WO2024173223A1 - Methods on supporting dynamic model selection for wireless communication - Google Patents

Abstract

A wireless transmit/receive unit (WTRU) is disclosed that includes a processor configured to store one or more AI/ML model configurations. Each AI/ML model configuration may have an associated identifier. The WTRU may include a transmitter configured to transmit a message to a gNB indicating AI/ML model reconfiguration. The AI/ML model reconfiguration may be based on NACKs. The AI/ML model reconfiguration may be based on trigger conditions. The WTRU may transmit a message indicating a preferred AI/ML model. The preferred AI/ML model may be based on one or more of a configured threshold, a RDRP, a speed of the WTRU, a pathloss, a zone ID, a cell ID, or a TRP ID. The WTRU may evaluate one or more AI/ML model configurations. The WTRU may receive, from the gNB, a confirmation message of the AI/ML model reconfiguration.

Description

METHODS ON SUPPORTING DYNAMIC MODEL SELECTION FOR WIRELESS COMMUNICATION

RELATED APPLICATIONS

[0001] The present application claims the benefit of and priority to U S. Provisional Patent Application No. 63/445,452, entitled "Methods on Supporting Dynamic Model Selection for Wireless Communication,” filed February 14, 2023, the entirety of which is incorporated by reference herein.

BACKGROUND

[0002] New implementations of wireless communications systems and protocols are able to deliver high throughput, low latency data exchanges, but may require complex configurations to achieve these benefits. For example, to efficiently increase signal strength without wasting power and/or to reduce interference or support additional user equipment (UE) or other devices, some implementations of communication systems use spatial filtering or beamforming to steer wireless signals to target recipient devices. Establishing this steering may require complex calculations, particularly for mobile devices that move in relation to each other, resulting in quickly changing transmission angles (and corresponding phasing and amplitude adjustments for multi-antenna arrays). Similarly, communication channel configurations (e.g. timing, frequencies, etc.) may need to change to avoid interference, meet application demands, etc.

SUMMARY

[0003] To provide the benefits of next generation wireless communications systems and predict changing parameters and configurations so that they can be applied in advance of congestion, interference, or other such conditions, implementations of the systems and methods discussed herein utilize artificial intelligence/machine learning (AI/ML) systems. These AI/ML systems may be based on trained models, enabling prediction of changing environmental conditions and adjustment of beamforming parameters or channel state information that can be applied ideally prior to any adverse effects experienced by the communicating devices. However, different AI/ML models may provide better performance depending on changing conditions. For example, an AI/ML model optimized for predicting beam management adjustments as devices or other entities move within an area may not necessarily make accurate predictions when interference from a local airport radar system is present.

[0004] Accordingly, in some aspects, the present disclosure is directed to implementations of systems and methods for dynamically identifying or selecting AI/ML models or model types to provide optimized predictions based on changing environmental conditions, device movement, or combinations of these or other changes. In some implementations, trained models may be generated (and/or updated) by UEs, base stations (gNBs), or other network devices (e g. application servers, AI/ML processing nodes, etc.) and may be provided to the devices (e.g. UEs and gNBs or other devices) for execution, prediction of conditions or characteristics, and adjustment of parameters. In some implementations, these predictions may enable devices to apply parameter adjustments in advance of requests or notifications from partner devices, such as changing retransmission window parameters immediately in response to packet loss rate changes without having to first exchange configuration change details, or other such modifications In some implementations, the devices (e.g. UEs, gNBs or other devices) may monitor conditions or performance of the network and may dynamically determine whether to change AI/ML models, and if so, what model to utilize. This may significantly reduce management overhead, and enable faster mitigation or recovery from changing channel conditions or issues.

[0005] In some embodiments for receiving configuration of AI/ML models with AI/ML model types, A UE is configured with one or more AI/ML models with model IDs, wherein each AI/ML model is associated/configured with a set of CSI parameters (potentially for each CSI type orCSI config e.g., forCSI and for BM). Alternatively, the UE indicates its capability (e.g., required CSI parameters for each AI/ML model).

[0006] In some embodiments, a UE indicates a need of AI/ML model reselection If one or more of number of NACKs (e.g., number of consecutive NACKs > threshold), RSRP e.g., RSRP < threshold), change of UE speed (e g., Current UE speed - Avg. UE speed within a measurement window > threshold), change of UE position (e.g., Current UE position - Avg. UE position within a measurement window > threshold), change of pathloss (e g., Current pathloss - Avg. pathloss within a measurement window > threshold), hypothetical PDCCH BLER < threshold, measured SINR or differential SINR (e.g., from scheduled DMRS ports) and etc.

[0007] In some embodiments, a UE indicates a type of AI/ML model for reselection based on the triggered conditions. If measured RSRP > threshold and number of consecutive NACKs > threshold, reselect AI/ML model for CSI. If measured RSRP < threshold, reselect AI/ML model for CSI. If number of failed measured hypothetical PDCCH BLER < threshold, reselect AI/ML model for CSI. Otherwise, reselect AI/ML for BM.

[0008] In some embodiments, a UE indicates AI/ML model type for reselection. The UE indicates an AI/ML model type for reselection to a gNB and receives the one or more RSs associated with the reported type. AI/ML model for CSI may be reselected after reselection of AI/ML model for BM if AI/ML model for BM is requested.

[0009] In some embodiments, a UE identifies a preferred model. UE identification of a preferred model, possibly with a model type indication (or CSI config ID), based on measuring one or more of configured thresholds, RSRP, UE speed, pathloss, zone ID, cell ID, TRP ID. For example, if UE speed is < X and Pathloss > Y (e.g., indoor), then determines Model #1. If UE speed is < X and Pathloss < Y (e g., walking outdoor), then determines Model #2. If UE speed is > X and Pathloss < Y (e.g., high speed outdoor), then determines Model #3.

[0010] In some embodiments, the UE indicates a preferred set of CSI parameters to a gNB for future reporting. For example, Model #1 : X number of beam IDs and corresponding L1-RSRPs for the current measurement (N1). Model #2: X1-1 number of beam IDs and corresponding L1-RSRPs for the current measurement and X1-2 number of beam IDs and corresponding L1-RSRPs for N1+N2. Model #3: X2-1, X2-2, X2-3 and X2-4 number of beam IDs and corresponding L1-RSRPs for N1/N2/N3/N4.

[0011] In some embodiments, a UE evaluates associated AI/ML models. Alternatively, the UE activates a set of associated AI/ML models, of the one or more AI/ML models, with the currently identified AI/ML model/model type for reselection to evaluate prediction accuracy of the AI/ML models. D details for the evaluation e.g., beam prediction accuracy/RSRP difference and RS transmission for the evaluation, different RS transmission for the activated models.

[0012] In some embodiments, a gNB confirms AI/ML model type for reselection. The UE receives a confirmation of the UE request and activates the indicated model for the reported request type.

[0013] In some embodiments, the UE behavior after receiving the gNB confirmation of AI/ML model type for reselection changes. For example, The UE may deactivate the previous model or the activated models for evaluation, activates the newly indicated model and apply the associated set of CSI parameters for next CSI report including a preferred size of measurement/prediction window.

BRIEF DESCRIPTION OF THE DRAWINGS

[0014] A more detailed understanding may be had from the following description, given by way of example in conjunction with the accompanying drawings, wherein like reference numerals in the figures indicate like elements, and wherein:

[0015] FIG. 1A is a system diagram illustrating an example communications system in which one or more disclosed embodiments may be implemented;

[0016] FIG. 1 B is a system diagram illustrating an example wireless transmit/receive unit (WTRU) that may be used within the communications system illustrated in FIG. 1A according to an embodiment;

[0017] FIG. 1C is a system diagram illustrating an example radio access network (RAN) and an example core network (CN) that may be used within the communications system illustrated in FIG. 1A according to an embodiment;

[0018] FIG. 1D is a system diagram illustrating a further example RAN and a further example CN that may be used within the communications system illustrated in FIG. 1A according to an embodiment;

[0019] FIG. 2 is a block diagram of a system for hybrid beamforming, according to some implementations;

[0020] FIG. 3 is a block diagram of a system for dynamic model selection for wireless communications, according to some implementations;

[0021] FIG. 4 is a flow chart of a method for dynamic model selection for wireless communications, according to some implementations; and

[0022] FIG. 5 is a logic diagram of an example implementation of dynamic model selection for wireless communications.

DETAILED DESCRIPTION

[0023] FIG. 1A is a diagram illustrating an example communications system 100 in which one or more disclosed embodiments may be implemented. The communications system 100 may be a multiple access system that provides content, such as voice, data, video, messaging, broadcast, etc., to multiple wireless users. The communications system 100 may enable multiple wireless users to access such content through the sharing of system resources, including wireless bandwidth. For example, the communications systems 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), singlecarrier FDMA (SC-FDMA), zero-tail unique-word discrete Fourier transform Spread OFDM (ZT-UW-DFT-S- OFDM), unique word OFDM (UW-OFDM), resource block-filtered OFDM, filter bank multicarrier (FBMC), and the like.

[0024] As shown in FIG. 1A, the communications system 100 may include wireless transmit/receive units (WTRUs) 102a, 102b, 102c, 102d, a radio access network (RAN) 104, a core network (ON) 106, a public switched telephone network (PSTN) 108, the Internet 110, and other networks 112, though itwill be appreciated that the disclosed embodiments contemplate 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 By way of example, the WTRUs 102a, 102b, 102c, 102d, any of which may be referred to as a station (STA), may be configured to transmit and/or receive wireless signals and may include a user equipment (UE), a mobile station, a fixed or mobile subscriber unit, a subscription-based unit, a pager, a cellular telephone, a personal digital assistant (PDA), a smartphone, a laptop, a netbook, a personal computer, a wireless sensor, a hotspot or Mi-Fi device, an Internet of Things (loT) device, a watch or other wearable, a head-mounted display (HMD), a vehicle, a drone, a medical device and applications (e.g., remote surgery), an industrial device and applications (e.g., a robot and/or other wireless devices operating in an industrial and/or an automated processing chain contexts), a consumer electronics device, a device operating on commercial and/or industrial wireless networks, and the like. Any of the WTRUs 102a, 102b, 102c and 102d may be interchangeably referred to as a UE.

[0025] The communications systems 100 may also include a base station 114a and/or a base station 114b. Each of the base stations 114a, 114b may be any type of device configured to wirelessly interface with at least one of the WTRUs 102a, 102b, 102c, 102d to facilitate access to one or more communication networks, such as the CN 106, the Internet 110, and/or the other networks 112. By way of example, the base stations 114a, 114b may be a base transceiver station (BTS), a NodeB, an eNode B (eNB), a Home Node B, a Home eNode B, a next generation NodeB, such as a gNode B (gNB), a new radio (NR) NodeB, a site controller, an access point (AP), a wireless router, and the like. While the base stations 114a, 114b are each depicted as a single element, it will be appreciated that the base stations 114a, 114b may include any number of interconnected base stations and/or network elements.

[0026] The base station 114a may be part of the RAN 104, which may also include other base stations and/or network elements (not shown), such as a base station controller (BSC), a radio network controller (RNC), relay nodes, and the like. The base station 114a and/or the base station 114b may be configured to transmit and/or receive wireless signals on one or more carrier frequencies, which may be referred to as a cell (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 a wireless service to a specific geographical area that may be relatively fixed or that may change over time. The cell may further be divided into cell sectors. For example, the cell associated with the base station 114a may be divided into three sectors. Thus, in one embodiment, the base station 114a may include three transceivers, i.e., one for each sector of the cell. In an embodiment, the 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 desired spatial directions.

[0027] The base stations 114a, 114b may communicate with one or more of the WTRUs 102a, 102b, 102c, 102d over an air interface 116, which may be any suitable wireless communication link (e.g., radio frequency (RF), microwave, centimeter wave, micrometer wave, infrared (IR), ultraviolet (UV), visible light, etc.). The air interface 116 may be established using any suitable radio access technology (RAT).

[0028] More specifically, as noted above, the communications system 100 may be a multiple access system and may employ one or more channel access schemes, such as CDMA, TDMA, FDMA, OFDMA, SC-FDMA, and the like. For example, the base station 114a in the RAN 104 and the WTRUs 102a, 102b, 102c may implement a radio technology such as Universal Mobile Telecommunications System (UMTS) Terrestrial Radio Access (UTRA), which may establish the air interface 116 using wideband CDMA (WCDMA). WCDMA may include communication protocols such as High-Speed Packet Access (HSPA) and/or Evolved HSPA (HSPA+). HSPA may include High-Speed Downlink (DL) Packet Access (HSDPA) and/or High-Speed Uplink (UL) Packet Access (HSUPA).

[0029] In an embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement a radio technology such as Evolved UMTS Terrestrial Radio Access (E-UTRA), which may establish the air interface 116 using Long Term Evolution (LTE) and/or LTE-Advanced (LTE-A) and/or LTE-Advanced Pro (LTE-A Pro).

[0030] In an embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement a radio technology such as NR Radio Access , which may establish the air interface 116 using NR.

[0031] In an embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement multiple radio access technologies. For example, the base station 114a and the WTRUs 102a, 102b, 102c may implement LTE radio access and NR radio access together, for instance using dual connectivity (DC) principles. Thus, the air interface utilized by WTRUs 102a, 102b, 102c may be characterized by multiple types of radio access technologies and/or transmissions sent to/from multiple types of base stations (e.g , an eNB and a gNB).

[0032] In other embodiments, the base station 114a and the WTRUs 102a, 102b, 102c may implement radio technologies such as IEEE 802.11 (i.e , Wireless Fidelity (WiFi), IEEE 802.16 (i.e., Worldwide Interoperability for Microwave Access (WiMAX)), CDMA2000, CDMA2000 1X, CDMA2000 EV-DO, Interim Standard 2000 (IS-2000), Interim Standard 95 (IS-95), Interim Standard 856 (IS-856), Global System for Mobile communications (GSM), Enhanced Data rates for GSM Evolution (EDGE), GSM EDGE (GERAN), and the like. [0033] The base station 114b in FIG 1A may be a wireless router, Home Node B, Home eNode B, or access point, for example, and may utilize any suitable RAT for facilitating wireless connectivity in a localized area, such as a place of business, a home, a vehicle, a campus, an industrial facility, an air corridor (e.g, for use by drones), a roadway, and the like. In one embodiment, the base station 114b and the WTRUs 102c, 102d may implement a radio technology such as IEEE 802.11 to establish a wireless local area network (WLAN). In an embodiment, the base station 114b and the WTRUs 102c, 102d may implement a radio technology such as IEEE 802.15 to establish a wireless personal area network (WPAN). In yet another embodiment, the base station 114b and the WTRUs 102c, 102d may utilize a cellular-based RAT (e.g, WCDMA, CDMA2000, GSM, LTE, LTE-A, LTE-A Pro, NR etc.) to establish a picocell or femtocell. As shown in FIG. 1A, the base station 114b may have a direct connection to the Internet 110. Thus, the base station 114b may not be required to access the Internet 110 via the CN 106.

[0034] The RAN 104 may be in communication with the CN 106, which may be any type of network configured to provide voice, data, applications, and/or voice over internet protocol (VoIP) services to one or more of the WTRUs 102a, 102b, 102c, 102d. The data may have varying quality of service (QoS) requirements, such as differing throughput requirements, latency requirements, error tolerance requirements, reliability requirements, data throughput requirements, mobility requirements, and the like. The CN 106 may provide call control, billing services, mobile location-based services, pre-paid calling, Internet connectivity, video distribution, etc, and/or perform high-level security functions, such as user authentication. Although not shown in FIG. 1A, it will be appreciated that the RAN 104 and/or the CN 106 may be in direct or indirect communication with other RANs that employ the same RAT as the RAN 104 or a different RAT. For example, in addition to being connected to the RAN 104, which may be utilizing a NR radio technology, the CN 106 may also be in communication with another RAN (not shown) employing a GSM, UMTS, CDMA 2000, WiMAX, E-UTRA, or WiFi radio technology.

[0035] The CN 106 may also serve as a gateway for the WTRUs 102a, 102b, 102c, 102d to access the PSTN 108, the Internet 110, and/or the other networks 112. The PSTN 108 may include circuit-switched telephone networks that provide plain old telephone service (POTS). The Internet 110 may include a global system of interconnected computer networks and devices that use common communication protocols, such as the transmission control protocol (TCP), user datagram protocol (UDP) and/or the internet protocol (IP) in the TCP/IP internet protocol suite. The networks 112 may include wired and/or wireless communications networks owned and/or operated by other service providers. For example, the networks 112 may include another CN connected to one or more RANs, which may employ the same RAT as the RAN 104 or a different RAT.

[0036] Some or all of the WTRUs 102a, 102b, 102c, 102d in the communications system 100 may include multi-mode capabilities (e.g, the WTRUs 102a, 102b, 102c, 102d may include multiple transceivers for communicating with different wireless networks over different wireless links). For example, the WTRU 102c shown in FIG. 1 A may be configured to communicate with the base station 114a, which may employ a cellularbased radio technology, and with the base station 114b, which may employ an IEEE 802 radio technology. [0037] FIG. 1 B is a system diagram illustrating an example WTRU 102. As shown in FIG. 1 B, the WTRU 102 may include a processor 118, a transceiver 120, a transmit/receive element 122, a speaker/microphone 124, a keypad 126, a display/touchpad 128, non-removable memory 130, removable memory 132, a power source 134, a global positioning system (GPS) chipset 136, and/or other peripherals 138, among others. It will be appreciated that the WTRU 102 may include any sub-combination of the foregoing elements while remaining consistent with an embodiment.

[0038] The processor 118 may be a general purpose processor, a special purpose processor, a conventional processor, a digital signal processor (DSP), a plurality of microprocessors, one or more microprocessors in association with a DSP core, a controller, a microcontroller, Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs), any other type of integrated circuit (IC), a state machine, and the like. The processor 118 may perform signal coding, data processing, power control, input/output processing, and/or any other functionality that enables the WTRU 102 to operate in a wireless environment. The processor 118 may be coupled to the transceiver 120, which may be coupled to the transmit/receive element 122. While FIG. 1 B depicts the processor 118 and the transceiver 120 as separate components, it will be appreciated that the processor 118 and the transceiver 120 may be integrated together in an electronic package or chip.

[0039] The transmit/receive element 122 may be configured to transmit signals to, or receive signals from, a base station (e.g., the base station 114a) over the air interface 116. For example, in one embodiment, the transmit/receive element 122 may be an antenna configured to transmit and/or receive RF signals. In an embodiment, the transmit/receive element 122 may be an emitter/detector configured to transmit and/or receive IR, UV, or visible light signals, for example. In yet another embodiment, the transmit/receive element 122 may be configured to transmit and/or receive both RF and light signals. It will be appreciated that the transmit/receive element 122 may be configured to transmit and/or receive any combination of wireless signals.

[0040] Although the transmit/receive element 122 is depicted in FIG. 1 B 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 over the air interface 116. [0041] The transceiver 120 may be configured to modulate the signals that are to be transmitted by the transmit/receive element 122 and to demodulate the signals that are received by the transmit/receive element 122. As noted above, the WTRU 102 may have multi-mode capabilities. Thus, the transceiver 120 may include multiple transceivers for enabling the WTRU 102 to communicate via multiple RATs, such as NR and IEEE 802.11 , for example.

[0042] The processor 118 of the WTRU 102 may be coupled to, and may receive user input data from, the speaker/microphone 124, the keypad 126, and/or the display/touchpad 128 (e.g., a liquid crystal display (LCD) display unit or organic light-emitting diode (OLED) display unit) The processor 118 may also output user data to the speaker/microphone 124, the keypad 126, and/or the display/touchpad 128. In addition, the processor 118 may access information from, and store data in, any type of suitable memory, such as the non-removable memory 130 and/or the removable memory 132. The 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. The removable memory 132 may include a subscriber identity module (SIM) card, a memory stick, a secure digital (SD) memory card, and the like. In other embodiments, the processor 118 may access information from, and store data in, memory that is not physically located on the WTRU 102, such as on a server or a home computer (not shown).

[0043] The processor 118 may receive power from the power source 134, and may be configured to distribute and/or control the power to the other components in the WTRU 102. The power source 134 may be any suitable device for powering the WTRU 102. For example, the power source 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.

[0044] The processor 118 may also be coupled to the GPS chipset 136, which may be configured to provide location information (e.g., longitude and latitude) regarding the current location of the WTRU 102. In addition to, or in lieu of, the information from the GPS chipset 136, the WTRU 102 may receive location information over the air interface 116 from a base station (e.g., base stations 114a, 114b) and/or determine its location based on the timing of the signals being received from two or more nearby base stations. It will be appreciated that the WTRU 102 may acquire location information by way of any suitable location-determination method while remaining consistent with an embodiment

[0045] The processor 118 may further be coupled to other peripherals 138, which may include one or more software and/or hardware modules that provide additional features, functionality and/or wired or wireless connectivity. For example, the peripherals 138 may include an accelerometer, an e-compass, a satellite transceiver, a digital camera (for photographs and/or video), a universal serial bus (USB) port, a vibration device, a television transceiver, a hands free headset, a Bluetooth® module, a frequency modulated (FM) radio unit, a digital music player, a media player, a video game player module, an Internet browser, a Virtual Reality and/or Augmented Reality (VR/AR) device, an activity tracker, and the like. The peripherals 138 may include one or more sensors. The sensors may be one or more of a gyroscope, an accelerometer, a hall effect sensor, a magnetometer, an orientation sensor, a proximity sensor, a temperature sensor, a time sensor; a geolocation sensor, an altimeter, a light sensor, a touch sensor, a magnetometer, a barometer, a gesture sensor, a biometric sensor, a humidity sensor and the like.

[0046] The WTRU 102 may include a full duplex radio for which transmission and reception of some or all of the signals (e g., associated with particular subframes for both the 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 either hardware (e.g., a choke) or signal processing via a processor (e.g., a separate processor (not shown) or via processor 118). In an embodiment, the WTRU 102 may include a half-duplex radio for which transmission and reception of some or all of the signals (e.g., associated with particular subframes for either the UL (e g., for transmission) or the DL (e g., for reception)).

[0047] FIG. 10 is a system diagram illustrating the RAN 104 and the CN 106 according to an embodiment. As noted above, the RAN 104 may employ an E-UTRA radio technology to communicate with the WTRUs 102a, 102b, 102c over the air interface 116. The RAN 104 may also be in communication with the ON 106.

[0048] The RAN 104 may include eNode-Bs 160a, 160b, 160c, though it will be appreciated that the RAN 104 may include any number of eNode-Bs while remaining consistent with an embodiment. The eNode-Bs 160a, 160b, 160c may each include one or more transceivers for communicating with the WTRUs 102a, 102b, 102c over the air interface 116. In one embodiment, the eNode-Bs 160a, 160b, 160c may implement MIMO technology. Thus, the eNode-B 160a, for example, may use multiple antennas to transmit wireless signals to, and/or receive wireless signals from, the WTRU 102a.

[0049] Each of the eNode-Bs 160a, 160b, 160c may be associated with a particular cell (not shown) and may be configured to handle radio resource management decisions, handover decisions, scheduling of users in the UL and/or DL, and the like. As shown in FIG. 1 C, the eNode-Bs 160a, 160b, 160c may communicate with one another over an X2 interface.

[0050] The CN 106 shown in FIG. 1C 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 the CN 106, it will be appreciated that any of these elements may be owned and/or operated by an entity other than the CN operator.

[0051] The MME 162 may be connected to each of the eNode-Bs 162a, 162b, 162c in the RAN 104 via an S1 interface and may serve as a control node. For example, the MME 162 may be responsible for authenticating users of the WTRUs 102a, 102b, 102c, bearer activation/deactivation, selecting a particular serving gateway during an initial attach of the WTRUs 102a, 102b, 102c, and the like. The MME 162 may provide a control plane function for switching between the RAN 104 and other RANs (not shown) that employ other radio technologies, such as GSM and/or WCDMA

[0052] The SGW 164 may be connected to each of the eNode Bs 160a, 160b, 160c in the RAN 104 via the S1 interface. The SGW 164 may generally route and forward user data packets to/from the WTRUs 102a, 102b, 102c. The SGW 164 may perform other functions, such as anchoring user planes during inter-eNode B handovers, triggering paging when DL data is available for the WTRUs 102a, 102b, 102c, managing and storing contexts of the WTRUs 102a, 102b, 102c, and the like.

[0053] The SGW 164 may be connected to the PGW 166, which may provide the WTRUs 102a, 102b, 102c with access to packet-switched networks, such as the Internet 110, to facilitate communications between the WTRUs 102a, 102b, 102c and IP-enabled devices.

[0054] The CN 106 may facilitate communications with other networks For example, the CN 106 may provide the WTRUs 102a, 102b, 102c with access to circuit-switched networks, such as the PSTN 108, to facilitate communications between the WTRUs 102a, 102b, 102c and traditional land-line communications devices. For example, the CN 106 may include, or may communicate with, an IP gateway (e.g., an IP multimedia subsystem (IMS) server) that serves as an interface between the CN 106 and the PSTN 108. In addition, the CN 106 may provide the WTRUs 102a, 102b, 102c with access to the other networks 112, which may include other wired and/or wireless networks that are owned and/or operated by other service providers.

[0055] Although the WTRU is described in FIGS. 1A-1 D as a wireless terminal, it is contemplated that in certain representative embodiments that such a terminal may use (e.g., temporarily or permanently) wired communication interfaces with the communication network.

[0056] In representative embodiments, the other network 112 may be a WLAN.

[0057] A WLAN in Infrastructure Basic Service Set (BSS) mode may have an Access Point (AP) for the BSS and one or more stations (STAs) associated with the AP. The AP may have access or an interface to a Distribution System (DS) or another type of wired/wireless network that carries traffic in to and/or out of the BSS. Traffic to STAs that originates from outside the BSS may arrive through the AP and may be delivered to the STAs. Traffic originating from STAs to destinations outside the BSS may be sent to the AP to be delivered to respective destinations. Traffic between STAs within the BSS may be sent through the AP, for example, where the source STA may send traffic to the AP and the AP may deliver the traffic to the destination STA The traffic between STAs within a BSS may be considered and/or referred to as peer-to-peer traffic. The peer-to- peer traffic may be sent between (e.g., directly between) the source and destination STAs with a direct link setup (DLS). In certain representative embodiments, the DLS may use an 802.11e DLS or an 802.11z tunneled DLS (TDLS). A WLAN using an Independent BSS (IBSS) mode may not have an AP, and the STAs (e.g., all of the STAs) within or using the IBSS may communicate directly with each other. The IBSS mode of communication may sometimes be referred to herein as an “ad-hoc” mode of communication.

[0058] When using the 802.11 ac infrastructure mode of operation or a similar mode of operations, the AP may transmit a beacon on a fixed channel, such as a primary channel. The primary channel may be a fixed width (e.g., 20 MHz wide bandwidth) or a dynamically set width. The primary channel may be the operating channel of the BSS and may be used by the STAs to establish a connection with the AP. In certain representative embodiments, Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) may be implemented, for example in 802.11 systems. For CSMA/CA, the STAs (e.g., every STA), including the AP, may sense the primary channel. If the primary channel is sensed/detected and/or determined to be busy by a particular STA, the particular STA may back off. One STA (e.g., only one station) may transmit at any given time in a given BSS.

[0059] High Throughput (HT) STAs may use a 40 MHz wide channel for communication, for example, via a combination of the primary 20 MHz channel with an adjacent or nonadjacent 20 MHz channel to form a 40 MHz wide channel.

[0060] Very High Throughput (VHT) STAs may support 20MHz, 40 MHz, 80 MHz, and/or 160 MHz wide channels The 40 MHz, and/or 80 MHz, channels may be formed by combining contiguous 20 MHz channels. A 160 MHz channel may be formed by combining 8 contiguous 20 MHz channels, or by combining two non- contiguous 80 MHz channels, which may be referred to as an 80+80 configuration. For the 80+80 configuration, the data, after channel encoding, may be passed through a segment parser that may divide the data into two streams. Inverse Fast Fourier Transform (IFFT) processing, and time domain processing, may be done on each stream separately The streams may be mapped on to the two 80 MHz channels, and the data may be transmitted by a transmitting STA. At the receiver of the receiving STA, the above described operation for the 80+80 configuration may be reversed, and the combined data may be sent to the Medium Access Control (MAC).

[0061] Sub 1 GHz modes of operation are supported by 802.11 af and 802.11 ah. The channel operating bandwidths, and carriers, are reduced in 802.11 af and 802.11ah relative to those used in 802.11n, and 802.11ac. 802.11 af supports 5 MHz, 10 MHz, and 20 MHz bandwidths in the TV White Space (TVWS) spectrum, and 802.11 ah supports 1 MHz, 2 MHz, 4 MHz, 8 MHz, and 16 MHz bandwidths using non-TVWS spectrum. According to a representative embodiment, 802.11 ah may support Meter Type Control/Machine- Type Communications (MTC), such as MTC devices in a macro coverage area. MTC devices may have certain capabilities, for example, limited capabilities including support for (e.g , only support for) certain and/or limited bandwidths The MTC devices may include a battery with a battery life above a threshold (e.g., to maintain a very long battery life).

[0062] WLAN systems, which may support multiple channels, and channel bandwidths, such as 802 11 n, 802.11ac, 802.11af, and 802.11 ah, include a channel which may be designated as the primary channel. The primary channel may have a bandwidth equal to the largest common operating bandwidth supported by all STAs in the BSS. The bandwidth of the primary channel may be set and/or limited by a STA, from among all STAs in operating in a BSS, which supports the smallest bandwidth operating mode. In the example of 802.11ah, the primary channel may be 1 MHz wide for STAs (e.g., MTC type devices) that support (e.g., only support) a 1 MHz mode, 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 sensing and/or Network Allocation Vector (NAV) settings may depend on the status of the primary channel. If the primary channel is busy, for example, due to a STA (which supports only a 1 MHz operating mode) transmitting to the AP, all available frequency bands may be considered busy even though a majority of the available frequency bands remains idle.

[0063] In the United States, the available frequency bands, which may be used by 802.11 ah, are from 902 MHz to 928 MHz. In Korea, the available frequency bands are from 917.5 MHz to 923.5 MHz. In Japan, the available frequency bands are from 916.5 MHz to 927.5 MHz. The total bandwidth available for 802.11 ah is 6 MHz to 26 MHz depending on the country code.

[0064] FIG. 1 D is a system diagram illustrating the RAN 104 and the CN 106 according to an embodiment. As noted above, the RAN 104 may employ an NR radio technology to communicate with the WTRUs 102a, 102b, 102c over the air interface 116. The RAN 104 may also be in communication with the CN 106.

[0065] The RAN 104 may include gNBs 180a, 180b, 180c, though it will be appreciated that the RAN 104 may include any number of gNBs while remaining consistent with an embodiment. The gNBs 180a, 180b, 180c may each include one or more transceivers for communicating with the WTRUs 102a, 102b, 102c over the air interface 116. In one embodiment, the gNBs 180a, 180b, 180c may implement MIMO technology. For example, gNBs 180a, 108b may utilize beamforming to transmit signals to and/or receive signals from the gNBs 180a, 180b, 180c. Thus, the gNB 180a, for example, may use multiple antennas to transmit wireless signals to, and/or receive wireless signals from, the WTRU 102a. In an embodiment, the gNBs 180a, 180b, 180c may implement carrier aggregation technology. For example, the gNB 180a may transmit multiple component carriers to the WTRU 102a (not shown). A subset of these component carriers may be on unlicensed spectrum while the remaining component carriers may be on licensed spectrum. In an embodiment, the gNBs 180a, 180b, 180c may implement Coordinated Multi-Point (CoMP) technology. For example, WTRU 102a may receive coordinated transmissions from gNB 180a and gNB 180b (and/or gNB 180c).

[0066] The WTRUs 102a, 102b, 102c may communicate with gNBs 180a, 180b, 180c using transmissions associated with a scalable numerology. For example, the OFDM symbol spacing and/or OFDM subcarrier spacing may vary for different transmissions, different cells, and/or different portions of the wireless transmission spectrum. The WTRUs 102a, 102b, 102c may communicate with gNBs 180a, 180b, 180c using subframe or transmission time intervals (TTIs) of various or scalable lengths (e.g., containing a varying number of OFDM symbols and/or lasting varying lengths of absolute time).

[0067] The gNBs 180a, 180b, 180c may be configured to communicate with the WTRUs 102a, 102b, 102c in a standalone configuration and/or a non-standalone configuration. In the standalone configuration, WTRUs 102a, 102b, 102c may communicate with gNBs 180a, 180b, 180c without also accessing other RANs (e.g., such as eNode-Bs 160a, 160b, 160c). In the standalone configuration, WTRUs 102a, 102b, 102c may utilize one or more of gNBs 180a, 180b, 180c as a mobility anchor point In the standalone configuration, WTRUs 102a, 102b, 102c may communicate with gNBs 180a, 180b, 180c using signals in an unlicensed band. In a non-standalone configuration WTRUs 102a, 102b, 102c may communicate with/connect to gNBs 180a, 180b, 180c while also communicating with/connecting to another RAN such as eNode-Bs 160a, 160b, 160c For example, WTRUs 102a, 102b, 102c may implement DC principles to communicate with one or more gNBs 180a, 180b, 180c and one or more eNode-Bs 160a, 160b, 160c substantially simultaneously. In the non- standalone configuration, eNode-Bs 160a, 160b, 160c may serve as a mobility anchor for WTRUs 102a, 102b, 102c and gNBs 180a, 180b, 180c may provide additional coverage and/or throughput for servicing WTRUs 102a, 102b, 102c.

[0068] Each of the gNBs 180a, 180b, 180c may be associated with a particular cell (not shown) and may be configured to handle radio resource management decisions, handover decisions, scheduling of users in the UL and/or DL, support of network slicing, DC, interworking between NR and E-UTRA, routing of user plane data towards User Plane Function (UPF) 184a, 184b, routing of control plane information towards Access and Mobility Management Function (AMF) 182a, 182b and the like. As shown in FIG. 1D, the gNBs 180a, 180b, 180c may communicate with one another over an Xn interface. [0069] The CN 106 shown in FIG. 1 D 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. While the foregoing elements are depicted as part of the CN 106, it will be appreciated that any of these elements may be owned and/or operated by an entity other than the CN operator.

[0070] The AMF 182a, 182b may be connected to one or more of the gNBs 180a, 180b, 180c in the RAN 104 via an N2 interface and may serve as a control node. For example, the AMF 182a, 182b may be responsible for authenticating users of the WTRUs 102a, 102b, 102c, support for network slicing (e.g., handling of different protocol data unit (PDU) sessions with different requirements), selecting a particular SMF 183a, 183b, management of the registration area, termination of non-access stratum (NAS) signaling, mobility management, and the like. Network slicing may be used by the AMF 182a, 182b in order to customize CN support for WTRUs 102a, 102b, 102c based on the types of services being utilized WTRUs 102a, 102b, 102c. For example, different network slices may be established for different use cases such as services relying on ultra-reliable low latency (URLLC) access, services relying on enhanced massive mobile broadband (eMBB) access, services for MTC access, and the like The AMF 182a, 182b may provide a control plane function for switching between the RAN 104 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.

[0071] The SMF 183a, 183b may be connected to an AMF 182a, 182b in the CN 106 via an N11 interface. The SMF 183a, 183b may also be connected to a UPF 184a, 184b in the CN 106 via an N4 interface. The SMF 183a, 183b may select and control the UPF 184a, 184b and configure the routing of traffic through the UPF 184a, 184b. The SMF 183a, 183b may perform other functions, such as managing and allocating UE IP address, managing PDU sessions, controlling policy enforcement and QoS, providing DL data notifications, and the like. A PDU session type may be IP-based, non-IP based, Ethernet-based, and the like.

[0072] The UPF 184a, 184b may be connected to one or more of the gNBs 180a, 180b, 180c in the RAN 104 via an N3 interface, which may provide the WTRUs 102a, 102b, 102c with access to packet-switched networks, such as the Internet 110, to facilitate communications between the WTRUs 102a, 102b, 102c and IP-enabled devices. The UPF 184, 184b may perform other functions, such as routing and forwarding packets, enforcing user plane policies, supporting multi-homed PDU sessions, handling user plane QoS, buffering DL packets, providing mobility anchoring, and the like.

[0073] The CN 106 may facilitate communications with other networks For example, the CN 106 may include, or may communicate with, an IP gateway (e.g., an IP multimedia subsystem (IMS) server) that serves as an interface between the CN 106 and the PSTN 108. In addition, the CN 106 may provide the WTRUs 102a, 102b, 102c with access to the other networks 112, which may include other wired and/or wireless networks that are owned and/or operated by other service providers In one embodiment, the WTRUs 102a, 102b, 102c may be connected to a local DN 185a, 185b through the UPF 184a, 184b via the N3 interface to the UPF 184a, 184b and an N6 interface between the UPF 184a, 184b and the DN 185a, 185b. [0074] In view of FIGs. 1A-1 D, and the corresponding description of FIGs. 1A-1 D, one or more, or all, of the functions described herein with regard to one or more of: WTRU 102a-d, Base Station 114a-b, eNode-B 160a-c, MME 162, SGW 164, PGW 166, gNB 180a-c, AMF 182a-b, UPF 184a-b, SMF 183a-b, DN 185a-b, and/or any other device(s) described herein, may be performed by one or more emulation devices (not shown). The emulation devices may be one or more devices configured to emulate one or more, or all, of the functions described herein. For example, the emulation devices may be used to test other devices and/or to simulate network and/or WTRU functions.

[0075] The emulation devices may be designed to implement one or more tests of other devices in a lab environment and/or in an operator network environment. For example, the one or more emulation devices may perform the one or more, or all, functions while being fully or partially implemented and/or deployed as part of a wired and/or wireless communication network in order to test other devices within the communication network. The one or more emulation devices may perform the one or more, or all, functions while being temporarily implemented/deployed as part of a wired and/or wireless communication network. The emulation device may be directly coupled to another device for purposes of testing and/or performing testing using over-the-air wireless communications.

[0076] The one or more emulation devices may perform the one or more, including all, functions while not being implemented/deployed as part of a wired and/or wireless communication network. For example, the emulation devices may be utilized in a testing scenario in a testing laboratory and/or a non-deployed (e.g., testing) wired and/or wireless communication network in order to implement testing of one or more components. The one or more emulation devices may be test equipment. Direct RF coupling and/or wireless communications via RF circuitry (e g., which may include one or more antennas) may be used by the emulation devices to transmit and/or receive data. The term wireless transmit/receive unit (WTRU) may be used interchangeably with the term user equipment (UE). Accordingly, anywhere UE is used, that terms means a WTRU, and vice versa. Additionally, the following acronyms may be used in this description. This list is not intended to be exhaustive and other acronyms may be utilized Furthermore, the acronyms may have alternate or other usages that will be apparent in context.

[0077] Af Sub-carrier spacing

[0078] gNB NR NodeB

[0079] AP Aperiodic; Access Point

[0080] BFR Beam Failure Recovery

[0081] BFD-RS Beam Failure Detection-Reference Signal

[0082] BLER Block Error Rate

[0083] BWP Bandwidth Part

[0084] CA Carrier Aggregation

[0085] CB Contention-Based (e.g. access, channel, resource) [0086] CDM Code Division Multiplexing

[0087] CG Cell Group

[0088] CoMP Coordinated Multi-Point transmission/reception

[0089] CP Cyclic Prefix

[0090] CPE Common Phase Error

[0091] CP-OFDM Conventional OFDM (relying on cyclic prefix)

[0092] CQI Channel Quality Indicator

[0093] CN Core Network (e.g. LTE packet core or NR core)

[0094] CRC Cyclic Redundancy Check

[0095] CSI Channel State Information

[0096] CSI-RS Channel State Information-Reference Signal

[0097] CU Central Unit

[0098] D2D Device to Device transmissions (e.g. LTE Sideiink)

[0099] DC Dual Connectivity

[0100] DCI Downlink Control Information

[0101] DL Downlink

[0102] DM-RS Demodulation Reference Signal

[0103] DRB Data Radio Bearer

[0104] DU Distributed Unit

[0105] EN-DCE-UTRA - NR Dual Connectivity

[0106] EPC Evolved Packet Core

[0107] FD-CDM Frequency Domain-Code Division Multiplexing

[0108] FDD Frequency Division Duplexing

[0109] FDM Frequency Division Multiplexing

[0110] ICI Inter-Cell Interference

[0111] ICIC Inter-Cell Interference Cancellation

[0112] IP Internet Protocol

[0113] LBT Listen-Before-Talk

[0114] LCH Logical Channel

[0115] LCID Logical Channel Identity

[0116] LCP Logical Channel Prioritization

[0117] LLC Low Latency Communications [0118] LoS Line of Sight [0119] LTE Long Term Evolution e.g. from 3GPP LTE R8 and up [0120] MAC Medium Access Control [0121] MAC CE Medium Access Control Control Element [0122] NACK Negative ACK [0123] MBMS Multimedia Broadcast Multicast System [0124] MCG Master Cell Group [0125] MCS Modulation and Coding Scheme [0126] MIMO Multiple Input Multiple Output [0127] MTC Machine-Type Communications [0128] MR-DC Multi-RAT Dual Connectivity [0129] NAS Non-Access Stratum [0130] NCB-RS New candidate beam-Reference Signal [0131] NE-DC NR-RAN - E-UTRA Dual Connectivity [0132] NR New Radio [0133] NR-DC Dual Connectivity with New Radio [0134] OFDM Orthogonal Frequency-Division Multiplexing [0135] OOB Out-Of-Band (emissions) [0136] P_cmax Total available UE power in a given transmission interval [0137] Pcell Primary cell of Master Cell Group [0138] PCG Primary Cell Group [0139] PDU Protocol Data Unit [0140] PER Packet Error Rate [0141] PHY Physical Layer [0142] PLMN Public Land Mobile Network [0143] PLR Packet Loss Rate [0144] PRACH Physical Random Access Channel [0145] PRB Physical Resource Block [0146] PRS Positioning Reference Signal [0147] Pscell Primary cell of a Secondary cell group [0148] PSS Primary Synchronization Signal [0149] PT-RS Phase T racking-Reference Signal [0150] QoS Quality of Service (from the physical layer perspective)

[0151] RAB Radio Access Bearer

[0152] RAN PA Radio Access Network Paging Area

[0153] RACH Random Access Channel (or procedure)

[0154] RAR Random Access Response

[0155] RAT Radio Access Technology

[0156] RB Resource Block

[0157] RCU Radio access network Central Unit

[0158] RF Radio Front end

[0159] RE Resource Element

[0160] RLF Radio Link Failure

[0161] RLM Radio Link Monitoring

[0162] RNTI Radio Network Identifier

[0163] ROM Read-Only Mode (for MBMS)

[0164] RRC Radio Resource Control

[0165] RRM Radio Resource Management

[0166] RS Reference Signal

[0167] RTT Round-Trip Time

[0168] SCG Secondary Cell Group

[0169] SCMA Single Carrier Multiple Access

[0170] SCS Sub-Carrier Spacing

[0171] SDU Service Data Unit

[0172] SOM Spectrum Operation Mode

[0173] SP Semi-persistent

[0174] SpCell Primary cell of a master or secondary cell group.

[0175] SRB Signaling Radio Bearer

[0176] SS Synchronization Signal

[0177] SRS Sounding Reference Signal

[0178] SSS Secondary Synchronization Signal

[0179] SUL Supplementary UpLink

[0180] SWG Switching Gap (in a self-contained subframe)

[0181] TB Transport Block [0182] TBS Transport Block Size

[0183] TCI Transmission Configuration Index

[0184] TDD Time-Division Duplexing

[0185] TDM Time-Division Multiplexing

[0186] Tl Time Interval (in integer multiple of one or more symbols)

[0187] TTI Transmission Time Interval (in integer multiple of one or more symbols)

[0188] TRP Transmission / Reception Point

[0189] TRPG Transmission / Reception Point Group

[0190] TRS Tracking Reference Signal

[0191] TRx Transceiver

[0192] UL Uplink

[0193] URC Ultra-Reliable Communications

[0194] URLLC Ultra-Reliable and Low Latency Communications

[0195] V2X Vehicular communications

[0196] WLAN Wireless Local Area Networks and related technologies (IEEE 8O2.xx domain)

[0197] To provide the benefits of next generation wireless communications systems and predict changing parameters and configurations so that they can be applied in advance of congestion, interference, or other such conditions, implementations of the systems and methods discussed herein utilize artificial intelligence/machine learning (AI/ML) systems. These AI/ML systems may be based on trained models, enabling prediction of changing environmental conditions and adjustment of beamforming parameters or channel state information that can be applied ideally prior to any adverse effects experienced by the communicating devices. However, different AI/ML models may provide better performance depending on changing conditions. For example, an AI/ML model optimized for predicting beam management adjustments as devices or other entities move within an area may not necessarily make accurate predictions when interference from a local airport radar system is present.

[0198] Accordingly, in some aspects, the present disclosure is directed to implementations of systems and methods for dynamically identifying or selecting AI/ML models or model types to provide optimized predictions based on changing environmental conditions, device movement, or combinations of these or other changes. In some implementations, trained models may be generated (and/or updated) by UEs, base stations (gNBs), or other network devices (e g. application servers, AI/ML processing nodes, etc.) and may be provided to the devices (e.g. UEs and gNBs or other devices) for execution, prediction of conditions or characteristics, and adjustment of parameters. In some implementations, these predictions may enable devices to apply parameter adjustments in advance of requests or notifications from partner devices, such as changing retransmission window parameters immediately in response to packet loss rate changes without having to first exchange configuration change details, or other such modifications In some implementations, the devices (e.g. UEs, gNBs or other devices) may monitor conditions or performance of the network and may dynamically determine whether to change AI/ML models, and if so, what model to utilize. This may significantly reduce management overhead, and enable faster mitigation or recovery from changing channel conditions or issues.

[0199] In some embodiments disclosed herein, an AI/ML model selection/reselection method based on dynamic AI/ML model determination is disclosed. A UE is configured with one or more AI/ML models with model IDs, wherein each AI/ML model is associated with an AI/ML model type (e.g., BM or CSI), one or more RSs to measure qualities for AI/ML model reselection, one or more RSs to reselect AI/ML model for BM and one or more RSs to reselect AI/ML model for CSI. The UE identifies need of AI/ML model reselection based on the following: If one or more of number of NACKs (e.g , number of consecutive NACKs > threshold), RSRP e.g., RSRP < threshold), change of UE speed (e.g., Current UE speed - Avg UE speed within a measurement window > threshold), change of UE position (e.g., Current UE position - Avg. UE position within a measurement window > threshold), change of pathloss (e.g., Current pathloss - Avg. pathloss within a measurement window > threshold), hypothetical PDCCH BLER < threshold, measured SINR or differential SI NR (e g., from scheduled DM RS ports) and etc.

[0200] The UE then identifies a type of AI/ML model for reselection based on the triggered conditions, such as the following: If measured RSRP > threshold and number of consecutive NACKs > threshold, reselect AI/ML model for CSI. If measured RSRP < threshold, reselect AI/ML model for CSI. If number of failed measured hypothetical PDCCH BLER < threshold, reselect AI/ML model for CSI. Otherwise, reselect AI/ML for BM.

[0201] The UE identifies an AI/ML model based on the conditions Alternatively, the UE activates a set of associated AI/ML models, of the one or more AI/ML models, with the currently identified AI/ML model/model type for reselection to evaluate prediction accuracy of the AI/ML models. -> details for the evaluation e.g., beam prediction accuracy/RSRP difference and RS transmission for the evaluation, different RS transmission for the activated models.. The UE indicates an AI/ML model type for reselection to a gNB and receives the one or more RSs associated with the reported type. The UE measures the one or more RSs and indicates a preferred model ID to the gNB based on the measurement (AI/ML model for CSI may be reselected after reselection of AI/ML model for BM if AI/ML model for BM is requested). The UE receives a confirmation of the UE request and activates the indicated model for the reported request type. Add UE behavior after activating/selecting/deciding AI/ML model.

[0202] In some embodiments, a CSI parameter switching is based on the associated AI/ML model. A UE is configured with one or more AI/ML models with model IDs, wherein each AI/ML model is associated/configured with a set of CSI parameters (potentially for each CSI type or CSI config e.g., for CSI and for BM). Or the UE indicates its capability (e.g., required CSI parameters for each AI/ML model). The UE identifies a preferred model, possibly with a model type indication (or CSI config ID), based on measuring one or more of configured thresholds, RSRP, UE speed, pathloss, zone ID, cell ID, TRP ID and etc. For example, if UE speed is < X and Pathloss > Y (e.g., indoor), then determines Model #1 If UE speed is < X and Pathloss < Y (e.g., walking outdoor), then determines Model #2. If UE speed is > X and Pathloss < Y (e.g., high speed outdoor), then determines Model #3.

[0203] Based on the determined model, the UE indicates a preferred model ID or a set of CSI parameters to a gNB. The UE receives a confirmation from the gNB (e.g., via CORESET/SS associated with the indication). The UE deactivates the previous model, activates the newly indicated model and apply the associated set of CSI parameters for next CSI report including a preferred size of measurement/prediction window. E.g., Model #1 : X number of beam IDs and corresponding L1-RSRPs for the current measurement (N1). Model #2: X1-1 number of beam IDs and corresponding L1-RSRPs for the current measurement and X1-2 number of beam IDs and corresponding L1-RSRPs for N1 +N2. Model #3: X2-1 , X2-2, X2-3 and X2-4 number of beam IDs and corresponding L1 -RSRPs for N 1/N2/N3/N4.

[0204] As discussed above, implementations of the systems and methods used herein utilize artificial intelligence and/or machine learning systems. Artificial intelligence may be broadly defined as the behavior exhibited by machines. Such behavior may e.g., mimic cognitive functions to sense, reason, adapt and act. Machine learning may refer to type of algorithms that solve a problem based on learning through experience (‘data’), without explicitly being programmed (‘configuring set of rules’). Machine learning can be considered as a subset of Al. Different machine learning paradigms may be envisioned based on the nature of data or feedback available to the learning algorithm. For example, a supervised learning approach may involve learning a function that maps input to an output based on labeled training example, wherein each training example may be a pair consisting of input and the corresponding output. For example, unsupervised learning approach may involve detecting patterns in the data with no pre-existing labels. For example, reinforcement learning approach may involve performing sequence of actions in an environment to maximize the cumulative reward. In some solutions, it is possible to apply machine learning algorithms using a combination or interpolation of the above- mentioned approaches. For example, semi-supervised learning approach may use a combination of a small amount of labeled data with a large amount of unlabeled data during training. In this regard semi-supervised learning falls between unsupervised learning (with no labeled training data) and supervised learning (with only labeled training data).

[0205] Deep learning refers to class of machine learning algorithms that employ artificial neural networks (specifically DNNs) which were loosely inspired from biological systems. The Deep Neural Networks (DNNs) are a special class of machine learning models inspired by human brain wherein the input is linearly transformed and pass-through non-linear activation function multiple times. DNNs typically consists of multiple layers where each layer consists of linear transformation and a given non-linear activation functions. The DNNs can be trained using the training data via back-propagation algorithm. Recently, DNNs have shown state-of- the-art performance in variety of domains, e.g., speech, vision, natural language, etc., and for various machine learning settings supervised, un-supervised, and semi-supervised. The term AI/ML based methods/processing may refer to realization of behaviors and/or conformance to requirements by learning based on data, without explicit configuration of sequence of steps of actions. Such methods may enable learning complex behaviors which might be difficult to specify and/or implement when using legacy methods.

[0206] In Rel-15, New Radio (NR) has introduced radio access technology (RAT) in frequency range 2 (FR2), where FR2 denotes the frequency range of 24 25 - 52.6 GHz. One of the key challenges of using FR2 is higher propagation loss. Since propagation loss increases as carrier frequency increases, FR2 experiences higher propagation loss than lower frequency range systems. In order to overcome the higher propagation loss, highly directional beamformed transmission and reception may be used to increase efficiency.

[0207] Beamforming gain can be achieved by adding or subtracting one signal from another signal. Since higher beamforming gain can be achieved as more signals are added or subtracted, a large number of antenna elements may be utilized for highly directional beamformed transmissions. Controlling signal addition or signal subtraction can be done by controlling phases of antenna elements.

[0208] Beamforming methods can be generally categorized into three types (e.g , analog beamforming, digital beamforming and hybrid beamforming) based on the phase controlling types. Figure 2 is a block diagram of a system for hybrid beamforming, according to some implementations The structure shown in Figure 2 may be implemented at either a base station or a WTRU/UE. While digital beamforming controls a phase of a signal by applying digital precoding, analog beamforming controls the phase of the signal through phase shifters. Generally, digital beamforming provides good flexibility (e.g., applying different phases for different frequency resource blocks), but requires more complex implementations. In contrast to digital beamforming, analog beamforming provides relatively simple implementations, but has limitations (e.g., same analog beam for all frequency resources). Given these trade-offs, hybrid beamforming is a good architecture to achieve large beamforming gains with reasonable implementation complexity. Hybrid beamforming provides enough flexibility with reasonable implementation complexity by combining analog beamforming and digital beamforming.

[0209] Since beam width of a beam decreases as beamforming gain increases, the beam can only cover a limited area. Therefore, the BS and the UE need to utilize multiple beams to cover the entire cell. For example, broadcast signals such as synchronization signal blocks (SSBs) can be transmitted along all directions (e.g., via beam sweeping) to cover the entire cell. For unicast transmissions between the BS and the UE, procedures to optimize the beam direction to the UE are provided through beam management (BM). Beam management may include selection and maintenance of the beam direction for unicast transmission (including control channel and/or data channel) between the BS and the UE or any other devices

[0210] Beam management procedures can be categorized into beam determination, beam measurement and reporting, beam switching, beam indication, and beam recovery. In beam determination, the BS and the UE or other devices find a beam direction to ensure good radio link quality for the unicast control and data channel transmission. Once a link is established, a device (e.g., the UE) measures the link quality of multiple transmission (TX) and reception (RX) beam pairs and reports the measurement results to the other device (e.g. BS). Furthermore, UE mobility, orientation, and channel blockages can change the radio link quality of TX and RX beam pairs. When the quality of the current beam pair degrades, the BS and the UE can switch to another beam pair with better radio link quality. To do so, the BS and/or the UE can monitor the quality of the current beam pair along with some other beam pairs and perform switching when necessary. When the BS assigns a TX beam to the UE via DL control signaling, a beam indication procedure may be used. Beam recovery entails a recovery procedure when a link between the BS and the UE can no longer be maintained.

[0211] In some implementations, AI/ML systems may be used for one or more of the following use cases: CSI feedback enhancement, e g., overhead reduction, improved accuracy, prediction; beam management, e.g., beam prediction in time, and/or spatial domain for overhead and latency reduction, beam selection accuracy improvement; and positioning accuracy enhancements for different scenarios including, e.g., those with heavy NLOS conditions. However, use of these systems may require additional implementation details discussed below.

[0212] Firstly, identification of adequate AI/ML model type for selection/reselection may be needed. If an AI/ML model for beam management is not working, then accurate prediction from AI/ML model for CSI does not provide proper performance as a best beam is not properly selected. In addition, identification of adequate AI/ML model, including whether the model can be generalized to difference conditions or measurements, may be needed.

[0213] The following cases are considered for verifying the generalization performance of an AI/ML model over various scenarios/configurations as a starting point:

• Case 1 : The AI/ML model is trained based on training dataset from one Scenario#A/Configuration#A, and then the AI/ML model performs inference/test on a dataset from the same Scenario#A/Configuration#A.

• Case 2: The AI/ML model is trained based on training dataset from one Scenario#A/Configuration#A, and then the AI/ML model performs inference/test on a different dataset than Scenario#A/Configuration#A, e.g., Scenario#B/Configuration#B, Scenario#A/Configuration#B.

• Case 3: The AI/ML model is trained based on training dataset constructed by mixing datasets from multiple scenarios/configurations including Scenario#A/Configuration#A and a different dataset than Scenario#A/Configuration#A, e.g., Scenario#B/Configuration#B, Scenario#A/Configuration#B, and then the AI/ML model performs inference/test on a dataset from a single Scenario/Configuration from the multiple scenarios/configurations, e.g., Scenario#A/Configuration#A, Scenario#B/Configuration#B, Scenario#A/Configuration#B. It is noted that companies to report the ratio for dataset mixing, and the number of the multiple scenarios/configurations can be larger than two.

[0214] In some implementations, he following case for generalization verification, can be optionally considered:

• Case 2A: The AI/ML model is trained based on training dataset from one Scenario#A/Configuration#A, and then the AI/ML model is updated based on a fine-tuning dataset different than Scenario#A/Configuration#A, e.g., Scenario#B/Configuration#B, Scenario#A/Configuration#B. After that, the AI/ML model is tested on a different dataset than Scenario#A/Configuration#A, e.g., subject to Scenario#B/Configuration#B, Scenario#A/Configuration#B. The company may report the fine-tuning dataset setting (e.g., size of dataset) and the improvement of performance. The feasibility of fine-tuning on the UE/Network side is for future study.

[0215] According to evaluation results, in general, Case 3 (training by using data set with mixed scenario/configuration) shows better generalization performance than Case 2 (training by using data set with difference scenario/configuration). In addition, the performance of Case 3 is usually degraded as compared to the performance of Case 1 , however, in some cases, the performance of Case 3 outperforms the performance of Case 1.

[0216] However, training one AI/ML model with all possible scenarios and configurations may be too complex or time- and resource-consuming. Therefore, in some implementations, multiple AI/ML models may be configured and one AI/ML model may need to be selected for each AI/ML model type (e.g., BM or CSI).

[0217] Accordingly, one problem dealt with by this disclosure is how a UE identifies a need for selecting/reselecting an AI/ML model, a proper AI/ML model type, and a proper AI/ML model for the selected AI/ML model type.

[0218] Various solution implementations are provided herein for efficiently identifying a need of selection/reselection. In some implementations, these solutions may enable UE identification of a need for AI/ML model reselection. In a solution, a UE may identify need of AI/ML model reselection by measuring and evaluated parameters.

[0219] In other implementations, solutions may enable UE identification of a proper AI/ML model type for reselection. In a solution, a UE may identify an AI/ML model type based on measurements and evaluated parameters

[0220] In still other implementations, solutions may enable UE identification of a proper AI/ML model for the determined AI/ML model type. In a solution, a UE may identify an AI/ML model based on measurements and evaluated parameters.

[0221] In still other implementations, solutions may enable UE indication of AI/ML model type and AI/ML model for reselection. In a solution, a UE may indicate AI/ML model type ID and AI/ML model ID for reselection as a part of CSI reporting

[0222] In yet still other implementations, solutions may enable CSI reporting parameter determination based on the determined AI/ML model type and AI/ML model ID In a solution, a UE may determine a set of CSI parameters to be reported based on the determined AI/ML model type and AI/ML model ID.

[0223] Implementations of these solutions may be combined in any set, without limitation.

[0224] Referring briefly to FIG. 3, illustrated is a block diagram of a system 300 for dynamic model selection for wireless communications, according to some implementations. System 300 may comprise a UE or WTRU 102, base station 114, etc., as discussed above, or any other such device System 300 may comprise one or more processors 118, memory devices 130, 132, antennas or antenna arrays 122, and analog and/or digital beam formers, as discussed above in connection with FIG. 2. [0225] In some implementations, system 300 may comprise a monitor 302. Monitor 302 may comprise an application, service, server, daemon, routine, or other executable logic for measuring and/or monitoring channel characteristics (e.g. noise, received signal strength, interference, packet loss rates, block error rates, etc.) and/or physical or electrical characteristics (e.g. position of system 300, change in position or velocity of system 300, change in velocity or acceleration of system 300, change in acceleration or jerkiness of system 300, orientation of system 300 or antenna arrays 122, angle of beam arrival, angle of beam departure, temperature, power usage, etc.) Monitor 302 may comprise hardware, software, or a combination of hardware and software, such as software embodied in an ASIC or FPGA, hardware sensors read by software processes, etc. Monitor 302 may be executed by one or more processors 118, and may make measurements periodically, randomly, constantly, in response to application or device requests, or any other schedule. Monitor 302 may store measurements in a logging database 304 or file, in any suitable format (e.g. array, flat file, indexed list, bitmap, etc.)

[0226] In some implementations, system 300 may comprise a selector 306. Selector 306 may comprise an application, service, server, daemon, routine, or other executable logic for determining whether a new AI/ML model should be selected, and if so, what model should be selected. Selector 306 may be embodied in hardware, software, or a combination of hardware and software. In some implementations, selector 306 may utilize measurements of channel characteristics or conditions and/or physical or electrical characteristics measured by monitor 302 and/or stored in a log 304. For example, in some implementations, selector 306 may compare measurements to one or more thresholds. In some implementations, selector 306 may comprise an AI/ML system, such as a DNN, decision tree, SVM, or any other type of AI/ML algorithm. For example, in some implementations, selector 306 may predict, based on past measurements of channel noise or signal strength, whether the channel is degrading and will likely become unusable in the future, and if so, whether a new AI/ML model should be utilized for CSI configuration or beam management. In other implementations, other measurements and/or combinations of measurements and thresholds may be utilized In many implementations, selector 306 and monitor 302 may be part of the same program or module Upon determining to select a new AI/ML model, selector 306 may select an AI/ML model from models stored in a database, file, or other storage 308. For example, models 308 may comprise a set of hyperparameters, weights, biases, neuron layer configurations, or any other such parameters of trained AI/ML models.

[0227] In some implementations, system 300 may execute one or more AI/ML engines 310. AI/ML engines 310 may comprise applications, services, servers, daemons, routines, or other executable logic for executing an AI/ML algorithm using a model 308. In some implementations, system 300 may execute a plurality of AI/ML engines 310 simultaneously. For example, upon determining to select a new AI/ML model, in some implementations discussed in more detail below, system 300 may execute a plurality of models simultaneously or in sequence and compare predictions to identify a model with a highest accuracy, sensitivity, selectivity, or quality, based on the present conditions or change in conditions. The system may then utilize the identified model and deactivate or disable the others until a subsequent time at which the selector 306 determines a new model is needed.

[0228] As discussed above, in some implementations, one or more processors 118 of system 300 may execute applications and/or AI/ML algorithms. In some implementations, system 300 may include one or more co-processors for executing AI/ML algorithms, such as a tensor processing unit (TPU), graphics processing unit (GPU), or other such processor. Such co-processors may be optimized for efficient or high-speed execution of AI/ML algorithms or particular types of models.

[0229] In some implementations, a UE may transmit or receive a physical channel or reference signal according to at least one spatial domain filter. The term “beam” may be used herein to refer to a spatial domain filter. The UE may transmit a physical channel or signal using the same spatial domain filter as the spatial domain filter used for receiving an RS (such as CSI-RS) or a SS block. The UE transmission may be referred to as “target”, and the received RS or SS block may be referred to as “reference” or “source”. In such case, the UE may be said to transmit the target physical channel or signal according to a spatial relation with a reference to such RS or SS block.

[0230] The UE may transmit a first physical channel or signal according to the same spatial domain filter as the spatial domain filter used for transmitting a second physical channel or signal. The first and second transmissions may be referred to as “target” and “reference” (or “source”), respectively. In such case, the UE may be said to transmit the first (target) physical channel or signal according to a spatial relation with a reference to the second (reference) physical channel or signal.

[0231] A spatial relation may be implicit, configured by RRC or signaled by MAC CE or DCI. For example, a UE may implicitly transmit PUSCH and DM-RS of PUSCH according to the same spatial domain filter as an SRS indicated by an SRI indicated in DCI or configured by RRC. In another example, a spatial relation may be configured by RRC for an SRS resource indicator (SRI) or signaled by MAC CE for a PUCCH. Such spatial relation may also be referred to as a “beam indication”.

[0232] The UE may receive a first (target) downlink channel or signal according to the same spatial domain filter or spatial reception parameter as a second (reference) downlink channel or signal. For example, such association may exist between a physical channel such as PDCCH or PDSCH and its respective DM-RS. At least when the first and second signals are reference signals, such association may exist when the UE is configured with a quasi-colocation (QCL) assumption type D between corresponding antenna ports. Such association may be configured as a TCI (transmission configuration indicator) state. A UE may be indicated an association between a CSI-RS or SS block and a DM-RS by an index to a set of TCI states configured by RRC and/or signaled by MAC CE. Such indication may also be referred to as a “beam indication”.

[0233] In this disclosure, the term “a RS resource set’ or similar terms may be interchangeably used with a resource configuration, a RS resource, and/or a beam group. Similarly, in this disclosure, the term “beam” may be interchangeably used with TCI state, TCI state group, and/or a beam pair; and beam reporting or similar terms may be interchangeably used with CSI measurement, CSI reporting, and/or beam measurement. Likewise, a “beam ID” may be interchangeably used with a beam index and/or a beam pair ID; and a reference signal (or beam reference signal) may be interchangeably used with one or more of following: a sounding reference signal (SRS), channel state information - reference signal (CSI-RS), demodulation reference signal (DM-RS), phase tracking reference signal (PT-RS), and/or a synchronization signal block (SSB). A “channel” (or a physical channel) may be interchangeably used with one or more of following: PDCCH, PDSCH, Physical uplink control channel (PUCCH), Physical uplink shared channel (PUSCH), Physical random access channel (PRACH), Physical sidelink control channel (PSCCH), Physical sidelink shared channel (PSSCH), Physical sidelink feedback channel (PSFCH), and/or Physical broadcasting channel (PBCH).

[0234] Referring now to FIG. 4, illustrated is a flow chart of a method 400 for dynamic model selection for wireless communications, according to some implementations. Method 400 may be executed by a UE, WTRU, base station, or other network node or device. In some implementations, a device may be configured with one or more models for AI/ML prediction, and/or may be configured with parameters for CSI and beam reporting. For example, in some embodiments, one or more of the following configurations may be used for CSI or beam reporting configuration as described below.

[0235] A UE may be configured with one or more CSI report configurations.

■ Report configuration type (e.g., periodic, semi-persistent on PUCCH, semi- persistent on PUSCH or aperiodic)

■ Report quantity (e.g., CRI-RI-PM l-CQI, CRI-RI-M , CRI-RI-i 1 -CQI, CRI-RSRP, SSB- Index-RSRP, CRI-RI-LI-PMI-CQI, CRI-SINR, SSB-lndex-SINR)

■ Report frequency configuration

• CQI format indicator (wideband CQI or subband CQI)

• PMI format indicator (wideband PMI or subband PMI)

• CSI reporting band

■ Time restriction for channel measurements

■ Time restriction for interference measurements

■ Codebook config

■ Group based beam reporting

■ CQI table

■ Subband size

■ Non-PMI port indication

■ Report slot config/offset list

■ CSI report periodicity and offset

■ One or more PUCCH resources for CSI reporting

■ Port Index

[0236] In some embodiments, one or more of following configurations may be used for measurement configuration of beam reporting: • A UE may be configured with one or more CSI measurement configurations o The CSI measurement configurations may include one or more of following:

■ RS for channel measurement

■ RS for interference measurement (zero power or non-zero power)

■ Report trigger size

■ Aperiodic trigger state list

■ Semi-persistent on PUSCH trigger state list

■ Associated CSI resource configs

■ Associated CSI report configs

[0237] In some embodiments, one or more of following configurations may be used for CSI resource configuration:

• A UE may be configured with one or more CSI resource configurations o The CSI resource configuration may include one or more of following:

■ CSI resource config ID

■ One or more RS resource sets for channel measurement

■ One or more RS resource sets for interference measurement

■ Bandwidth part ID

■ Resource type (e.g., aperiodic, semi-persistent or periodic)

[0238] In some embodiments, one or more of the following configurations may be used for RS resource set:

• A UE may be configured with one or more RS resource sets o The RS resource set configuration may include one or more of following:

■ RS resource set ID

■ One or more RS resources for the RS resource set

■ Repetition (i.e., on or off)

■ Aperiodic triggering offset (e.g., one of 0-6 slots)

■ TRS info (e.g., true or not)

[0239] In some embodiments, one or more of the following configurations may be used for RS resource:

• A UE may be configured with one or more RS resources o The RS resource configuration may include one or more of following:

■ RS resource ID

■ Resource mapping (e g., REs in a PRB)

■ Power control offset (e.g., one value of -8, ... , 15)

■ Power control offset with SS (e g., -3 dB, 0 dB, 3 dB, 6 Db)

■ Scrambling ID

■ Periodicity and offset QCL information (e.g., based on a TCI state)

[0240] In some embodiments, at step 402, a UE may indicate its capability for AI/ML models which the UE support. For example, in embodiments at 402, the UE may indicate one or more of the following:

• Supported model types: the UE may indicate one or more types of AI/ML model which the UE support. For example, the UE may indicate one or more of CSI, BM, positioning and etc. as supported AI/ML model types (or AI/ML model functionalities)

• Supported RS types: the UE may indicate one or more types of RSs which the UE support measurement. The indication may be separately indicated for AI/ML model reselection detection (i.e., detecting need of AI/ML model reselection), AI/ML model type evaluation (i.e., identifying a type of AI/ML model type for reselection) and AI/ML model reselection evaluation (i.e., evaluating qualities of AI/ML models). For example, one or more of the following RSs may be indicated as supported RS types:

• SRS (e.g., from other UEs)

• CSI-RS

• DM-RS

• PT-RS

• SSB

• Supported measurement types: a UE may indicate one or more types of measurements which the UE supports. The indication may be separately indicated for AI/ML model reselection detection (i.e., detecting need of AI/ML model reselection), AI/ML model type evaluation (i.e., identifying a type of AI/ML model type for reselection) and AI/ML model reselection evaluation (i.e., evaluating qualities of AI/ML models). For example, one or more of the following measurement types may be indicated as supported RS types:

• RSRP or L1-RSRP

• RSRQ or L1-RSRQ

• SINR or L1-SINR

• Hypothetical PDCCH BLER

• Hypothetical PUCCH BLER

• CQI

• PDSCH ACK/NACK

• Supported AI/ML models: a UE may indicate one or more AI/ML models which the UE supports. For example, the UE may indicate one or more of the following:

• BM for only spatial domain prediction

• BM for both spatial/temporal domain prediction

• BM for only temporal domain prediction

• CSI compression without temporal prediction

• CSI compression with temporal prediction [0241] In some implementations at step 402, the UE may indicate a type of CSI parameter set which the UE supports. For example, the UE may indicate one or more of the following:

• Set #1 (BM for only spatial domain prediction): X number of beam IDs and corresponding L1-RSRPs for the current measurement (N1).

• Set #2 (BM for both spatial/temporal domain prediction) : X1-1 number of beam IDs and corresponding L1-RSRPs for the current measurement and X1-2 number of beam IDs and corresponding L1-RSRPs for N1 +N2.

• Set #3 (BM for only temporal domain prediction): X2-1 , X2-2, X2-3 and X2-4 number of beam IDs and corresponding L1-RSRPs for N1/N2/N3/N4

• Set #4 (CSI compression without temporal prediction): Rl, X3-1 number of PMIs and corresponding CQIs for each codeword.

• Set #5 (CSI compression with temporal prediction): Rl#1 , X3-1 number of PMIs and corresponding CQIs for each codeword for the current measurement (N 1 ) and Rl#2, X3-2 number of PMIs and corresponding CQIs for each codeword for N1+N3.

[0242] The capability reporting may be via a CSI report or other management report, in various implementations.

[0243] The term “AI/ML model functionality” may be interchangeably used with AI/ML model type, AI/ML model use case, and AI/ML model function, or similar terms. Likewise, “AI/ML model reselection” may be interchangeably used with AI/ML model switching, AI/ML model determination, AI/ML model activation/deactivation, and AI/ML model fallback.

[0244] In some embodiments, a UE may be pre-configured with one or more AI/ML models with model IDs or, in other embodiments, at step 404 may receive one or more AI/ML models and corresponding model IDs. For example, in some implementations, AI/ML models may be pre-loaded into memory of a UE or other device. In other implementations, AI/ML models may be downloaded or received via a broadcast or other transmission (e g. from a base station, network node, application server, etc.). For example, in some implementations, AI/ML models may be periodically updated by an application server performing continued or periodic training of the AI/ML models based on new data from other network devices. Accordingly, in some implementations, AI/ML models may be received periodically or in response to transmission of capabilities at step 402. In some embodiments, the AI/ML configurations received may be based on the reported UE capability (e.g. a subset of AI/ML models may be provided to a UE, based on its capabilities, thus reducing network bandwidth and memory required to transmit and store models not usable or beyond the capabilities of the UE). AI/ML model configurations may include one or more of the following:

• Model type ID: model type ID may indicate a type of AI/ML model. For example, the UE may be configured with one of CSI, BM, positioning, etc.. • Model ID: model ID may indicate an ID of AI/ML model. For example, the model ID may be an ID among AI/ML models for each AI/ML model type. In another example, the model ID may be an ID among AI/ML models for all AI/ML model types.

• Associated set of CSI parameters/CSI parameter type: each AI/ML model may be associated with a set of CSI parameters or CSI parameter type. For example, the following CSI parameter set may be defined.

• Set #1 (BM for only spatial domain prediction): X number of beam IDs and corresponding L1-RSRPs for the current measurement (N1).

• Set #2 (BM for both spatial/temporal domain prediction) : X1-1 number of beam IDs and corresponding L1-RSRPs for the current measurement and X1-2 number of beam IDs and corresponding L1-RSRPs for N1 +N2.

• Set #3 (BM for only temporal domain prediction): X2-1, X2-2, X2-3 and X2-4 number of beam IDs and corresponding L1-RSRPs for N1/N2/N3/N4

• Set #4 (CSI compression without temporal prediction): Rl, X3-1 number of PMIs and corresponding CQIs for each codeword.

• Set #5 (CSI compression with temporal prediction): Rl#1 , X3-1 number of PMIs and corresponding CQIs for each codeword for the current measurement (N 1 ) and Rl#2, X3-2 number of PMIs and corresponding CQIs for each codeword for N1+N3.

Configurations of CSI parameter sets may be based on one or more of CSI parameter set ID, CSI parameter indication (e.g., RI-PMI-CQI) and etc. The associated set of CSI parameters with an AI/ML model may be predefined for each AI/ML model. For example, when the UE indicates BM for only spatial domain prediction, Set #1 may be predefined as an associated CSI parameter set.

• Associated set of RS resources and/or RS resource sets: in some embodiments, each AI/ML model or AI/ML model type may be associated with a set of RS resources or RS resource sets. For example, the UE may be configured with a first set of RS resources/resource sets associated with a first AI/ML model and a second set of RS resources/resource sets associated with a second AI/ML model. In another example, the UE may be configured with a first set of RS resources/resource sets associated with a first AI/ML model type and a second set of RS resources/resource sets associated with a second AI/ML model type.

[0245] In one embodiment, a required RS resource type may be determined based on an AI/ML model type. For example, one or more RS resources may be configured for a first AI/ML model type (e.g., for CSI). One or more RS resource set may be configured for a second AI/ML model type (e.g., for BM or positioning). In another embodiment, the configuration may be separately indicated for AI/ML model reselection detection (i.e., detecting need of AI/ML model reselection), AI/ML model type evaluation (i.e., identifying a type of AI/ML model type for reselection) and AI/ML model reselection evaluation (i.e., evaluating qualities of AI/ML models). For example, SSB, DMRS or PT-RS may be configured for AI/ML model reselection detection and SSB orCSI- RS may be configured for AI/ML model reselection evaluation. [0246] A UE may need to reselect AI/ML model based on the performance of the AI/ML model it is currently using or based on the detection of the change of one or more parameters or measurements that may indicate or affect the performance of AI/ML model. To this end, the UE may use one or more of the following solutions to determine the need of AI/ML model reselection.

[0247] In some embodiments, at step 406, the UE may measure one or more signals and channels for detecting the need of AI/ML model reselection. The UE may use different types of signals and channels for AI/ML model reselection detection (i.e., detecting need of AI/ML model reselection), AI/ML model type evaluation (i.e., identifying a type of AI/ML model type for reselection) and AI/ML model reselection evaluation (i.e., evaluating qualities of AI/ML models) For example, the UE may use a first type of RS (e.g., DMRS or PT- RS) for the AI/ML model reselection detection, a second type of RS (e.g., SSB) for AI/ML model type evaluation and a third type of RS (e.g., CSI-RS) for AI/ML model reselection evaluation. At step 406, the UE may measure one or more of the following signals and channels for detecting the need of AI/ML model reselection:

• SRS (e.g., from other UEs)

• CSI-RS

• DM-RS

• PT-RS

• SSB

• PDCCH

• PDSCH

• PUCCH (e.g., from other UEs)

• PUSCH (e.g., from other UEs)

• PRACH (e.g., from other UEs)

[0248] The evaluation of an AI/ML model type or AI/ML model may be referred to as one or more of following:

• Performance monitoring of AI/ML model type or AI/ML model to determine whether AI/ML model type (or AI/ML mode) switching, activation/deactivation, or reselection is needed.

• A procedure triggered by one or more conditions (e.g., throughput performance degradation, latency increase, higher number of HARQ-NACK, packet failure, etc.)

[0249] In one embodiment, at step 408, the UE may determine the need for reselecting an AI/ML model upon the detection of one or more NACKs (unsuccessful reception of code blocks or transport blocks). For example, in some such embodiments, when the UE detects or identifies that a number of NACKs within a preconfigured time window received from the gNB (e.g., via RRC signaling or MAC-CE indication) exceeds a preconfigured/indicated threshold number (e.g , via RRC signaling or MAC-CE indication), at step 408, the UE may determine that AI/ML model reselection is needed. The time window may be configured to be a moving time window in the immediate past (e.g., x ms in the immediate past, y slots in the immediate past) from the time instance AI/ML model reselection is evaluated by the UE or from the time instance the last measurement was taken at step 406 (that is, as shown, steps 406-408 may be repeated iteratively or periodically until the measurements indicate that reselection is needed). The duration of the time window may be configured via one or more of RRC signaling, MAC-CE indication, and/or DCI indication.

[0250] In another embodiment, at step 408, the UE may determine the need for reselecting an AI/ML model based on the number of consecutive NACKs. For example, if the UE detects or identifies a number of consecutive NACKs that exceeds a preconfigured threshold number by the gNB (e.g., via RRC signaling or MAC-CE indication), at step 408, the UE may determine that AI/ML model reselection is needed.

[0251] In an embodiment, at step 408, the UE may determine a need for AI/ML model reselection based on one or more measurements (e.g., measurement associates with one or more RSs) made at step 406, or estimated quantities by using one or more measurements (e.g. via a first AI/ML model, such as a prediction model that predicts future channel conditions or likely measurements) The UE may be configured with one or more thresholds via RRC signaling, MAC-CE indication, or DCI indication to determine the need of AI/ML model reselection by comparing the measurements or estimated quantities. For example, if a measured quantity or an estimated quantity > a threshold, then at step 408 the UE may determine that AI/ML model reselection is needed. If the measured quantity or the estimated quantity < the threshold, then at step 408, the UE may determine that AI/ML model reselection is not needed. The measured or estimated qualities may include one or more of the following but not limited to:

• RSRP (e.g., RSRP associates with one or more configured RSs).

• Differential RSRP (e.g., differential RSRP associates with one or more configured RSs). For example, the UE may compute the differential RSRP (e.g., L1-RSRP) of a RS (e g., a beam failure detection RS) considering the measurements of two time measurement instances (e.g., two last measurements instances). The UE may determine that reselection of AI/ML model is needed if the computed differential RSRP > the preconfigured threshold by the gNB. The UE may determine that reselection of AI/ML model is not needed if the computed differential RSRP < the preconfigured threshold by the gNB.

• RSRQ (e.g., RSRQ associates with one or more configured RSs).

• SINR (e.g., SINR associates with one or more configured RSs).

• Differential SINR. For example, the UE may compute the differential SINR by using measurements associated with one or more configured RSs at two time instances. The UE may determine that reselection of AI/ML model is needed if the computed differential SINR > the preconfigured threshold by the gNB. The UE may determine that reselection of AI/ML model is not needed if the computed differential SINR < the preconfigured threshold by the gNB. • Speed of the UE

• Position of the UE

• Path loss

• Hypothetical PDCCH BLER

[0252] In another embodiment, at step 408, the UE may determine a need for AI/ML model reselection based on a change of LoS condition. For example, if the UE determines that the current LoS condition is different from the LoS condition when the current AI/ML model was selected, the UE may determine that AI/ML model reselection is needed.

[0253] In an embodiment, at step 408, the UE may determine the need of AI/ML model reselection by comparing one or more current measurements (e.g., measurement associated with one or more RSs) or current estimated quantities by using one or more of the current measurements made at step 406, with the respective average measurements or the average estimated quantities over a preconfigured time window (e.g. over past iterations of step 406).

[0254] The time window may be configured to be a moving time window that considers the immediate past (e g., x ms in the immediate past, y slots in the immediate past) from the time instance AI/ML model reselection is evaluated or from the time instance the last measurement was taken. The duration of the time window may be configured via RRC signaling, MAC-CE indication, or DCI indication.

[0255] The UE may be configured with one or more thresholds via RRC signaling, MAC-CE indication, and/or DCI indication.

[0256] In some implementations, the UE may compute the difference between the measurements or estimated quantities and the corresponding average measurements or the average estimated quantities (computed differential values), and then may compare the computed differential values with the preconfigured thresholds to determine a need for AI/ML model reselection at step 408.

[0257] For example, if the difference between measured quantity or the estimated quantity and the corresponding average measurements or the average estimated quantity > a threshold, then at step 408 the UE may determine that AI/ML model reselection is needed. If the difference between measured quantity or the estimated quantity and the corresponding average measurements or the average estimated quantity < a threshold, then at step 408 the UE may determine that AI/ML model reselection is not needed.

[0258] As discussed above, the measured or estimated qualities may include one or more of the following but not limited to:

• Speed of the UE: for example, if the difference between the current UE speed and average UE speed over the preconfigured measurement window > the preconfigured threshold, the UE may determine that AI/ML model reselection is needed. Otherwise, the UE may determine that the AI/ML model reselection is not needed.

• Position of the UE: for example, if the difference between the current position of the UE and average position of the UE over the preconfigured measurement window > the preconfigured threshold, the UE may determine that AI/ML model reselection is needed. Otherwise, the UE may determine that AI/ML model reselection is not needed.

• Path loss: for example, if the difference between the current path loss experienced by the UE and average path loss of the UE over the preconfigured measurement window > the preconfigured threshold, the UE may determine that AI/ML model reselection is needed. Otherwise, the UE may determine that AI/ML model reselection is not needed.

• Hypothetical PDCCH block error rate (BLER): for example, if the difference between the hypothetical BLER of the last received PDCCH and the average hypothetical BLER of the PDCCHs received over the preconfigured measurement window > the preconfigured threshold, the UE may determine that AI/ML model reselection is needed. Otherwise, the UE may determine that the AI/ML model reselection is not needed.

• SINR: for example, if the difference between the current SINR (e.g., measured SINR from a scheduled DMRS ports) and the average SINR over the preconfigured measurement window > the preconfigured threshold, the UE may determine that AI/ML model reselection is needed. Otherwise, the UE may determine that the AI/ML model reselection is not needed.

• RSRP: for example, if the difference between the current RSRP (e.g., L1-RSRP of a configured RS) and the average RSRP over the preconfigured measurement window > the preconfigured threshold, the UE may determine that AI/ML model reselection is needed. Otherwise, the UE may determine that the AI/ML model reselection is not needed.

• LoS probability: for example, if the difference between the current LoS probability and the average LoS probability over the preconfigured measurement window is higher than a threshold, the UE may determine that AI/ML model resection is needed.

[0259] Upon determining a need for AI/ML model reselection, in some implementations such as where selection of a new model is performed by the base station or gNB, an application server, an AI/ML server, or other network node, the UE may indicate the need of model reselection to the gNB To this end, the UE may use one or a combination of the following embodiments.

[0260] In an embodiment, the UE may indicate the determination of the need of reselecting AI/ML model to the gNB via PUCCH or PUSCH (MAC-CE) For example, the UE may indicate the need for reselecting an AI/ML model by a single bit transmission, where bit value T may indicate that AI/ML model reselection is needed. The bit value ‘O’ may indicate that AI/ML model reselection is not needed.

[0261] In another embodiment, at step 418, the UE may indicate the parameters, measurements, or soft information (e.g., difference between the current or the instantaneous RSRP and average RSRP within the configured time window, current RSRP, current LoS condition) along with its decision for the need of the AI/ML model reselection For example, if the indicated information (e.g., RSRP) < threshold or the indicated information (e.g., Pathloss > threshold), the UE may implicitly indicate the need of model reselection [0262] In another embodiment, the UE may be configured with one or more preambles. Each preamble corresponds to one or more preconfigured condition (e.g , the L1-RSRP of a configured RS falls below the preconfigure threshold, the difference between instantaneous L1-RSRP and the average L1-RSRP over the configured time window exceeds the preconfigured threshold, the change in the LoS condition etc.,) needed to be satisfied to determine the need for AI/ML model reselection. Upon the detection of the need for AI/ML model reselection at step 408, in some embodiments at step 418, the UE may transmit the preamble corresponding to one or more of the conditions used to determine the need of reselecting AI/ML model.

[0263] In an embodiment, the indication of the need for reselecting AI/ML model by the UE may trigger additional measurement procedures including RS transmissions to measure or estimate additional measurements (e.g., RSRP, speed, location, SINR, etc.,) to support AI/ML model reselection process. One or more RS resources may be associated RSs based on the AI/ML configuration.

[0264] In an embodiment, at step 420, the UE may monitor for a confirmation indication or configuration (e g., via a PDCCH indication possibly in an associated CORESETs/SearchSpace within a preconfigured monitoring window immediately after the indication of the need of reselecting AI/ML model to the gNB or via a MAC-CE indication). The UE may receive an indication or selection of a new AI/ML model for use at 420, and at step 422 may switch to using the new model.

[0265] In another embodiment, the UE may perform at least an initial selection. For example, at step 410, a UE may identify one or more model types (e.g., CSI, BM, positioning and etc.) for reselection of AI/ML model. The UE may be configured with one or more thresholds (e.g., via one or more of RRC, MAC CE and DCI). For example, if a measured quality > a threshold, then at step 410, the UE may determine a first type (e g., AI/ML model for CSI) for reselection. Otherwise, at step 410, the UE may determine a second type of AI/ML model (e g., AI/ML model for BM) for reselection. The measured quality may be one or more of the following:

• RSRP

• RSRQ

• SINR

• Hypothetical PDCCH/PUCCH BLER

• Pathloss

[0266] For example, and referring briefly ahead to FIG. 5, illustrated is a logic diagram 500 of an example implementation of dynamic model selection for wireless communications. Although one particular set of logic is illustrated, other configurations may be utilized in other implementations. The logic may be executed as a series of gates as shown, or may be executed via a decision tree or other structure. In the implementation illustrated, measurements made at step 406 may be compared to thresholds at step 408 to determine whether model reselection is required. If no measurement is above a threshold, then steps 406-408 may be repeated. If one or more measurements are above a threshold, then the logic maybe evaluated at step 410 For example, if measured RSRP > a first threshold and number of consecutive NACKs (or hypothetical PDCCH BLER) > a second threshold, at step 410, the UE may determine a first type of AI/ML model (e.g , reselecting AI/ML model for CSI 502). If measured RSRP < the first threshold (possible with number of consecutive NACKs (or hypothetical PDCCH BLER) > a second threshold), then at step 410, the UE may determine the second type of AI/ML model (e.g., reselecting AI/ML model for BM 504).

[0267] In another embodiment, if measured Pathloss < a first threshold and number of consecutive NACKs (or hypothetical PDCCH BLER) > a second threshold, at step 410, the UE may determine a first type of AI/ML model (e g., reselecting AI/ML model for CSI 502) If measured Pathloss > the first threshold (possible with number of consecutive NACKs (or hypothetical PDCCH BLER) > a second threshold), then at step 410, the UE may determine the second type of AI/ML model (e.g , reselecting AI/ML model for BM 504).

[0268] In another embodiment, if measured RSRP > a first threshold and measured RSRQ (or SINR) < a second threshold, at step 410, the UE may determine a first type of AI/ML model (e.g , reselecting AI/ML model for CSI 502). If measured RSRP < the first threshold and measured RSRQ (or SINR) < the second threshold, then at step 410, the UE may determine the second type of AI/ML model (e.g., reselecting AI/ML model for BM 504).

[0269] In another embodiment, if a number of failed measured hypothetical PDCCH BLER < a first threshold, at step 410, the UE may determine the first type of AI/ML model (e.g., reselecting AI/ML model for CSI 502). Otherwise (e.g., one or more of measured RSRP < a second threshold, Pathloss > a third threshold, measured RSRQ (or SINR) < a fourth threshold and etc.), at step 410, the UE may determine the second type of AI/ML model (e.g., reselecting AI/ML for BM 504).

[0270] Returning to FIG. 4, in an embodiment, a UE may be configured with a measurement type for AI/ML model type identification for AI/ML model reselection. Based on the indicated measurement type, the UE may use the measurement type for AI/ML model type identification. For example, the UE may be indicated one or more of the following:

• RSRP

• RSRQ

• SINR

• Hypothetical PDCCH/PUCCH BLER

• Pathloss

[0271] In an embodiment, at step 412, a UE may indicate the determined one or more AI/ML model types for AI/ML model reselection (e.g., to a gNB). The indication may be based on one or more of the following:

• PUCCH: in some implementations, the UE may indicate the determined AI/ML model types by using PUCCH. For example, the UE may be configured/indicated with one or more PUCCH resources (e.g., via one or more of RRC, MAC CE and DCI). If one PUCCH resource is configured for all AI/ML model types, the PUCCH resource may be used for all AI/ML model types. In this case, the UE may indicate AI/ML model type ID as a part of UCI. If one PUCCH resource is configured for each AI/ML model type, the UE may indicate AI/ML model type in a PUCCH resource associated with the determined AI/ML model type or the currently activated AI/ML model type. The PUCCH transmission may be one or more of scheduling request (SR), HARQ ACK/NACK report and CSI reporting.

• PUSCH: in some implementations, the UE may indicate the determined AI/ML model types by using PUSCH. For example, the UE may be configured/indicated with one or more PUSCH resources (e.g. , via one or more of RRC (e.g., configured grant), MAC CE and DCI (e.g., dynamic grant)). If the UE receives indications of both dynamic grant and configured grant, the UE may prioritize one grant For example, the UE may prioritize earlier PUSCH resources In another example, the UE may prioritize based on a first type (e.g., dynamic) than a second type (e.g., configured).

• PRACH: in some implementations, the UE may indicate the determined AI/ML model types by using PRACH. For example, the UE may be configured/indicated with one or more PRACH resources (e.g., via one or more of RRC, MAC CE and DCI). If one PRACH resource is configured for all AI/ML model types, the PRACH resource may be used for all AI/ML model types. In this case, a PRACH sequence may be associated with each AI/ML model type For example, if the UE may determine a first AI/ML model type for reselection, the UE may transmit a first PRACH sequence. If the UE may determine a second AI/ML model type for reselection, the UE may transmit a second PRACH sequence. If one PRACH resource is configured for each AI/ML model type, the UE may indicate AI/ML model type in a PRACH resource associated with the determined AI/ML model type or the currently activated AI/ML model type.

[0272] In an embodiment, the UE indication of the one or more AI/ML model types may trigger additional measurement procedures including RS transmission. For example, a first set of RS resources may be transmitted to the UE if the first type of AI/ML model is indicated. A second set of RS resources may be transmitted to the UE if the second type of AI/ML model is indicated. The association between the RS resources may be based on one or more of the following:

• Predefined RS resources: for example, the first RS resources and the second RS resources may be predefined for the first AI/ML model type and the second AI/ML model type, respectively.

• Configured RS resources: for example, the first RS resources may be configured as associated RS resources for the first AI/ML model type and the second RS resources may be configured as associated RS resources for the first AI/ML model type.

• Dynamic indication: for example, the UE may receive an indication of RS resources which may be transmitted for the indicated AI/ML model type. For example, the UE may receive an indication of one or more RS resources from a set of configured RS resources (e.g., via RRC and/or MAC CE).

[0273] In an embodiment, one or more parameters of the RS resources may be dynamically determined while other parameters are predefined and/or configured. For example, a parameter may be dynamically indicated (e.g., from one or more configured candidate parameters (e.g., via RRC and/or MAC CE). In another example, the UE may determine a parameter based on the indicated AI/ML model type. For example, the UE may determine a first parameter if the first Al/M L model type is indicated and a second parameter if the second AI/ML model type is indicated. The one or more parameters may be one or more of the following:

• RS resource offset or RS resource set offset (e.g., between the UE indication of AI/ML model type and the RS resources/the RS resource sets)

• RS Density

• Number of repetitions

• Periodicity

• Transmission type (e.g., periodic, semi-persistent or dynamic)

[0274] In some such embodiments, at step 412, the UE may send to the gNB the results of its determination of the AI/ML model for (re)selection (e.g., model ID) and/or the parameters used to identify the AI/ML model (e g., measurements, performance monitoring parameters, static parameters about the model, any other parameters, etc.). The UE may send an indication of a preferred model for (re)selection and/or the related measurements/parameters to the gNB via any of the following message types:

• RRC signaling and/or NAS messages (e.g., SRBO, SRB1, SRB2, SRB3, SRB4)

• UL MAC CE (e.g., existing MAC CE, new MAC CE, regular BSR, periodic BSR, padding BSR, enhanced BSR, pre-emptive BSR, etc.)

• UCI (e.g., single bit SR, multi-bit SR, feedback, ACK/NACK, CSI report)

• PUCCH

• PUSCH

• Application layer signaling/messages

[0275] In an embodiment, a UE identification of AI/ML model for (re)selection is based on measurements. A UE may be configured with resources to make measurements The UE may compare the measurement to thresholds received from the gNB. If the measurement is less or greater than the threshold, the UE may determine that it needs an AI/ML model and the preferred AI/ML model that is suitable for the UE. Measurements made by the UE for selecting an AI/ML model may include one or more of:

• L1 or L3 measurement such as RSRP, RSSI, RSRQ, SINR, Rl, CQI, PMI, LI

• Doppler, Doppler spread, delay spread, number of multipaths

• Channel coherence time, channel coherence bandwidth

• UE Speed, position, direction of motion, velocity

• Interference/pathloss measurements

• Number of ACKs/NACKs Beam measurements such as L1-RSRP, beam direction, beamwidth, beam ID, number of beam IDs, corresponding L1-RSRP of beams, etc.

• Whether the path is line or sight (LOS) or non-line of sight (NLOS), or LoS probability

• BLER

• Throughput

[0276] For example, if UE speed is < X and Pathloss > Y (e.g , indoor), then the UE may determine Model #1. If UE speed is < X and Pathloss < Y (e.g., walking outdoor), then the UE may determine Model #2. If UE speed is > X and Pathloss < Y (e.g., high speed outdoor), then the UE may determine Model #3. If SINR > X and/or LOS probability > Y (e.g. relatively good channel environment), then the UE may determine Model #4. Otherwise, Model #5 If coherence time > Model #6, otherwise Model #7.

[0277] In one embodiment, the UE may maintain the number/percentage of NACKs over a time period If number/percentage of NACKs > X, the UE may determine Model #4. The time period may be dynamically determined (e.g., sliding window).

[0278] In one embodiment, the UE may transmit to the gNB a preferred set of beam measurements (e.g., beam IDs and corresponding L1-RSRPs) based on which the gNB may determine that Model #5 is suitable for the UE.

[0279] In one embodiment, the UE may transmit to the gNB information on its location and/or mobility (e.g., UE speed, position, direction of motion, etc.) based on which the gNB may determine that Model #6 is suitable for the UE.

[0280] In one embodiment, the UE may measure a channel coherence time. A small value of the channel coherence time < X may be indicative of a fast-fading channel. The UE may determine Model #7 accordingly.

[0281] In one embodiment, the UE may determine a set of CSI parameters (e.g., for CSI reporting associated with AI/ML model and/or AI/ML model type). For example, if the UE determines a first AI/ML model, the UE may determine a first set of CSI parameters for UE reporting. If the UE determines a second AI/ML model, the UE may determine a second set of CSI parameters for UE reporting. For example, one or more of the following may be used.

• Model #1 : X number of beam IDs and corresponding L1-RSRPs for the current measurement (N1)

• Model #2: X1-1 number of beam IDs and corresponding L1-RSRPs for the current measurement and X1-2 number of beam IDs and corresponding L1-RSRPs for N1+N2

• Model #3: X2-1 , X2-2, X2-3 and X2-4 number of beam IDs and corresponding L1- RSRPs for N1/N2/N3/N4

[0282] Alternatively, the UE may indicate a preferred set of CSI parameters to a gNB for future reporting. E.g., 1st set: X number of beam IDs and corresponding L1-RSRPs for the current measurement (N1). 2nd set: X1-1 number of beam IDs and corresponding L1-RSRPs for the current measurement and X1-2 number of beam IDs and corresponding L1-RSRPs for N1+N2. 3rd set: X2-1 , X2-2, X2-3 and X2-4 number of beam IDs and corresponding L1-RSRPs for N1/N2/N3/N4.

[0283] In another embodiment, at step 416, the UE may execute a plurality of AI/ML models and evaluate their performance (e.g. accuracy, specificity, selectivity, etc.) In some embodiment, UE may identify an AI/ML model for (re)selection based on a function of an AI/ML model The gNB may have several AI/ML models for different functions. The UE may identify the AI/ML model for (re)selection based on the function of the AI/ML model, such as:

• CSI compression

• CSI prediction

• Beam prediction

• Obstacle prediction

• Position prediction

• UE speed prediction

• UE rotation prediction

[0284] For example, the UE may want to perform CSI compression and as such select an AI/ML model for CSI compression.

[0285] In some embodiments, the UE identification of an AI/ML model for (re)selection may be based on other parameters. A UE may identify an AI/ML model for (re)selection based one or more of the following parameters:

• UE Zone ID, UE Cell ID, TRP/gNB ID: for example, a UE may request for a model that has been trained (possibly by a neighboring UE) in the same zone/cell/TRP/gNB.

• UE capabilities: e.g., antenna configuration supported by UE, whether the UE supports TDD/FDD/full-duplex/half-duplex, etc. The UE may determine if an AI/ML model is suitable if it is applicable to at least one of its capabilities, for example, if the model was trained using the same antenna configuration, if the ML model was trained on a full-duplex or half-duplex configuration and so on.

• Model type: e.g., Recurrent Neural Network (RNN), Convolutional Neural Network (CNN), Deep Neural Network (DNN), Long Short Term Memory (LSTM), one-sided model v/s two-sided model, etc.) The UE may determine if an AI/ML model is suitable if it uses a model type that the UE supports. For example, a UE may only support one-sided models and may not be able to support a 2-sided model with an encoder at UE and decoder at gNB.

• UE traffic type: for example, a UE may determine if an AI/ML model is suitable if it is applicable to the UE’s traffic type, e.g., periodic/aperiodic, burst start/end/duration, throughput etc. The UE may select the model only if it has been trained on a similar type of traffic. • System parameters: e.g , subcarrier spacing, CP length, waveform type, carrier ID, bandwidth part ID, slot number, frame number, and a time index.

[0286] In some embodiments, a UE identification of AI/ML model for (re)selection is based on model performance monitoring. The UE may identify a model based on performance metrics, e.g., Normalized Mean Square Error (NMSE), Cosine Similarity, etc. The NMSE/Cosine Similarity may be measured, for example, by considering the difference between the ML model output and the traditional method of computation. The accuracy of the ML model may be inversely proportional to the NMSE value (computed at the UE or the gNB). The UE may request for the performance of the available models at the gNB. In one example, the UE may only select a model that has a low NMSE < X. In one example, the UE may only select a model that has a high Cosine similarity coefficient > X.

[0287] In some embodiments, a UE identification of AI/ML model for (re)selection is based on static information about the model. A UE may identify an AI/ML model for (re)selection based on static parameters about the model. The UE may send the identified/preferred model to the gNB and/or the raw parameters used for selection (e.g., the UE may send a request for a model < X where X may be reported in Mbytes/Mbits/Kbytes, etc.) Static model parameters used for identification/selection of model may include any one or more of:

• Model size

• Training metrics

• Overhead associated with model

• Latency for model to output results

[0288] In one example, the UE may identify a model of size < X (where X may be measured in Mbytes/Mbits /Kbytes, etc.). The UE may send the ID of the selected model based on model size, and/or the value of X. In one example, the UE may identify a model based on the bandwidth required to download the model. The UE may send the ID of the selected model based on available bandwidth or report the available bandwidth to the gNB). In one example, the UE may identify a model for (re)selection based on training metrics (e.g., time to train, number of iterations to achieve convergence, amount of data required for training, retraining frequency, etc.)

[0289] In one embodiment, the UE identification of a model for (re)selection may be a one-stage or multiplestage process. The (re)selection process to identify a model for download may be a one-stage or multiple stage process. In one example, upon sending the request for a model to the gNB (e.g., via model ID), the UE may receive an ACK for model download or it may receive the model directly in the DL. In one example, upon sending the parameters used to identity the model, the UE may receive some options of models that are suitable for the UE, based on the transmitted parameters. For example, if the UE reports SIN R to the gNB, it may receive from the gNB some model options (e.g., Models #3 and #4) that are suitable for the SINR range indicated by the UE. The UE may also receive from the gNB other parameters (e.g., antenna configuration that the models were trained on). The UE may then downselect based on these parameters and send a request to the gNB for the preferred AI/ML model.

[0290] In one example, the UE may require a model for CSI prediction. The UE may request for the CSI predictions models available at the gNB and may receive from the gNB one or more options of models for CSI prediction. The UE may also send to the gNB some additional parameters (e.g., UE antenna configuration, SI NR range, UE coherence bandwidth, etc.) that may allow the gNB to do a first round of selection amidst the available models and only send to the UE models that may be suitable to the UE (e.g., match the UE antenna configuration, trained for the SINR range measured at the UE). In a first step, the UE may only receive metadata about the AI/ML models available at the gNB (e.g., model function, configuration, model size, parameters/conditions under which the model was trained, etc.). Once all rounds of the selection process are complete (at the UE and/or gNB), the UE may receive the selected AI/ML model in the DL.

[0291] Herein accuracy, validity, predication accuracy, or prediction validity of an AI/ML model may be used interchangeably. Herein, a beam resource may consist of a TCI state, CSI-RS or a SSB for downlink, an SRS resource or TCI state for uplink.

[0292] In an embodiment, at step 416, a UE may select one or more AI/ML models and determine to activate the selected AI/ML models. The activated AI/ML models may be chosen, determined, and/or selected from a set of associated AI/ML models. The activated AI/ML models may be selected from a list of candidate AI/ML models and/or AI/ML model types that the UE has identified for AI/ML model reselection. In an example, the UE may determine to evaluate the prediction accuracy for the activated AI/ML models.

[0293] In some embodiments, methods to evaluate the accuracy of an AI/ML model are disclosed. A UE may determine the accuracy of the predictions made by an AI/ML model. The UE may determine or be configured with one or more resources to measure one or more parameters to determine the accuracy of an AI/ML model. The measurement resources may be one or more beam resources, reference signals, and time and frequency resources to perform the measurements.

[0294] In an example, an AI/ML model may use one or more beam resources from a first set of beam resources to perform beam predication and/or beam selection for one or more beam resources in a second set of beam resources. The UE may be configured with measurement resources in the first and second sets of beam resources to determine the accuracy of the AI/ML model. In an example, the UE may be configured with periodic, semi-persistent, or aperiodic measurement resources in the second set of beam resources to measure and compare with the predicted beam resources at the output of the AI/ML model (e.g , that may be based on the first set of beam resources).

[0295] The UE may determine the accuracy of an AI/ML model based on one or more of the following metrics: measurements, transmission performance, and/or requested reference signals, each discussed below. [0296] Referring first to measurements, in an embodiment, the UE may determine the accuracy based on one or more measured parameters based on one or more reference signals in addition to corresponding threshold values. The UE may measure at least one of the RSRP, RSSI, RSRQ, CQI, PMI, Rl, LI, probability of LOS, Doppler shift, Doppler spread, average delay, delay spread, and so forth. The UE may determine the accuracy of AI/ML model (e.g., beam prediction) by comparing the measured parameters with corresponding threshold.

[0297] In another example, the UE may compare the measurements based on AI/ML model (e g., beam predictions) with the measurements based on legacy methods (e.g., non-AI/ML-based) (e g., beam selections). One or more of the following may apply

[0298] The UE may perform a first measurement on one or more parameters (e.g., RSRP of the best N beams) for the beam resources (e.g., from the second set) that were predicted by an AI/ML model (e.g., based on the first set).

[0299] The UE may perform a second measurement on one or more parameters (e.g , RSRP of the best N beams) that were selected based on legacy methods (e.g., non-AI/ML-based, e g., beam selections) based on the received and/or configured reference signals (e g., based on the second set).

[0300] The UE may measure the difference between the first and second measurements (e g., comparing the RSRP of the best N beams).

[0301] The UE may compare the measured difference with a threshold

[0302] In case the measured difference is lower than the threshold, the UE may determine that the AI/ML model has an acceptable accuracy and/or performance.

[0303] Otherwise, if the measured difference is higher than the threshold, the UE may determine that the AI/ML model does not have an acceptable accuracy and/or performance.

[0304] Referring now to transmission performance, for example, the UE may determine the accuracy of an AIML model based on the performance of transmissions (e.g., for the predicted beam resources). The transmissions performance may be determined based on at least one of BLER, hypothetical PDCCH BLER, HARQ-ACK/NACK performance (e.g., rate of consecutive ACKs or consecutive NACKs), latency, data transmission throughput, spectral efficiency, and so forth.

[0305] In another example, the UE may determine the accuracy of an AIML model based on the performance of the transmission functions (e.g., for the predicted beam resources). The performance of transmission functions may be determined based on one or more of: the rate or the number of beam failure detections, the rate or the number of radio link failures, link adaptation performance and the rate or the number of ping-pong affect (e.g., going back and forth in the selected best beam resources, selected CQI, PMI, and so forth), persistent LBT failure, persistent existence of interference (e.g., cross-link interference (CLI)), rate or number of (consecutive) cell (re)selections, rate or number of failed random-access procedures, and so forth.

[0306] The UE may be configured with one or more parameters to measure and compare with corresponding thresholds to determine the accuracy of the AI/ML model (e.g., RSRP measurement of the best N beams and comparing with RSRP measurement of M actual best beams, and so forth). [0307] The UE may be configured with one or more counters and/or timers, and corresponding limits and/or thresholds to measure and compare the number of consecutive failures, the rate or the number of the occasions that an event is triggered, or the time duration of a procedure or an event. As such, the UE may determine the accuracy of the AI/ML model to be acceptable in case the counter or the timer is lower than the limit Otherwise, the UE may determine the accuracy of the AI/ML model to be not acceptable in case the counter or the timer is higher than the limit.

[0308] Referring now to requested reference signals, for example, the UE may request resources (e.g. , DL reference signals) to determine the validity of an AIML model The UE may indicate to the gNB the type of resource required, the AI L model (e.g., AIML model index), the associated function. As such, the UE may determine the accuracy of an AI/ML model based on the number of the requested RS transmissions for the evaluation and/or the number of different RS transmission for the activated models. The UE may determine the accuracy of the AI/ML model to be acceptable in case the number of requested reference signals for validation is lower than a corresponding limit Otherwise, the UE may determine the accuracy of the AI/ML model to be not acceptable if the number of requested reference signals for validation is higher than the corresponding limit. [0309] In other embodiments, a UE may compare the prediction accuracy of the activated AI/ML models. A UE may compare the accuracy of one or more activated AI/ML models. The UE may determine an accuracy metric for the activated AI/ML models The UE may order the activated AI/ML models based on the determined accuracy metric (e.g , in descending order) The accuracy metric may be based on one or more measured parameters, reference signals, counters, and/or timers that were determined or configured for evaluating the accuracy (e.g., separate accuracy metrics, joint optimization metrics, or combination of one or more measured parameters).

[0310] At step 410 following step 416, in such embodiments, the UE may report the accuracy or the order of accuracy for the activated AI/ML models. Alternatively, the UE may (implicitly) report the AI/ML model with highest accuracy level, when the UE selects and reports the model among the activated AI/ML models at steps 410 and 412.

[0311] In some embodiments, a gNB may confirm an AI/ML model type and/or AI/ML model for reselection at step 414. The UE may be configured to receive an indication and/or configuration from gNB associated with AI/ML model reselection at the UE. In one embodiment, the UE may receive the indication for AI/ML model reselection in a DCI. For example, the UE may receive an indication in DCI that indicates that the UE shall apply the model reselection as per the previous UE report/recommendation. In another example, the UE may receive an indication in DCI that configures the UE to apply a specific AI/ML model. In one embodiment, the UE may receive the indication for AI/ML model reselection in a MAC CE. For example, the indication may include the target model identity that the UE should apply. For example, the indication may include the function/use case for which the AI/ML model is applicable. For example, the function/use case may be associated with CSI feedback, Beam management, etc. For example, the indication may include the sub-use case associated with the target model identity and the function/use case. For example, the sub-use case may be associated with spatial prediction, temporal prediction, compression only, compression and prediction etc. In another embodiment, the UE may receive the configuration for AI/ML model reselection in an RRC message. For example, the UE may receive a plurality of configuration sets -wherein each configuration set may include parameters for an AI/ML model (e.g., model ID, model type, or the likes), applicable function/use case and optionally sub-use case. For example, the configuration set may include parameterization for the function/use case/sub-case. For example, the configuration set for the CSI feedback use case may include CSI-RS resource, measurement and/or reporting configuration etc. For example, the configuration set for the beam management use case may include measurement window, prediction window, set A and/or set B parameters etc. In a solution, the MAC CE may semi-statically activate/deactivate applicable configuration sets. In another solution, the DCI may dynamically indicate the configuration (possibly within the active configuration sets) that the UE should apply for AI/ML model reselection.

[0312] In some embodiments, UE behavior/actions may be defined upon gNB indication. The UE may be configured to perform one or more actions based on the AI/ML model reselection indication from gNB. The UE may apply the AI/ML model reselection indication from the gNB even if it does not correspond to the UE recommendation for AI/ML model reselection. The UE may continue using the current AI/ML model if it doesn’t receive any indication from gNB In an embodiment, the indication from a gNB may indicate an AI/ML model different than the currently used AI/ML model at the UE. Upon receiving such indication, the UE may deactivate the currently used AI/ML model and activate the indicated AI/ML model. Possibly the UE may discard any feedback/reports generated via the deactivated models that are yet to be transmitted. In an embodiment, the UE may receive additional configuration of use case specific parameters (e.g., CSI compression/prediction configuration, beam management/prediction configuration). For example, the UE may be configured to apply the use case specific parameters after reselecting to the new AI/ML model

[0313] In an embodiment, the AI/ML model reselection indication may carry implicit use case specific configuration. For example, a UE may be configured to infer the use case specific configuration from the AI/ML model ID (or configuration thereof). In another solution, the AI/ML model reselection indication may carry implicit AI/ML model ID (or configuration thereof). For example, the UE may be configured to infer the AI/ML model ID from the use case specific configuration.

[0314] In an embodiment, at step 422, the UE may be configured to activate the AI/ML model and/or apply the new CSI and/or Beam management configuration immediately upon receiving the gNB indication For example, upon reselection to the indicated AI/ML model for beam management, the UE may update the set A and/or set B and/or observation window and/or prediction window configuration as applicable to the reselected AI/ML model. In an embodiment, the UE may be configured to activate the AI/ML model and/or apply the new CSI/Beam management configuration within a time offset T relative to the timing of the reception of gNB indication. In another solution, the UE may be configured to activate the AI/ML model and/or apply the new CSI/Beam management configuration before the next CSI and/or beam reporting opportunity relative to the timing of reception of gNB indication For example, the UE may be configured to report the next CSI and/or beam report using the reselected AI/ML model and/or the associated configuration thereof. In yet another embodiment, the UE may be configured to activate the AI/ML model and/or apply the new CSI/Beam management configuration before the N CSI and/or beam reporting opportunity. For example, the value of N may be preconfigured. For example, the value of time offset T and/or the value of N may be preconfigured and/or signaled in the gNB indication for AI/ML model reselection. For example, the value of T and/or value of N may be a function of UE capability. For example, different values of T and/or N may be configured for different functions/use cases/sub-use cases/AI/ML models/parameter configurations.

[0315] In one embodiment, the UE may be preconfigured with rules to activate AI/ML model and/or apply the new CSI and/or Beam Management (BM) configuration as a function of type and/or number of AI/ML models associated with the gNB indication. For example, if the UE is configured to reselect AI/ML models for both CSI and BM, the UE may first reselect the AI/ML model for BM and upon successful reselection of BM AI/ML model, the UE may reselect the AI/ML model for CSI. Herein successful reselection may include activation of the new AI/ML model and/or applying the new configuration associated with the function (e.g., CSI/BM) and/or performing measurements using the new configuration and/or performing inference using the new AI/ML model and/or transmission of report associated with the function.

[0316] Although features and elements are described above in particular combinations, one of ordinary skill in the art will appreciate that each feature or element can be used alone or in any combination with the other features and elements. In addition, the methods described herein may 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 over wired or wireless connections) and computer-readable storage media. Examples of computer-readable storage media include, but are not limited to, a read only memory (ROM), a random access memory (RAM), a register, cache memory, semiconductor memory devices, magnetic media such as internal hard disks and removable disks, magnetooptical media, and optical media such as CD-ROM disks, and digital versatile disks (DVDs). A processor in association with software may be used to implement a radio frequency transceiver for use in a WTRU, UE, terminal, base station, RNC, or any host computer.

Claims

CLAIMS What is Claimed:

1. A user equipment (UE) comprising: one or more processors configured to: measure one or more communication channel conditions, trigger reselection of an artificial intelligence/machine learning (AI/ML) model based on at least a first criterion related to the measured one or more communication channel conditions, and determine, responsive to the triggered reselection, an AI/ML model type for AI/ML model reselection based on at least a second criterion related to the measured one or more communication channel conditions; and one or more transceivers configured to: transmit, responsive to the triggered reselection, a message to a gNB indicating the AI/ML model is to be reselected, and receive, from the gNB responsive to receipt of the message, one or more reference signals; wherein the one or more processors are further configured to: measure the one or more reference signals, and select an AI/ML model of the determined AI/ML model type based on the measured one or more reference signals; and wherein the one or more transceivers configured to transmit, to the gNB, an identifier of the selected AI/ML model.

2. The UE of any preceding claim, wherein the first criterion comprises a number of negative acknowledgements (NACKs) exceeding a first threshold or a received signal strength being below a second threshold.

3. The UE of any preceding claim, wherein the second criterion is different from the first criterion.

4. The UE of claim any preceding claim, wherein the selection of the AI/ML model is further based on a reference signal received power (RSRP), a speed of the UE, a pathloss, a zone identifier (ID), a cell ID, or a transmission/reception point (TRP) ID.

5. The UE of any preceding claim, wherein the one or more transceivers are further configured to receive, from the gNB, a confirmation message of the AI/ML model reconfiguration.

6. The UE of any preceding claim, wherein the one or more processors are further configured to: responsive to the triggered reselection, execute a plurality of AI/ML models of the determined AI/ML type; evaluate, for each of the plurality of AI/ML models, a prediction accuracy of subsequent measurements of the one or more communication channel conditions; and select the AI/ML model having a highest prediction accuracy.

7. The UE of claim 6, wherein the one or more processors are further configured to execute the plurality of

AI/ML models using as model inputs the measured one or more communication channel conditions or a subset of subsequent measurements of the one or more communication channel conditions.

8. A method, comprising: measuring, by one or more processors of a user equipment (UE), one or more communication channel conditions; triggering, by the one or more processors, reselection of an artificial intelligence/machine learning (AI/ML) model based on at least a first criterion related to the measured one or more communication channel conditions; determining, by the one or more processors responsive to the triggered reselection, an AI/ML model type for AI/ML model reselection based on at least a second criterion related to the measured one or more communication channel conditions; responsive to the triggered reselection, transmitting, by one or more transceivers of the UE, a message to the gNB indicating the AI/ML model is to be reselected; receiving, by the one or more transceivers from the gNB responsive to receipt of the message, one or more reference signals; measuring, by the one or more processors, the one or more reference signals; selecting, by the one or more processors, an AI/ML model of the determined AI/ML model type based on the measured one or more reference signals; and transmitting, by the one or more transceivers, an identifier to a gNB of the selected AI/ML model.

9. The method of claim 8, wherein the first criterion comprises a number of negative acknowledgements (NACKs) exceeding a first threshold or a received signal strength being below a second threshold.

10. The method of claim 8 or 9, wherein the second criterion is different from the first criterion.

11. The method of any of claims 8 through 10, wherein the selection of the AI/ML model is further based on a reference signal received power (RSRP), a speed of the U E, a pathloss, a zone identifier (ID), a cell ID, or a transmission/reception point (TRP) ID.

12. The method of any of claims 8 through 11 , further comprising receiving, by the one or more transceivers from the gNB, a confirmation message of the AI/ML model reconfiguration.

13. The method of any of claims 8 through 12, further comprising: responsive to the triggered reselection, executing, by the one or more processors, a plurality of AI/ML models of the determined AI/ML type; for each of the plurality of AI/ML models, evaluating, by the one or more processors, a prediction accuracy of subsequent measurements of the one or more communication channel conditions; and selecting, by the one or more processors, the AI/ML model having a highest prediction accuracy.

14. The method of claim 13, further comprising executing the plurality of AI/ML models using as model inputs the measured one or more communication channel conditions or a subset of subsequent measurements of the one or more communication channel conditions.

15. A wireless transmit/receive unit (WTRU) configured to perform any of the methods of claims 8 through 14.

16. One or more processors, configured to perform any of the methods of claims 8 through 14.

17. A wireless circuit, configured to perform any of the methods of claims 8 through 14.

18. A network device, configured to perform any of the methods of claims 8 through 14.

19. An electronic device, configured to perform any of the methods of claims 8 through 14.

20. A non-transitory computer readable medium comprising instructions that, when executed by one or more processors, cause a device to perform any of the methods of claims 8 through 14.