WTRU is between the first and second direction thresholds (e.g., Y11 < WTRU direction < Y12), if the azimuth of the WTRU is between the first and second direction thresholds (Z11 < WTRU azimuth < Z12), etc., then the base station may determine that the WTRU is within the first range. Alternatively, if the position of the WTRU is between the third and fourth position thresholds (e.g., X21 < WTRU position < X22), if the direction of the WTRU is between the third and fourth direction thresholds (e.g., Y21 < WTRU direction < Y22), if the azimuth of the WTRU is between the third and fourth azimuth thresholds (e.g., Z21 < WTRU azimuth < Z22), etc., then the base station may determine that the WTRU is within the second range. The WTRU may receive configuration information about one or more ranges and thresholds of the position, direction, azimuth, etc. of the WTRU from the base station via one or more of RRC, MAC CE, and / or DCI.

[0106] Figure 6 Diagram of example AI / ML models 602a, 602b for hybrid inference on the base station side. As Figure 6 shown, the example AI / ML models 602a, 602b may be used by the base station 518 (see Figure 5A ) to predict the input and output BPT, as described above in reference to Figure 5A and Figure 5B described.

[0107] exist Figure 6 In the example shown, AI / ML model 602a receives input 604a, and AI / ML model 602b receives input 604b. Inputs 604a and 604b can be the optimal beam reported to the base station based on WTRU measurements. (Return to reference) Figure 5A In the example shown, input 604a is the best beam 18 reported to base station 518 in report 502 from first WTRU 514, and input 604b is the best beam 60 reported to base station 518 in report 508 from second WTRU 516. While first WTRU 514 reports only one beam (beam 18) as the best beam, and second WTRU 516 reports only one beam (beam 60) as the best beam, in some embodiments, the WTRU may report one or more best beams (e.g., those with the highest RSRP) and the corresponding measured RSRP. In some embodiments, inputs 604a and 604b may also include side information that the base station may have received from first WTRU 514 and / or second WTRU 516.

[0108] The base station can perform inference by using one or more best-reported measurement beams as inputs to AI / ML models 604a, 604b. The outputs 606a and 608a from AI / ML model 602a and the outputs 604b and 608b from AI / ML model 602b can be or include the best-predicted output BPT corresponding to the WTRU and the corresponding beam index, as well as the best-predicted input BPT and the corresponding beam index, all of which can be used for inference at the WTRU. For example, returning to the reference... Figure 5A AI / ML model 602a provides an input beam pattern 610 and beam indices (1, 18, 35, 52) as a first output 606a, and provides an output beam pattern 612 and beam indices (2, 19, 34, 51) as a second output 608a. AI / ML model 602b provides an output beam pattern 614 and beam indices (26, 43, 60) as a first output 606b, and an output beam pattern 616 and beam indices (44, 59) as a second output 608b. AI / ML models 602a and 602b can be different AI / ML models implemented in base station 518 or the same AI / ML model.

[0109] Figure 7 It is used for hybrid inference in Figure 5A A diagram of example AI / ML models 702a and 702b on the WTRU side of the first WTRU 514 and the second WTRU 516. Figure 7As shown, example AI / ML models 702a and 702b can be constructed from a first WTRU 514 and a second WTRU 516, respectively (see...). Figure 5A Used to predict the overall optimal beam, as mentioned above. Figure 5A and Figure 5B As stated above.

[0110] exist Figure 7 In the example shown, the first WTRU 514 and the second WTRU 516 (see...) Figure 5A The beam indices of base station 518, as indicated in messages 504 and 510, are used as inputs to AI / ML models 702a and 702b, respectively. Recall that base station 518 is able to use... Figure 6 The AI / ML models 602a and 602b shown determine which beam indices. The first and second WTRUs 514 and 516 can each select an AI / ML model for prediction from a larger group of AI / ML models configured in each WTRU. The WTRU can use output beam pattern information received in messages 504 and 506 from base station 518 to select an AI / ML model for prediction. In some embodiments, each AI / ML model configured in the WTRU corresponds to one of the output beam patterns configured in the WTRU, or a pair of one of the output beam patterns configured in the WTRU and one of the input beam patterns configured in the WTRU. The WTRU can select a configured AI / ML corresponding to the received output beam pattern information or a pair of received output beam pattern information and input beam pattern information. Figure 7 In the example shown, AI / ML model 702a corresponds to a pair of inputs BPT710 and outputs BPT712 (which can correspond to...). Figure 6 The AI / ML model 702b corresponds to a pair of input BPT 716 and output BPT 718 (which can correspond to input BPT 610 and output BPT 612). Figure 6 The input BPT 614 and output BPT 616 are shown in the figure.

[0111] exist Figure 7In the example shown, AI / ML model 702a receives inputs 700a, 700b, 700c, and 700d, and AI / ML model 702b receives inputs 701a, 701b, and 701c. The inputs 700a, 700b, 700c, and 700d of AI / ML model 702a at the first WTRU 514 can be beam indices (1, 18, 35, 52) of the input beam pattern 710 indicated to the first WTRU 514 in message 504. The inputs 701a, 701b, and 701c of AI / ML model 702b at the second WTRU 516 can be beam indices (26, 43, 60) of the input beam pattern 716 indicated to the first WTRU 514 in message 510.

[0112] The first WTRU 514 can perform inference by using the beam index of the input beam pattern indicated by the base station as inputs 700a, 700b, 700c, and 700d of the AI / ML model 702a. The second WTRU 516 can similarly perform inference by using the beam index of the input beam pattern indicated by the base station as inputs 701a, 701b, and 701c of the AI / ML model 702b. Based on measurements at the WTRU, initial predictions at the base station, and second predictions at the WTRU, the output 704 from the AI / ML model 702a and the output 708 from the AI / ML model 702b can be or include the beam index of the optimal overall beam for a particular WTRU. For example, refer to [reference]. Figure 5A The AI / ML model 702a at the first WTRU 514 provides beam index 19 as output 704, and the AI / ML model 702b at the second WTRU 516 provides beam index 44 as output 708. Each WTRU can then use the output beam index for transmission or reception, as described elsewhere herein, for example.

[0113] In some embodiments, the WTRU can determine an AI / ML model and / or weights for WTRU side beam prediction. For example, this determination may be based on one or more of a received model ID, associated information, beam pattern type, and / or beam pattern. The WTRU may use an AI / ML model and / or weights associated with the received model ID (e.g., indicated via one or more of RRC, MACCE, or DCI). The WTRU may use an AI / ML model and / or weights associated with the current WTRU position, WTRU velocity, WTRU orientation, WTRU azimuth, etc. (e.g., based on indicated range). The WTRU may use an AI / ML model and / or weights associated with the received, configured, and / or indicated input and output beam pattern types and / or beam patterns.

[0114] The WTRU can perform a second prediction on the WTRU side based on one or more of the following: measurements of a second set of B-beams (e.g., the measured quality of the second set (e.g., RSRP)), determined, configured, and / or indicated input and / or output beam pattern types, determined, configured, and / or indicated beam indices, determined, configured, and / or indicated association information, and / or determined, configured, and / or indicated AI / ML models and / or weights. For example, in Figure 7 In the example shown, as described above, the first WTRU 514 predicts beam 19 as the optimal beam, and the second WTRU 516 predicts beam 44. The WTRU can (e.g., via CRI) report one or more of the predicted optimal beams.

[0115] In some embodiments, the WTRU may receive an indication that a second inference and / or corresponding report is not required. For example, the WTRU may receive such an indication if the optimal beam reported based on measurements of the first set B beams is the actual optimal beam (e.g., based on a first inference from the base station side). Based on this indication, the WTRU may deactivate the determination / measurement of the second set B beams, the prediction / inference of one or more optimal beams from the second set, and / or the indication of one or more of the one or more optimal beams from the second set.

[0116] For example, the indication can be an explicit indication. For instance, the WTRU can receive an indication with a first value (e.g., value zero) that indicates no second beam prediction is required. In another example, the WTRU can receive an indication with a second value (e.g., value one) that indicates the need for second beam prediction. As yet another example, the WTRU can receive an indication of a first value (e.g., value 0 and / or NACK) on the WTRU's reporting beam and an indication of a second value (e.g., value 1 or ACK) on the WTRU's reporting beam.

[0117] As another example, the indication can be implicit. For instance, the WTRU can receive the indication via beam pattern type and / or beam pattern. For example, input BPT 0 or output BPT 0 could indicate an ACK for the beam reported by the WTRU and / or lack of support for second beam prediction on the WTRU side.

[0118] Based on one or more reported optimal beams, the WTRU can receive one or more DL channels / signals and / or transmit one or more UL channels / signals.

[0119] Additionally or alternatively, after the WTRU performs a second inference on its side and reports the predicted optimal beam, the WTRU may perform an accuracy measurement. If the WTRU performs an accuracy measurement based on the first predicted optimal beam, the WTRU may measure one or more accuracy parameters from the base station on one or more receive channels / signals. One or more accuracy parameters may be, for example, RSRP and / or the assumed BLER. The WTRU may determine that the measured accuracy parameters are not within the acceptable range for the first predicted optimal beam, for example due to false detections during inference and beam prediction, environmental changes, or WTRU mobility. The WTRU may send an indication that it needs more information about the AI / ML model and requests further information. The WTRU may indicate invalid accuracy (e.g., flag indication) and one or more measurement parameters of the first predicted beam as supplementary information to the base station. During hybrid inference, the WTRU may similarly perform an accuracy measurement on the measured optimal beam. For example, if the WTRU performs an accuracy measurement on a periodic set B, the WTRU may receive a configuration (e.g., time period) to measure all beams in set B (e.g., a long time period). The WTRU can measure and report up to N beams from the received beam (e.g., the highest measured RSRP). The WTRU can (e.g., based on inference at the base station) receive instructions from the base station to schedule the reception of further information.

[0120] The WTRU can receive information from the base station (e.g., in response to a request for additional information from the WTRU). In the example of a single beam transfer, the WTRU can receive an AI / ML model transfer for a single beam (e.g., transferring trained weight coefficients for beam #3 of the first WTRU). In the example of an updated model transfer, the WTRU can receive an indication of the updated AI / ML model and the corresponding group-beam-based model for receiving the indicated first BPT (e.g., updated trained weight coefficients for BPT 2 of the first WTRU). In the example of a second BPT indication, the WTRU can receive an indication using the second BPT and the corresponding configuration (e.g., the second BPT for the first WTRU could be BPT 3 for beam indices (2, 3, 18, 19)).

[0121] The WTRU can predict a second optimal predicted beam (e.g., beam 3) based on the received information and report the second optimal predicted beam to the base station. If the measured accuracy parameters are within an acceptable range, the WTRU can indicate that it has selected the predicted beam as the optimal beam. The WTRU can use the predicted optimal beam for DL ​​reception (e.g., PDCCH, PDSCH, CSI-RS, etc.) and / or UL transmission (e.g., PRACH, PUCCH, SRS, etc.).

[0122] The WTRU can be configured to receive one or more first AI / ML models (e.g., input and / or output BPT) and corresponding configurations (e.g., from the base station). The WTRU can predict first beam resources (e.g., as the optimal beam) based on the first AI / ML models and can report the predicted first beam resources (e.g., to the base station).

[0123] In some embodiments, the WTRU can measure one or more parameters on the received signal and / or channel based on a first beam resource. Based on the first beam resource, the WTRU can be configured with one or more UL transmit and DL receive signals. Furthermore, the WTRU can be configured with one or more resource signals (RS) based on the first beam resource. For example, the WTRU can measure RSRP, SINR, RSRQ, CQI, PDCCH assumption BLER, etc.

[0124] If the WTRU determines that one or more of the measured parameters are outside the acceptable range, it can send an indication that it needs more information about the AI / ML model and may request further information. For example, if one or more of the measured RSRP, SINR, RSRQ, CQI, etc., are below the corresponding threshold, the WTRU can determine that the measured parameter is outside the acceptable range. As another example, if the measured assumed BLER is above the corresponding threshold, the WTRU can determine that the measured parameter is outside the acceptable range. The WTRU can report one or more of the measured parameters of the first beam as supplementary information.

[0125] In some embodiments, the WTRU can be configured to measure one or more parameters based on a first set of B-beams, wherein the WTRU can determine whether the measured parameters are within acceptable ranges. For example, the WTRU can be configured to measure the first set of B-beams over a time period (e.g., a long time period). For example, the WTRU can measure the first set of B-beams and report up to N beam resources (e.g., those with the highest measured RSRP). The WTRU can receive indications from the base station to schedule the reception of further information.

[0126] In some embodiments, the WTRU may receive further information about the AI / ML model. In the example of single-beam transfer, the WTRU may receive indication and / or scheduling information for receiving AI / ML model transfer for a single beam resource and / or beam pattern. For example, in the example of updated model transfer, the WTRU may receive indication and / or scheduling information (e.g., updated trained weight coefficients) for receiving an updated AI / ML model of the indicated AI / ML model. In the example of a second BPT indication, for example, the WTRU may receive an indication to use a second AI / ML model. These indications may include corresponding configurations.

[0127] The WTRU can predict a second optimal predicted beam based on received information and an updated AI / ML model. The WTRU can report the second predicted beam to the base station. The WTRU can also measure one or more parameters based on the second beam resources. If the measured accuracy parameters are within acceptable limits, the WTRU can indicate that it has selected the predicted beam as the optimal beam. The WTRU can use the predicted optimal beam for DL ​​reception (e.g., PDCCH, PDSCH, CSI-RS, etc.) and / or UL transmission (e.g., PRACH, PUCCH, SRS, etc.).

[0128] In some embodiments, the WTRU receives configuration information indicating one or more BPTs for inference on the WTRU side. The WTRU may receive configuration information about one or more subsets of the set of B beams for measurement for each BPT. Additionally or alternatively, the WTRU may receive instructions or configurations to perform inference and report the predicted beam for each configured BPT, wherein the purpose of the WTRU receiving the configured inference and beam reporting is for base station training. Alternatively or additionally, the WTRU may perform inference based on a configured subset of the set of B beams measured for each configured BPT. Additionally or alternatively, in addition to one or more auxiliary information (UE azimuth, location, etc.), the WTRU may report the predicted beam for each configured BPT of each configuration subset of the set of B beams. The WTRU may determine that the priority of the report transmission for base station training purposes is a low-priority UL compared to other configured and / or scheduled UL or DL ​​timings. The priority may be a function of the number of BPTs or the BPT category.

[0129] Base stations may need to be trained to enable hybrid inference. For example, a base station may need to be trained to predict, select, and / or indicate the best AI / ML model and corresponding configuration for a given set of scenarios based on reported information from WTRUs. This could be part of a broader data collection process conducted by the network, where different base stations / TRPs can collect data on one or more scenarios from one or more WTRUs to build a database that allows for accurate prediction of BPT to the WTRU for hybrid inference.

[0130] As part of the training process, the WTRU can be configured with or otherwise receive configurations. These configurations can be used to receive, measure, predict, and report one or more beam resources. For example, one or more of the following configurations can be used: BPT, one or more best serving beams (e.g., beam index), RSRP / RSRQ of one or more best beams, RSRP / RSRQ difference of one or more best beams, AoA / AoD of one or more best beams, beams with RSRP / RSRQ above a threshold, one or more worst beams (which may correspond to congestion), RSRP / RSRQ of one or more worst beams, RSRP / RSRQ difference of one or more worst beams, beams with RSRP / RSRQ below a threshold, congestion detected by WTRU (e.g., sudden drop in L1-RSRP, path loss measurement, calculated CSI parameters (e.g., low CQI), etc.), information about congestion that helps NW build its database (e.g., congestion duration, location, affected beam set, etc.), CSI parameters (e.g., CQI, PMI, RI) and / or any changes thereof, WTRU location / positioning / coordinates, WTRU relative location, mobility status, LOS / NLOS / path loss conditions, and / or timestamps. The aforementioned best or worst beam can refer to either the transmission beam or the receiving beam.

[0131] For example, the WTRU can be configured with one or more input and / or output beampts (BPTs). The WTRU can perform inference based on measurement beams from a subset of a set of B-beams, which is determined by the WTRU based on each of the configured input BPTs. The WTRU can select an AI / ML model for inference based on the configured output BPTs. In other words, the WTRU can perform inference and predict the optimal beam using the configured input BPTs and configured output BPTs one by one. In this way, the WTRU can determine one or more pairs of input and output BPTs that lead to the prediction of the optimal beam, and the WTRU can report these pairs to the base station.

[0132] For example, the WTRU can report the optimal beam and the BPT (Browser Predictor Target) that the WTRU uses to predict the optimal beam to the base station. This information can help the base station configure the component beams that make up part of each BPT. For example, the WTRU can report the RSRP (Reverse Target Probability) and / or RSRQ (Reverse Target ...

[0133] For example, for each configured input and / or output BPT, the WTRU can refer to the best predicted beam to report the differential RSRP and / or RSRQ of one or more predicted beams. The WTRU can report the beam index of the reported beam. In another example, the WTRU can report variations in the measured RSRP and / or RSRQ of the best beam based on different WTRU position, azimuth, etc. In yet another example, based on inferences based on different input and / or output BPTs, the WTRU can report variations in the measured RSRP and / or RSRQ of the best beam.

[0134] For example, the WTRU can report the AOA and / or AOD of the best predicted beam for each configuration's input and / or output BPT, for example, based on different scenarios such as the WTRU's location, orientation, etc. For example, the WTRU can report one or more beam resources for which the predicted RSRP and / or RSRQ are higher than a threshold for each configuration's input and / or output BPT (e.g., based on different scenarios such as the WTRU's location, orientation, etc.). For example, the WTRU can report the worst predicted serving beam for each configuration's input and / or output BPT. For example, by excluding the worst serving beam, reporting such information can help the base station configure the constituent beams that form part of each BPT.

[0135] For example, the WTRU can report the RSRP and / or RSRQ of the worst predicted beam for each configuration's input and / or output BPT. Alternatively, the WTRU can report the differential RSRP and / or RSRQ of the worst predicted beam for each configuration's input and / or output BPT. In one example, the WTRU can report the differential RSRP and / or RSRQ of the worst predicted beam for each configuration's input and / or output BPT based on the decrease in RSRP.

[0136] For example, for each configuration's input and / or output BPT, the WTRU can report predicted beams with measured RSRP and / or RSRQ below a threshold. For example, the WTRU can report detected or predicted blocking for each configuration's input and / or output BPT.

[0137] WTRU reports of optimal input and / or output BPT and WTRU locations can, for example, help the NW configure the fingerprint of the optimal BPT for each location. For instance, the WTRU can report its location, orientation, mobility status, etc. This information, along with the reported optimal beam, RSRP, RSRQ, and / or optimal input and output BPTs, allows the NW to construct a fingerprint map of signal strength at different locations and configure the BPTs accordingly (e.g., the number of beams in each output BPT, the number of ensemble B-beams in each input BPT, etc.).

[0138] The relative position of the WTRU can indicate, for example, the relative position of the WTRU with respect to a specific TRP and / or base station. The WTRU can send SRS to the TRP and / or base station, which can assess parameters such as propagation delay, path loss, signal strength, etc., to measure the relative position and / or radio conditions of the WTRU (e.g., the position of the WTRU relative to another WTRU). Mobility status can refer to the speed and direction of the moving WTRU.

[0139] The WTRU can be configured to report a timestamp corresponding to a measurement for any of the aforementioned parameters. In some cases, the WTRU can be configured to report data as soon as a measurement is performed. In other cases, since the data is for base station training purposes, the WTRU can be configured to collect data over time and report it immediately (e.g., via a data collection framework like Minimum Drive Test (MDT)). In this case, the timestamp can help the NW construct an accurate fingerprint of the optimal BPT for each location.

[0140] The WTRU can receive configuration information indicating one or more input and / or output beamforming points (BPTs) for inference at the WTRU side. In some embodiments, the WTRU can (e.g., from the network) receive instructions or configurations to perform inference and report the predicted beams of the input and / or output BPTs for each configuration. The WTRU can (e.g., from the NW) receive a list of configured input and / or output BPTs and report one or more predicted optimal beams and one or more associated input and / or output BPTs used to predict the optimal beams to the network.

[0141] For example, the WTRU can infer the configuration subset of the set of B-beams based on the measurements of the input BPT for each configuration. In addition to one or more auxiliary information (such as WTRU azimuth, position, etc. listed above), the WTRU can report the predicted beam for the output BPT for each configuration subset of the set of B-beams.

[0142] As another example, the WTRU may report the aforementioned data and / or auxiliary information via any one or more of the following frameworks: recorded MDT, instant MDT, RRC (e.g., RRC measurement report, UE auxiliary information), early idle and / or inactive measurement report, beam reporting framework, CSI reporting framework, LPP (LTE positioning protocol) for location / positioning information, etc.

[0143] As another example, the WTRU can receive instructions that the purpose of the measurement and report is for model training in the network. For example, the WTRU can receive this instruction after receiving a configured list of BPTs from the NW. For example, the WTRU can receive this instruction after receiving the resource for performing the measurement (e.g., a CSI-RS resource).

[0144] Knowing that measurements may be used for base station training, the WTRU can determine that the priority of report transmissions used for base station training is a lower priority UL timing compared to other configured and / or scheduled UL or DL ​​timings. In one example, the priority can be a function of the number of BPTs or the BPT category.

[0145] For example, the WTRU can be configured to determine the period at which measurements are performed. In another example, it can be event-triggered. For instance, the configuration of the base station receiving CSI-RS resources and / or resource sets, as well as CSI reports used for training AI / ML models, can trigger the WTRU to perform measurements.

[0146] As another example, the WTRU can perform measurements of the above parameters only when connected and report them back to the network (e.g., via instant MDT, CSI reporting frame, beam reporting frame, etc.). In another example, the WTRU can perform measurements in idle and / or inactive states and use different frames to report large amounts of data at once (e.g., via a recorded MDT frame).

[0147] In one example, the WTRU can assist base station training using conventional methods of measuring the best serving beam (e.g., the beam with the highest RSRP) and reporting it to the NW (possibly with other information such as the WTRU's location). For example, the WTRU can receive the configuration of CSI-RS resources and / or resource sets, along with CSI reports, for training AI / ML models at the base station for hybrid inference. The WTRU can perform periodic measurements to identify the best serving beam (e.g., the beam with the highest RSRP) and report (e.g., up to four) of the best beams to the base station. The WTRU can also send its location and each report of one or more of the best beams to the network, allowing the network to build a database of candidate BPTs and their constituent beams for each of its locations.

[0148] In another example, the WTRU can assist base station training with an AI / ML model at the WTRU. For instance, the AI / ML model at the WTRU can output the best BPT from a pre-configured set of BPTs from the NW under different scenarios (e.g., different WTRU locations, velocities, LOS / NLOS conditions), and the WTRU can report this best BPT to the NW to help the NW train.

[0149] In any of the examples above, this data collection process can be performed before deploying the corresponding model for inference to assist in model training within the network. Fine-tuning and / or retraining can also be performed periodically based on the addition of new scenarios and / or updates to existing scenarios.

[0150] To maintain the accuracy of AI / ML-based beam prediction, it may be important to regularly monitor beam prediction accuracy and improve it or take alternative measures when necessary. One factor that may lead to a loss of beam prediction accuracy is the use of outdated AI / ML models and / or input sets (e.g., the AI / ML model and selected inputs fail to capture / reflect the current behavior of the beam). AI / ML models can become outdated due to a number of factors, including changes in the environment (e.g., the movement of large objects in the environment or a WTRU previously moved from outdoors to indoors), WTRU movement (e.g., changes in the speed and direction of movement), WTRU rotation, and so on.

[0151] For example, consider the case where an AI / ML model and model input set (e.g., beam set) are selected when the WTRU is indoors. However, over time, the WTRU has moved and is now located in an outdoor environment. In this case, using the same AI / ML model and its input set, assuming an indoor propagation environment, may not provide accurate beam predictions because the environment has changed. To compensate for or manage this loss of beam prediction accuracy, appropriate steps may need to be taken (e.g., selecting an alternative AI / ML model, selecting an alternative input beam set, etc.).

[0152] Therefore, in the embodiments described herein, after the WTRU performs inference and predicts and reports the predicted optimal beam to the base station using, for example, the hybrid inference method described above, the WTRU can receive configuration information about one or more second candidate BPTs. For example, refer to Figure 5A and Figure 5B The first WTRU 514 can receive the third BPT ( Figure 5B(Not shown in the text) is used for beam indices (3, 4, 19, and 20) and / or for the fourth BPT as an indication of beam indices (3 and 18), as a candidate BPT. The WTRU can perform inference and predict one or more second beams based on the received candidate second BPTs or BPTs.

[0153] The WTRU can compare one or more measurement and / or prediction accuracy parameters (e.g., RSRP, assumed BLER, etc.) of a first beam with prediction accuracy parameters of a second prediction beam. The WTRU can receive a channel / signal from a base station based on a first (e.g., active) beam (e.g., beam 19), where the WTRU measures one or more accuracy parameters. Alternatively or additionally, the WTRU can predict one or more accuracy parameters of the first beam.

[0154] The WTRU can determine that one or more prediction accuracy parameters corresponding to at least one second prediction beam are better than the measured or predicted accuracy parameters of the first beam (e.g., beam 3). For example, the WTRU can determine that the difference between the predicted RSRP of the second beam and the measured / predicted RSRP of the first beam is greater than an RSRP threshold. As another example, the WTRU can determine that the difference between the prediction assumption BLER of the second beam and the measured / predicted assumption BLER of the first beam is greater than a BLER threshold. The WTRU can send a request to the base station to switch to the second BPT for further inference and prediction. The WTRU can send a request to the base station to switch to the second prediction beam. The WTRU can use the second prediction beam for DL ​​reception (e.g., PDCCH, PDSCH, CSI-RS, etc.) and / or UL transmission (e.g., PRACH, PUCCH, SRS, etc.).

[0155] For example, the WTRU can predict and report one or more first beams (e.g., the beam with the highest predicted RSRP) based on the hybrid inference method described above. In the example configuration, the WTRU can receive a first input and a first output BPT, as well as a corresponding configuration for beam prediction (e.g., a beam index). The WTRU can perform beam inference and predict the first or more beams (e.g., the beam with the highest predicted RSRP) based on the received first input BPT and the received first output BPT, as well as the corresponding configuration. The WTRU can report the predicted first beams to the base station (e.g., via UCI, MAC-CE, and / or RRC indication).

[0156] Refer again Figure 5A and Figure 5BThe first WTRU 514 can receive indications using an output beam pattern 552 with beam indices (2, 19, 34, and 51) and an input beam pattern 554 with beam indices (1, 18, 35, and 52). In this way, the WTRU can predict and report beam 19 as the first predicted beam (e.g., the beam with the highest predicted RSRP).

[0157] The WTRU can receive configuration information from one or more second candidate input BPTs and one or more second output BPTs that are different from the first configuration input and output BPTs. For example, in Figure 5B In the example configuration shown, the first WTRU 514A can receive output BPT3 using beam indices (3, 4, 19, and 20). Figure 5B (Unmarked) as the second candidate output BPT and the input BPT4 for beam indices (5, 22, 18, and 35) Figure 5B (Unmarked) serves as an indication of the second candidate input BPT. In another example, the first WTRU 514 can receive the output BPT4 (unmarked) using beam indices (3 and 18). Figure 5B (Unlabeled) as the second candidate output BPT and the input BPT2 with beam indices (1, 18, 35) Figure 5B (Unmarked) serves as an indication for the second candidate input BPT.

[0158] The WTRU can perform beam inference using one or more of the configured second candidate input and / or output BPTs and predict one or more second beams. For example, the first WTRU 5A can predict beam 3 as the optimal beam based on the configured second candidate input and / or output BPTs. The first WTRU 514 can compare one or more measured and / or predicted beam quality parameters (e.g., RSRP, assumed BLER, RSRQ, SINR) of the first or active beam with the predicted beam quality parameters of one or more second beams predicted using one or more second candidate BPTs.

[0159] The WTRU can use one or more channels (e.g., PDCCH) or signals (e.g., CSI-RS) received via the first beam to measure, determine, or estimate one or more channel / signal quality parameters associated with the first beam. For example, a first WTRU 514, which has selected and / or predicted beam 19 as the first and / or active beam, can use beam 19 to receive one or more channels and / or signals from the base station. The first WTRU 514 can use the received channels and / or signals received via beam 19 to determine the channel / signal quality parameters. The first WTRU 514 can predict one or more channel quality parameters for the first or active beam.

[0160] The WTRU can determine that one or more measured and / or predicted beam quality parameters associated with at least one second predicted beam are superior to the measured or predicted beam quality parameters of the first or active beam. For example, the WTRU can determine that the difference between the predicted RSRP of one of the second beams and the measured or predicted RSRP of the first beam is higher than a configured and / or pre-configured RSRP threshold. For example, the WTRU may receive or be configured and / or pre-configured with a corresponding threshold from the base station, for example via RRC signaling, MAC-CE, and / or DCI indication. As another example, the WTRU can determine that the difference between the predicted assumption BLER of a channel (e.g., PDCCH) associated with one of the second beams and the measured or predicted assumption BLER of the first beam is higher than a configured or pre-configured BLER threshold.

[0161] The WTRU can send reports to the base station indicating the predicted second beam and one or more predicted beam quality parameters. The WTRU can send these reports and / or indications via UCI, MAC-CE, and / or RRC signaling. For example, a report sent by the WTRU may include a request to switch to the second input and / or output BPT for further inference and prediction.

[0162] The WTRU can monitor and receive acknowledgments from the base station (e.g., via DCI, MAC-CE indication, or RRC signaling) that grant permission to switch to the second input and / or output BPT for further inference and prediction. After the configured and / or pre-configured BPT handover time (e.g., via RRC signaling, MAC-CE indication, or DCI indication configuration and / or pre-configuration), the WTRU can begin using the second input and / or output BPT for further inference and prediction.

[0163] The reports and / or indications that the WTRU can send may include requests to switch to the second predicted beam. The WTRU may send requests to the base station to switch to the second predicted beam or grant permission to switch to the second predicted beam for use in DL reception (e.g., PDCCH, PDSCH, CSI-RS, etc.) and / or UL transmission (e.g., PRACH, PUCCH, SRS, etc.). The WTRU may (e.g., from the gNB, within a configured and / or pre-configured monitoring window) monitor and receive acknowledgment indications for the use of the second beam. For example, the WTRU may monitor acknowledgment indications on the PDCCH (e.g., PDCCH received via a configured or pre-configured search space / CORESET) or MAC-CE indications or RRC signaling.

[0164] After the configured and / or pre-configured beam application time, the WTRU can begin using the predicted second beam for DL ​​reception (e.g., PDCCH, PDSCH, CSI-RS, etc.) and / or UL transmission (e.g., PRACH, PUCCH, SRS, etc.).

[0165] As mentioned above, WTRU (e.g., Figure 5A The first WTRU (514 or the second WTRU (516) shown) can be configured with information and related information about the AI / ML model in any of several different ways. In a particular example, the WTRU can be configured for and / or be able to be used for hybrid inference of beam management in an AI / ML system. The WTRU can receive configuration information about one or more Group Radio Network Temporary Identifiers (G-RNTIs) to monitor and receive DCIs of group common transfers (DCIs) based on beamgroups with trained weights and / or coefficients corresponding to one or more BPTs. For example, the WTRU may be connected to a cell and may need to receive all models used in the cell. The base station may change the model every hour (due to security concerns) and indicate the new model and / or BPT to the WTRU.

[0166] The WTRU can detect a DCI with CRC scrambling using G-RNTI based on one or more of the following: the WTRU receives a configuration of trained weights and / or coefficients requiring a first BPT (e.g., for hybrid inference), where the WTRU determines that it does not have the corresponding trained weights and / or coefficients; the WTRU receives an indication (e.g., paging, group common DCI, etc.) indicating that the trained weights and / or coefficients corresponding to the first BPT have been updated at the base station, and the WTRU needs to receive the group common transfer of the coefficients; the WTRU determines that the accuracy parameters of beam prediction based on the configured first BPT are invalid and / or unacceptable and / or the configured BPT has expired; and / or the WTRU determines that the accuracy parameters of beam prediction based on the configured first BPT are invalid and / or unacceptable.

[0167] When the WTRU receives a hybrid inference configuration requiring the first BPT (Boundary Point Test), and if the WTRU determines that it does not have the corresponding trained weights and / or coefficients, the WTRU can select a first G-RNTI associated with the first BPT. The WTRU can monitor to detect a DCI with CRC scrambled by the first G-RNTI to receive scheduled resources to receive the trained weights and / or coefficients corresponding to the first BPT.

[0168] When the WTRU receives an indication (e.g., paging, group common DCI, etc.) indicating that the trained weights and / or coefficients corresponding to the first BPT have been updated at the gNB and the WTRU needs to receive the group common transition of the coefficients, the WTRU may select a first G-RNTI associated with the first BPT. The WTRU may monitor to detect a DCI with CRC scrambled by the first G-RNTI to receive scheduled resources to receive the trained weights and / or coefficients corresponding to the first BPT.

[0169] When the WTRU determines that the accuracy parameters of beam prediction based on the configured first BPT are invalid and / or unacceptable and / or the configured BPT has expired, the WTRU can monitor to detect a DCI scrambled with a first G-RNTI using CRC, in order to receive scheduled resources to receive the trained weights and / or coefficients corresponding to the first BPT. When the WTRU determines that the accuracy parameters of beam prediction based on the configured first BPT are invalid and / or unacceptable, the WTRU can determine, request, or suggest a switch to a second BPT. The WTRU can select a second G-RNTI associated with the second BPT. The WTRU can monitor to detect a DCI with CRC scrambled using the second G-RNTI, in order to receive scheduled resources to receive the trained weights and / or coefficients corresponding to the second BPT.

[0170] The WTRU can use the received trained weights and / or coefficients corresponding to the configured BPT to infer and predict the optimal beam. The WTRU can report the predicted optimal beam and receive and / or transmit DL and / or UL based on the predicted optimal beam.

[0171] The WTRU can receive configuration information about one or more group indicators and / or identifiers to monitor and / or receive DCIs of one or more group common transfers (G-RNTIs) based on beamgroups corresponding to one or more AI / ML system models. For example, the WTRU can receive one or more group G-RNTIs to receive group common DCI signaling. The WTRU can be configured with a first G-RNTI associated with a first AI / ML system model, a second G-RNTI associated with a second AI / ML system model, and so on. For example, the WTRU can be configured with separate and different G-RNTIs for different AI / ML system models. For example, during initial cell access and / or after cell connection and in connected mode, the WTRU can receive configurations on the G-RNTIs via SIB, RRC, MAC-CE, or DCI.

[0172] Upon receiving a DCI scrambled with its CRC using a first G-RNTI, the WTRU can receive configured and / or dynamic grants via the DCI, indicating time and / or frequency resources for DL ​​grants. The WTRU can use the configured time and frequency resources (e.g., from the gNB) to receive, transfer, and / or download information including trained weights and / or coefficients corresponding to a first AI / ML system model associated with the indicated first G-RNTI. For example, if the WTRU has recently connected to the cell and needs to receive all AI / ML models used in that cell, the WTRU can determine to receive, download, and / or transfer this information. Alternatively, the WTRU can receive indications that the AI / ML model has been updated (e.g., from the base station) (e.g., because the base station updates the model based on a recent retraining, or because the base station changes the model every hour (e.g., due to security concerns)). In this way, the WTRU can determine to receive, download, and / or transfer information.

[0173] In some embodiments, the WTRU may receive one or more configuration information for operations (e.g., hybrid inference) based on one or more AI / ML models. For example, the WTRU may receive configuration to perform inference based on a first AI / ML model. The WTRU may determine that it does not have the corresponding trained weights, coefficients, and / or AI / ML model (e.g., in the WTRU's memory or storage device). In this case, the WTRU may select a first G-RNTI associated with the first AI / ML model. The WTRU may monitor to detect a DCI with CRC scrambled using the first G-RNTI to receive scheduled resources to receive trained weights and / or coefficients corresponding to the first AI / ML model.

[0174] In some embodiments, the WTRU may receive one or more indications, wherein the WTRU can determine that the first AI / ML model has been updated. For example, the WTRU may receive one or more indications from the base station via DCI, MAC-CE, and / or RRC, indicating that the first system model has been updated at the base station side, for example, due to training at the base station. Thus, the WTRU can select a first G-RNTI associated with the first AI / ML model. The WTRU may monitor to detect a DCI with CRC scrambled using the first G-RNTI to receive scheduled resources to receive trained weights and / or coefficients corresponding to the first AI / ML model.

[0175] In some embodiments, the WTRU can determine that one or more of the configured input and / or output BPTs have expired. For example, the WTRU can determine that the corresponding timer or counter has expired. As another example, the WTRU can determine the expiration of a BPT based on one or more events, such as beam failure detection, multiple consecutive retransmission requests from the base station, multiple consecutive NACKs, etc. In this way, the WTRU can monitor and detect the DCI with CRC scrambled using a first G-RNTI to receive scheduled resources to receive the trained weights and / or coefficients corresponding to the first BPT.

[0176] In some embodiments, the WTRU may determine that the accuracy parameters inferred based on a first AI / ML model are not within an acceptable range. For example, the WTRU may determine that parameters based on one or more measurements and / or predictions of the received signal and / or the channel based on the predicted first beam are not within an acceptable range. For example, the WTRU may determine that measured RSRP, RSRQ, SINR, etc., based on RS measurements according to the predicted first beam are below a corresponding configured threshold, which can be configured via RRC, MAC-CE, and / or DCI. As another example, the WTRU may determine that the assumed BLER based on measurements of the PDCCH received based on the predicted first beam is above a corresponding configured threshold, which can be configured via RRC, MAC-CE, and / or DCI.

[0177] The WTRU can determine and send a request to the base station to perform inference using a second AI / ML model (e.g., as described above). Upon receiving confirmation from the base station that the second AI / ML model has been used, the WTRU can select a second G-RNTI associated with the second AI / ML model. The WTRU can monitor and detect a DCI with CRC scrambled using the second G-RNTI to receive scheduled resources to receive trained weights and / or coefficients corresponding to the second AI / ML model. For example, the WTRU can use the received, transferred, and / or downloaded AI / ML model to perform inference and predict the optimal beam. The WTRU can report the predicted optimal beam and receive DL signaling and / or transmit UL signaling based on the predicted optimal beam.

[0178] In some embodiments, the WTRU may receive configurations (e.g., via RRC) regarding one or more generalized input and / or output BPTs, which can be used to indicate small and beamgroup-based AI / ML models. These configurations may include one or more BPT configurations and / or BPT indications. The WTRU may receive one or more configuration information regarding one or more input and / or output BPTs, including, for example, one or more of the following: BPT category, BPT ID, number of beams in the BPT, BPT mesh structure, focused beams in the BPT mesh, beam index in the BPT, trained weights and / or coefficients of the output BPT. The WTRU may receive one or more indications using one or more input and / or output BPTs (e.g., for hybrid inference at the WTRU), including BPT IDs and corresponding beam indices. The WTRU may determine one or more of the input BPT IDs and / or output BPT IDs based on one or more of the following: received input BPT ID, received output BPT ID, offset value, TCI state, absolute value, and / or generic default number. WTRU can perform, determine, and / or report beam measurements and / or beam IDs based on the determined input BPT and output BPT.

[0179] In some embodiments, the WTRU may be configured to receive configuration and / or indications (e.g., via RRC, MAC-CE, and / or DCI) regarding one or more input and / or output BPTs. For example, during initial cell access and / or after cell connection and in connected mode, the WTRU may receive configurations regarding the AI / ML model, input BPTs, and / or output BPTs. The WTRU may receive the configurations via cell-specific and / or WTRU-specific signaling. Configurations on input and / or output BPTs may include one or more of the following: BPT class, BPT ID, number of beams in the BPT, BPT mesh structure, focused beams in the BPT mesh, and / or beam indexes in the BPT.

[0180] The WTRU can receive indications about whether a configured BPT is an input or output BPT. The WTRU can receive the index or identifier associated with each BPT. The WTRU can receive input BPT 1, BPT 2, etc., and / or output BPT1, BPT2, etc. The WTRU can receive, or be configured with, the number of beams considered in each input and / or output BPT. For example, in... Figure 2 In the examples shown, the number of beams can be 16, 9, 15, and 15 for inputs BPT 202, BPT 204, BPT 206, and BPT 208, respectively. As another example, in... Figure 3In the example shown, the number of beams for output BPT302, output BPT304, output BPT306, and output BPT308 can be 8, 8, 4, and 4, respectively.

[0181] WTRU can be configured with the structure and size of the grid for each input and / or output BPT. For example, the BPT can be based on a square grid, rectangular grid, circular grid, hexagonal grid, etc. Figure 2 The input BPT 202 has a 4×4 square grid structure. As another example, Figure 3 The output BPT 304 has a 4×2 rectangular grid structure.

[0182] The WTRU can receive, or otherwise configure, focused beams for input and / or output BPTs for each configuration. For example, the focused beams in an input BPT can be associated with a subset of the set of B-beams included in the respective input BPT. Figure 2 The green beam in the input BPT 206 can be a focused beam. As another example, the focused beam in the output BPT can be associated with a beam for which trained weights and / or coefficients are provided as part of the BPT definition. For example, Figure 3 The beams 1f, 4f, 5f, and 8f in the output BPT 304 can be focused beams.

[0183] The WTRU can determine and / or configure beam indexing based on a numbering procedure, which can be a default numbering procedure. For example, the WTRU can consider numbering the beams within the configured BPT grid structure, starting from an initial value (e.g., zero) up to the total number of beams in the BPT. For example, the indication of the focused beam can therefore be based on the numbering procedure. Figure 2 The input BPT 204 can have 9 beams located in a 3×3 square grid structure. Thus, the WTRU can determine or configure the beam numbers, for example {0, 1, 2, ..., 8}. Therefore, the WTRU can determine or configure such that for the input BPT 204, the focused beams are beams {0, 4, 8}.

[0184] The WTRU can receive and / or otherwise configure trained weights and / or coefficients for the focused beam in each output BPT. For example, the WTRU can use the configured trained weights and / or coefficients as a model for inference on the WTRU side.

[0185] The WTRU can receive indication information to perform inference using one or more input and / or output BPTs and / or BPT pairs. For example, a BPT pair can be an indication of an input and output BPT pairing, where the WTRU can determine the indicated input and output BPTs by receiving the BPT pair indication. As another example, the WTRU can receive indication information from the base station via DCI, MAC-CE, and / or RRC. The indication information may include an indication of the BPT ID to be used. For inference purposes, the indication information may also include a beam index to be considered for the input and output BPTs. Based on one or more of absolute values, offset values, generic default numbers, and / or TCI states, the WTRU can receive an indication of the beam index to be used with one or more input and / or output BPTs.

[0186] The WTRU can receive indications of the absolute beam IDs of one or more BPTs. For example, the WTRU can receive indications of the absolute beam IDs to be used with the indicated input and output BPTs. Return to Reference Figure 7 For example, the first WTRU (e.g., Figure 5A The first WTRU (514) can be configured with an input BPT 202, wherein the beam index to be used with the configured BPT 202 can be indicated based on the absolute beam index considered within the network (e.g., beam indices {1, 18, 35, and 52}). As another example, the first WTRU (e.g., Figure 5A The WTRU 514 in the network can be configured with an output BPT 304, wherein the beam index to be used with the configured BPT 304 can be indicated based on the absolute beam index considered within the network (e.g., beam indexes {2, 19, 34, and 51}).

[0187] The WTRU can receive an indication of the first and / or anchor absolute beam ID and one or more offsets of the first beam ID to determine additional beam IDs that will be used with the indicated input and / or output BPT. For example, the first WTRU (e.g., Figure 5A The first WTRU (514) can be configured with an input BPT 202, wherein the beam index to be used with the configured BPT 202 can be indicated based on the absolute beam index of the first and / or anchor beam (e.g., beam 1a) and one or more offset values, wherein each offset value is indicated based on a previous beam index (e.g., offset value {17, 17, 17}). In this way, the WTRU can use the absolute beam index of the first and / or anchor beam and determine the beam indices of other beam indices accordingly (e.g., beam indices {1, 18, 35, and 52}).

[0188] The WTRU may receive and / or otherwise configure one or more settings for determining the indicated input and output BPT beam indices based on generic and / or default numbers in the BPT. Return to Reference Figure 2 and Figure 3 For example, a WTRU can receive and / or be configured to use... Figure 2 Input BPT 202 and Figure 3 The output BPT304 indicates the beam index. The WTRU can use the default or configured and / or pre-configured beam index {0, 1, 2, ..., 15} as the beam index in the input BPT 202. In this way, the WTRU can use the beam index {0, 5, 10, 15} as the focusing beam index in the input BPT 202.

[0189] Figure 8A A diagram showing an example of a beam index based on a generic default number. Figure 8B This is another example of a diagram indicating the beam index based on the general default number.

[0190] The WTRU can configure or determine the beam index and focusing beam index corresponding to the configured and / or indicated output BPT based on the indicator beam index of the input BPT. The WTRU can determine, configure, or receive an indication of the first and / or anchor focusing beam of the output BPT based on the indicator beam index of the input BPT. That is, for example, based on the indicator beam index of the input BPT, the WTRU can receive an indication that the first and / or anchor focusing beam of output BPT 2 is beam #1. Thus, the WTRU can determine other beam indices of output BPT2 as, for example, beam indices {1, 2, 5, 6, 9, 10, 13, 14}. This can be done, for example... Figure 8A As can be seen, beams 0, 5, 10, and 15 are beam indices corresponding to the input BPT, and beams 1, 6, 9, and 14 are beam indices corresponding to the output BPT.

[0191] If the WTRU does not receive an indication of the first and / or anchor focus beam index of the configured output BPT, the WTRU can determine that the first and / or anchor focus beam index of the configured output BPT is the same as the first beam index of the configured input BPT. That is, for example, a WTRU configured with input BPT 1 and output BPT 2 but without a configured first and / or anchor focus beam for the output BPT can determine that the first focus beam for the indicated output BPT 2 is the same as the first beam index (e.g., beam #0) for the input BPT. Thus, the WTRU can determine other beam indices for output BPT 2 as, for example, beam indices {0, 1, 4, 5, 8, 9, 12, 13}. This can be done in, for example... Figure 8BAs can be seen, beams 0, 5, 10, and 15 are beam indices corresponding to the input BPT, and beams 0, 5, 8, and 13 are beam indices corresponding to the output BPT.

[0192] The WTRU can determine the beams belonging to a BPT (e.g., an input BPT) based on BPT information (e.g., BPT ID and / or the number of beams in the BPT) and the active TCI state set. For example, the WTRU can be configured to select beams associated with TCI states #1, 2, and 3 of the active TCI state set as beams belonging to a first input BPT consisting of three beams. For example, the WTRU can be configured to select beams associated with TCI states #3 and 4 of the active TCI state set as beams belonging to a second input BPT consisting of two beams. Furthermore, the WTRU can be configured to assign BPT-specific beam IDs to the beams of a BPT. For example, the WTRU can assign beam IDs in descending order of the determined L1-RSRP of the RS associated with the beams belonging to the BPT. Additionally or alternatively, the WTRU can receive a configuration of a mapping between the indicated TCI state and the configured BPT (e.g., an output BPT).

[0193] Although the features and elements have been described above in specific combinations, those skilled in the art will appreciate that each feature or element can be used alone or in any combination with other features and elements. Furthermore, the methods described herein can be implemented in a computer program, software, or firmware incorporated in a computer-readable medium for execution by a computer or processor. Examples of computer-readable media include electronic signals (transmitted via wired or wireless connections) and computer-readable storage media. Examples of computer-readable storage media include, but are not limited to, read-only memory (ROM), random access memory (RAM), registers, cache memory, semiconductor memory devices, magnetic media such as internal hard disks and removable disks, magneto-optical media, and optical media such as CD-ROMs and digital multifunction discs (DVDs). The processor associated with the software can be used to implement a radio frequency transceiver for use in a WTRU, UE, terminal, base station, RNC, or any host computer.

Claims

1. A wireless transmit / receive unit (WTRU) configured with a plurality of input beam patterns, a plurality of output beam patterns, a plurality of artificial intelligence / machine learning (AI / ML) models, and trained weights for each of the plurality of AI / ML models, wherein each of the plurality of input beam patterns includes a plurality of beam indices corresponding to a set of input beams, each of the plurality of output beam patterns includes a plurality of beam indices corresponding to a set of output beams, and wherein each of the plurality of AI / ML models corresponds to a corresponding one of the plurality of output beam patterns, the WTRU comprising: transceiver; and processor, The transceiver and the processor are configured to measure the signal quality of a subset of the input beam sets for each of the plurality of input beam patterns to identify at least one input beam with the highest signal quality. The transceiver and the processor are further configured to report the at least one input beam to the base station. The transceiver and the processor are further configured to receive from the base station an indication of one of the plurality of input beam patterns and an indication of one of the plurality of output beam patterns, for use in predicting the output beam based on the report. The transceiver and the processor are further configured to predict the output beam using an AI / ML model corresponding to one indicated in the plurality of output beam patterns and trained weights corresponding to the AI / ML model. The transceiver and the processor are further configured to report the predicted output beam to the base station.

2. The WTRU according to claim 1, wherein: The transceiver and the processor are further configured to receive additional configurations for at least one additional input beam pattern including a plurality of beam indices corresponding to another set of input beams and at least one additional output beam pattern including a plurality of beam indices corresponding to another set of output beams. The transceiver and processor are also configured to use an additional configuration to predict another output beam.

3. The WTRU according to claim 2, wherein: The transceiver and processor are also configured to compare at least one accuracy parameter of the predicted output beam with that of another predicted output beam. The transceiver and the processor are further configured to determine that at least one accuracy parameter of the predicted other output beam is better than at least one accuracy parameter of the predicted output beam, and The transceiver and processor are also configured to send messages to the base station, which include a request to switch from a predicted output beam to another predicted output beam.

4. The WTRU of claim 2 or 3, wherein the transceiver and the processor are further configured to send a message to the base station, wherein the message includes a request to switch to further prediction using the at least one other input beam pattern and the at least one other output beam pattern.

5. The WTRU according to any one of claims 1-4, wherein the transceiver and the processor are further configured to receive, via radio resource control (RRC) signaling, a configuration for at least one of the plurality of input beam patterns, the plurality of output beam patterns, the plurality of AI / ML models, or the trained weights of each of the plurality of AI / ML models.

6. The WTRU according to any one of claims 1-5, wherein: The transceiver and the processor are also configured to monitor the Physical Downlink Control Channel (PDCCH) to detect Downlink Control Information (DCI) scrambled with Cyclic Redundancy Check (CRC) using Group Radio Temporary Identifier (G-RNTI), thereby retrieving the trained weights of the AI / ML model based on one or more of the following: the WTRU receives an indication to use a specific input beam pattern or a specific output beam pattern and the WTRU does not have the corresponding trained weights; the WTRU receives an indication to update the training weights at the base station; or the WTRU determines that the accuracy parameters of the predicted output beam do not meet the defined criteria.

7. A method implemented in a wireless transmit / receive unit (WTRU) configured with a plurality of input beam patterns, a plurality of output beam patterns, a plurality of artificial intelligence / machine learning (AI / ML) models, and trained weights for each of the plurality of AI / ML models, wherein each of the plurality of input beam patterns includes a plurality of beam indices corresponding to a set of input beams, wherein each of the plurality of output beam patterns includes a plurality of beam indices corresponding to a set of output beams, and wherein each of the plurality of AI / ML models corresponds to a corresponding one of the plurality of output beam patterns, the method comprising: Measure the signal quality of a subset of the input beam set for each of multiple input beam patterns to identify at least one input beam with the highest signal quality; Report the at least one input beam to the base station; The system receives an indication of one of multiple input beam patterns and an indication of one of multiple output beam patterns from the base station for use in predicting the output beam based on the report. The output beam is predicted using an AI / ML model corresponding to one of the indicated multiple output beam patterns and trained weights corresponding to the AI / ML model; and Report the predicted output beam to the base station.

8. The method of claim 7, further comprising: Receive at least one other input beam pattern including a plurality of beam indices corresponding to another set of input beams and another configuration including at least one other output beam pattern including a plurality of beam indices corresponding to another set of output beams; and Use another configuration to predict another output beam.

9. The method of claim 8, further comprising: Compare at least one accuracy parameter of the predicted output beam with that of another predicted output beam; Determine at least one accuracy parameter of the predicted other output beam that is better than at least one accuracy parameter of the predicted output beam; and A message is sent to the base station, wherein the message includes a request to switch from the predicted output beam to another predicted output beam.

10. The method of claim 7 or 8, further comprising sending a message to the base station, wherein the message includes a request to switch to further prediction using the at least one other input beam pattern and the at least one other output beam pattern.

11. The method according to any one of claims 6-10, further comprising receiving via radio resource control (RRC) signaling a configuration for at least one of the plurality of input beam patterns, the plurality of output beam patterns, the plurality of AI / ML models, or the trained weights of each of the plurality of AI / ML models.

12. The method according to any one of claims 6-11, further comprising: The Physical Downlink Control Channel (PDCCH) is monitored to detect Downlink Control Information (DCI) scrambled with Cyclic Redundancy Check (CRC) using Group Radio Temporary Identifier (G-RNTI), thereby retrieving the trained weights of the AI / ML model based on one or more of the following: the WTRU receives an indication to use a specific input beam pattern or a specific output beam pattern and the WTRU does not have the corresponding trained weights; the WTRU receives an indication that the trained weights have been updated at the base station; or the WTRU determines that the accuracy parameters of the predicted output beam do not meet the defined criteria.

13. A base station, comprising: transceiver; and processor, The transceiver and the processor are configured to send at least one configuration message to the base station, wherein the at least one configuration message includes configuration information for a plurality of input beam patterns, a plurality of output beam patterns, a plurality of artificial intelligence / machine learning (AI / ML) models, and trained weights for each of the plurality of AI / ML models, wherein each of the plurality of input beam patterns includes a plurality of beam indices corresponding to a set of input beams, wherein each of the plurality of output beam patterns includes a plurality of beam indices corresponding to a set of output beams, and wherein each of the plurality of AI / ML models corresponds to a corresponding one of the plurality of output beam patterns. The transceiver and the processor are further configured to receive reports from the WTRU of the measurement beam at the WTRU in a subset of the input beam sets of each of the plurality of input beam patterns. The transceiver and the processor are further configured to send an indication of one of the plurality of input beam patterns and an indication of one of the plurality of output beam patterns to the WTRU for predicting the output beam based on the report, and The transceiver and the processor are further configured to receive reports of predicted beams from the WTRU, the predicted beams being predicted using an AI / ML model corresponding to one indicated in the plurality of output beam patterns and trained weights corresponding to the AI / ML model. The transceiver and the processor are also configured to send downlink transmissions to the WTRU using the predicted beam.

14. The base station according to claim 13, wherein: The transceiver and the processor are further configured to transmit for another configuration including at least one additional input beam pattern corresponding to a plurality of beam indices of another input beam set and at least one additional output beam pattern corresponding to a plurality of beam indices of another output beam set, and The transceiver and the processor are also configured to receive reports from the WTRU of other output beams predicted using other configurations.

15. The base station according to claim 13 or 14, wherein: The transceiver and the processor are also configured to transmit downlink control information (DCI) scrambled with group radio temporary identifiers (G-RNTI) and cyclic redundancy check (CRC) on the physical downlink control channel (PDCCH), wherein the DCI includes scheduling information for WTRU to retrieve trained weights of the AI / ML model.

Images