Method and system for adaptive CSI quantization

Abstract

A wireless transmit / receive unit (WTRU) is configured to receive configuration information from a network node for alignment of one or more compressed channel state information (CSI) quantizers at the WTRU and the network node. The configuration information may indicate multiple CSI quantization settings and / or criteria for selecting from the multiple CSI quantization settings. The WTRU may perform one or more CSI measurements on one or more reference signals. The WTRU may encode the one or more CSI measurements. The WTRU may select a CSI quantization setting from the multiple quantization settings based on the criteria indicated in the configuration information. The WTRU may quantize the encoded one or more CSI measurements using the selected CSI quantization setting. The WTRU may send a feedback report to the network node including the quantized one or more CSI measurements and / or an indication of the selected CSI quantization setting.

Description

[Background technology]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application No. 63 / 421,865, filed November 2, 2022, the contents of which are incorporated herein by reference in their entirety.

[0002] Information compression methods such as channel state information (CSI) compression may be studied for future wireless communication systems. CSI data for multiple-input multiple-output (MIMO) communication to the receiving end may cause large overhead unless efficient compression techniques are used. Summary of the Invention [Means for solving the problem]

[0003] In examples, methods and / or procedures may be utilized to enable an adaptive quantization process for the output of an encoder portion of an autoencoder trained for information compression (e.g., channel state information (CSI) compression). Examples may include configuring an adaptive (e.g., first-order) CSI quantizer, selecting and / or determining associated parameters, feeding back the parameters, and / or monitoring and / or reporting its performance to the receiving end.

[0004] A wireless transmit / receive unit (WTRU) may receive configuration information from a network node for alignment of one or more compressed channel state information (CSI) quantizers at the WTRU and the network node. The configuration information may indicate multiple CSI quantization settings and / or criteria for selecting from the multiple CSI quantization settings. The WTRU may perform one or more CSI measurements on one or more reference signals. The WTRU may encode the one or more CSI measurements. The WTRU may select a CSI quantization setting from the multiple quantization settings based on the criteria indicated in the configuration information. The WTRU may quantize the encoded one or more CSI measurements using the selected CSI quantization setting. The WTRU may send a feedback report to the network node including the quantized one or more CSI measurements and / or an indication of the selected CSI quantization setting for alignment of the one or more compressed CSI quantizers at the WTRU and the network node.

[0005] The criteria may be based on one or more of a reference signal received power (RSRP) measurement, a CSI measurement, a type of feedback report, a resource used to send the feedback report, and / or a metric associated with one or more encoded CSI measurements.

[0006] The indication of the selected CSI quantization setting may include an index associated with multiple CSI quantization settings. The indication of the selected CSI quantization setting may include an indication of a quantization type, a quantization parameter, a quantizer granularity, and / or a quantizer resolution. The quantization type may include a single quantizer method and / or a multiple quantizer method. The quantizer type may include a distribution-based method, a cluster-based method, or uniform, non-uniform, or vector-based quantization. The quantization parameter may include a centroid of a distribution or cluster of encoder outputs.

[0007] The WTRU may be configured with a fallback (e.g., secondary) quantizer. The WTRU may select the fallback (e.g., secondary) quantizer based on a second criterion. The second criterion may include a current state of the WTRU's current speed, a received signal strength indicator (RSSI), or performance of one or more compressed CSI quantizers and / or the fallback (e.g., secondary) quantizer measured via cosine similarity and / or quality noise.

[0008] The WTRU may determine the quantizer parameters based on an index associated with a predefined lookup table. The WTRU may determine one or more parameters of a codebook for the predefined CSI quantizer. [Brief explanation of the drawings]

[0009] [Figure 1A] FIG. 1 is a system diagram illustrating an example communication system in which one or more disclosed embodiments may be implemented. [Figure 1B] 1B is a system diagram illustrating an exemplary wireless transmit / receive unit (WTRU) that may be used within the communications system shown in FIG. 1A according to one embodiment. [Figure 1C] FIG. 1B is a system diagram illustrating an example radio access network (RAN) and core network (CN) that may be used within the communication system of FIG. 1A according to one embodiment. [Figure 1D] FIG. 1B is a system diagram illustrating a further example RAN and CN that may be used within the communication system shown in FIG. 1A, according to one embodiment. [Figure 2] FIG. 1 illustrates an example of a channel state information (CSI) measurement configuration. [Figure 3] FIG. 1 illustrates an example of codebook-based precoding using feedback information. [Figure 4]FIG. 1 illustrates an example block diagram of a CSI compression framework. [Figure 5] 1A and 1B are diagrams illustrating examples of a uniform quantizer and a non-uniform quantizer. [Figure 6] FIG. 1 is an exemplary diagram of vector quantization. [Figure 7] FIG. 1 illustrates an example of CSI compression using quantization. [Figure 8] FIG. 10 is a diagram illustrating an example of a distribution of encoder output values. [Figure 9] FIG. 2 illustrates an example of a non-uniform quantization process. [Figure 10] FIG. 10 is a diagram illustrating an example of clustering. [Figure 11] FIG. 1 illustrates an example of a quantization process in a wireless transmit / receive unit (WTRU). [Figure 12] FIG. 1 illustrates an exemplary multi-quantizer method. [Figure 13] FIG. 1 illustrates an exemplary inverse quantizer for a distribution-based method. DETAILED DESCRIPTION OF THE INVENTION

[0010] 1A is a system diagram illustrating an example communication system 100 in which one or more disclosed embodiments may be implemented. The communication system 100 may be a multiple-access system providing content, such as voice, data, video, messaging, broadcast, etc., to multiple wireless users. The communication system 100 may enable the multiple wireless users to access such content through the sharing of system resources, including wireless bandwidth. For example, the communication system 100 may employ one or more channel access methods, such as code division multiple access (CDMA), time division multiple access (TDMA), frequency division multiple access (FDMA), orthogonal FDMA (OFDMA), single-carrier FDMA (SC-FDMA), zero-tail (ZT) unique-word (UW) discrete Fourier transform (DFT) spread OFDM (ZT UW DTS-s OFDM), unique word OFDM (UW-OFDM), resource block filtered OFDM, filter bank multicarrier (FBMC), etc.

[0011] 1A, communications system 100 may include wireless transmit / receive units (WTRUs) 102a, 102b, 102c, 102d, radio access networks (RANs) 104 / 113, core networks (CNs) 106 / 115, public switched telephone networks (PSTNs) 108, the Internet 110, and other networks 112, although it will be understood that the disclosed embodiments contemplate any number of WTRUs, base stations, networks, and / or network elements. Each of 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” and / or “STA,” may be configured to transmit and / or receive wireless signals, and may include (or be) a user equipment (UE), a mobile station, a fixed or mobile subscriber unit, a subscription-based unit, a pager, a cellular phone, 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 (IoT) device, a watch or other wearable, a head-mounted display (HMD), a vehicle, a drone, a medical device and application (e.g., remote surgery), an industrial device and application (e.g., robots and / or other wireless devices operating in an industrial and / or automated processing chain context), a consumer electronics device, a device operating on a commercial and / or industrial wireless network, etc. Any of the WTRUs 102a, 102b, 102c, and 102d may be referred to interchangeably as a UE.

[0012] The communications system 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 communications networks, such as, for example, the CN 106 / 115, the Internet 110, and / or the network 112. By way of example, the base stations 114a, 114b may be any of a base transceiver station (BTS), a Node B (NB), an eNodeB (eNB), a Home Node B (HNB), a Home eNodeB (HeNB), a gNode B (gNB), a NR Node B (NR NB), a site controller, an access point (AP), a wireless router, etc. Although the base stations 114a, 114b are each shown as a single element, it will be understood that the base stations 114a, 114b may include any number of interconnected base stations and / or network elements.

[0013] The base station 114a may be part of the RAN 104 / 113, 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, etc. The base station 114a and / or base station 114b may be configured to transmit and / or receive wireless signals on one or more carrier frequencies, sometimes referred to as a cell (not shown). These frequencies may be licensed spectrum, unlicensed spectrum, or a combination of licensed and unlicensed spectrum. A cell may provide coverage for wireless services in a particular geographic area, which may be relatively fixed or may change over time. A cell may be further 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 one embodiment, the base station 114a may employ multiple-input multiple-output (MIMO) technology and utilize multiple transceivers for each or any sector of the cell. For example, beamforming may be used to transmit and / or receive signals in desired spatial directions.

[0014] 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).

[0015] More particularly, as mentioned above, the communication 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, etc. For example, the base station 114a and the WTRUs 102a, 102b, 102c in the RAN 104 / 113 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 Packet Access (HSDPA) and / or High Speed ​​Uplink Packet Access (HSUPA).

[0016] In one 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).

[0017] In one embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement a radio technology, such as New Radio (NR) radio access, which may establish the air interface 116 using NR.

[0018] In one 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 jointly implement LTE radio access and NR radio access, e.g., using dual connectivity (DC) principles. Thus, the air interface utilized by the WTRUs 102a, 102b, 102c may be characterized by multiple types of radio access technologies and / or transmissions from / to multiple types of base stations (e.g., eNBs and gNBs).

[0019] In one embodiment, the base station 114a and the WTRUs 102a, 102b, 102c may implement a wireless technology such as IEEE 802.11 (i.e., Wireless Fidelity (Wi-Fi)), 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), or the like.

[0020] 1A may be, for example, a wireless router, a Home NodeB, a Home eNodeB, or an access point and may utilize any suitable RAT for facilitating wireless connectivity in a localized area, such as a business, a home, a vehicle, a campus, an industrial facility, an air corridor (e.g., for use by drones), a road, etc. 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 one 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 one 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 either a small cell, a picocell, or a femtocell. 1A, the base station 114b may have a direct connection to the Internet 110. Therefore, the base station 114b may not be required to access the Internet 110 via the CN 106 / 115.

[0021] The RAN 104 / 113 may be in communication with the CN 106 / 115, which may be any type of network configured to provide voice, data, application, and / or voice over internet protocol (VoIP) services to one or more of the WTRUs 102a, 102b, 102c, 102d. The data may have various quality of service (QoS) requirements, such as different throughput, latency, error resilience, reliability, data throughput, mobility, etc. The CN 106 / 115 may provide call control, billing services, mobile location-based services, prepaid 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 understood that the RAN 104 / 113 and / or the CN 106 / 115 may be in direct or indirect communication with other RANs employing the same RAT as the RAN 104 / 113 or a different RAT. For example, in addition to being connected to the RAN 104 / 113, which may utilize NR radio technology, the CN 106 / 115 may also be in communication with another RAN (not shown) that employs any of GSM, UMTS, CDMA2000, WiMAX, E-UTRA, or Wi-Fi radio technologies.

[0022] The CN 106 / 115 may also serve as a gateway for the WTRUs 102a, 102b, 102c, 102d to access the PSTN 108, the Internet 110, and / or other networks 112. The PSTN 108 may include a circuit-switched telephone network providing 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 TCP, User Datagram Protocol (UDP), and / or IP in the Transmission Control Protocol / Internet Protocol (TCP / IP) Internet protocol suite. The network 112 may include wired and / or wireless communication networks owned and / or operated by other service providers. For example, the network 112 may include another CN connected to one or more RANs, which may employ the same RAT as the RAN 104 / 114 or a different RAT.

[0023] 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. 1A may be configured to communicate with a base station 114a that may employ cellular-based wireless technology and may be configured to communicate with a base station 114b that may employ IEEE 802.11 wireless technology.

[0024] 1B is a system diagram illustrating an example WTRU 102. As shown in FIG. 1B, the WTRU 102 may include, among other things, 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 elements / peripherals 138. It will be understood that the WTRU 102 may include any sub-combination of the above elements while remaining consistent with an embodiment.

[0025] The processor 118 may be a general-purpose processor, a special-purpose processor, a conventional processor, a digital signal processor (DSP), multiple microprocessors, one or more microprocessors associated with a DSP core, a controller, a microcontroller, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) circuit, any other type of integrated circuit (IC), a state machine, etc. 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. 1B depicts the processor 118 and the transceiver 120 as separate components, it will be understood that the processor 118 and the transceiver 120 may be incorporated together, for example, in an electronic package or chip.

[0026] The transmit / receive element 122 may be configured to transmit signals to or receive signals from a base station (e.g., 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 one 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 one embodiment, the transmit / receive element 122 may be configured to transmit and / or receive both RF and light signals. It will be understood that the transmit / receive element 122 may be configured to transmit and / or receive any combination of wireless signals.

[0027] 1B as a single element, the WTRU 102 may include any number of transmit / receive elements 122. For example, 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.

[0028] The transceiver 120 may be configured to modulate signals to be transmitted by the transmit / receive element 122 and demodulate signals received by the transmit / receive element 122. As mentioned above, the WTRU 102 may have multi-mode capabilities. Thus, the transceiver 120 may include multiple transceivers to enable the WTRU 102 to communicate via multiple RATs, such as, for example, NR and IEEE 802.11.

[0029] The processor 118 of the WTRU 102 may be coupled to and may receive user input data from a speaker / microphone 124, a keypad 126, and / or a display / touchpad 128 (e.g., a liquid crystal display (LCD) display unit or an 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. Furthermore, the processor 118 may access information from and store data in any type of suitable memory, such as non-removable memory 130 and / or 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, etc. 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 home computer (not shown).

[0030] The processor 118 may receive power from the power source 134 and may be configured to distribute and / or control the power to 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 batteries (e.g., nickel-cadmium (NiCd), nickel-zinc (NiZn), nickel-metal hydride (NiMH), lithium-ion (Li-ion), etc.), solar cells, fuel cells, etc.

[0031] The processor 118 may also be coupled to a GPS chipset 136, which may be configured to provide location information (e.g., longitude and latitude) regarding the current location of the WTRU 102. In addition to, or instead of, information from the GPS chipset 136, the WTRU 102 may receive location information from base stations (e.g., base stations 114a, 114b) over the air interface 116 and / or determine its location based on the timing of when signals are received from two or more nearby base stations. It will be understood that the WTRU 102 may acquire location information via any suitable location determination method while remaining consistent with an embodiment.

[0032] The processor 118 may further be coupled to other elements / peripherals 138, which may include one or more software and / or hardware modules / units that provide additional features, functionality, and / or wired or wireless connectivity. For example, the elements / peripherals 138 may include an accelerometer, an electronic compass, a satellite transceiver, a digital camera (for photos and / or videos), 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, etc. The element / peripheral 138 may include one or more sensors, which 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, and / or a humidity sensor.

[0033] The WTRU 102 may include a full-duplex radio where transmission and reception of some or all of the signals (e.g., associated with a particular subframe for both the uplink (e.g., for transmission) and the downlink (e.g., for reception) may be parallel and / or simultaneous. The full-duplex radio may include an interference management unit for reducing and or substantially eliminating self-interference via either hardware (e.g., a choke) or signal processing via a processor (e.g., a separate processor (not shown) or via the processor 118). In one embodiment, the WTRU 102 may include a half-duplex radio that is for transmission and reception of some or all of the signals (e.g., associated with a particular subframe for either the uplink (e.g., for transmission) or the downlink (e.g., for reception)).

[0034] 1C is a system diagram illustrating the RAN 104 and the CN 106, according to one embodiment. As mentioned above, the RAN 104 may employ E-UTRA radio technology to communicate with the WTRUs 102a, 102b, and 102c over the air interface 116. The RAN 104 may also be in communication with the CN 106.

[0035] The RAN 104 may include eNodeBs 160a, 160b, and 160c, although it will be understood that the RAN 104 may include any number of eNodeBs while remaining consistent with an embodiment. The eNodeBs 160a, 160b, and 160c may each include one or more transceivers for communicating with the WTRUs 102a, 102b, and 102c over the air interface 116. In an embodiment, the eNodeBs 160a, 160b, and 160c may implement MIMO technology. Thus, the eNodeB 160a, for example, may use multiple antennas to transmit wireless signals to, and receive wireless signals from, the WTRU 102a.

[0036] Each of the eNodeBs 160a, 160b, and 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 uplink (UL) and / or downlink (DL), etc. As shown in FIG. 1C, the eNodeBs 160a, 160b, 160c may communicate with one another over an X2 interface.

[0037] 1C may include a mobility management entity (MME) 162, a serving gateway (SGW) 164, and a packet data network (PDN) gateway (PGW) 166. Although each of the above elements is shown as part of the CN 106, it will be understood that any one of these elements may be owned and / or operated by an entity other than the CN operator.

[0038] The MME 162 may be connected to each of the eNodeBs 160a, 160b, and 160c in the RAN 104 via an S1 interface and may act 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 initial attach of the WTRUs 102a, 102b, 102c, etc. 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.

[0039] The SGW 164 may be connected to each of the eNodeBs 160a, 160b, 160c in the RAN 104 via an 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 the user plane during inter-eNodeB handover, triggering paging when DL data is available for the WTRUs 102a, 102b, 102c, managing and storing the context of the WTRUs 102a, 102b, 102c, etc.

[0040] The SGW 164 may be connected to a PGW 166, which may provide the WTRUs 102a, 102b, 102c with access to packet-switched networks, such as the Internet 110, to facilitate communication between the WTRUs 102a, 102b, 102c and IP-enabled devices.

[0041] The CN 106 may facilitate communication 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 communication between the WTRUs 102a, 102b, 102c and legacy landline communication devices. For example, the CN 106 may include or communicate with an IP gateway (e.g., an IP Multimedia Subsystem (IMS) server) that acts as an interface between the CN 106 and the PSTN 108. Additionally, the CN 106 may provide the WTRUs 102a, 102b, 102c with access to other networks 112, which may include other wired and / or wireless networks owned and / or operated by other service providers.

[0042] Although the WTRU is described in Figures 1A-1D as a wireless terminal, it is contemplated that in certain representative embodiments, such a terminal may use a wired communication interface with the communication network (e.g., temporarily or permanently).

[0043] In a representative embodiment, the other network 112 may be a WLAN.

[0044] 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 to or interface with a distribution system (DS) or another type of wired / wireless network that carries traffic during and / or from the BSS. Traffic to the STA originating from outside the BSS may arrive through the AP and be delivered to the STA. Traffic originating from the STA to a destination outside the BSS may be sent to the AP for delivery to the respective destination. Traffic between STAs within the BSS may be sent through the AP, e.g., where a source STA may send traffic to the AP, and the AP may deliver the traffic to the destination STA. Traffic between STAs within the BSS may be considered and / or referred to as peer-to-peer traffic. Peer-to-peer traffic may be sent between (e.g., directly between) a source STA and a destination STA via a direct link setup (DLS). In some representative embodiments, the DLS may use 802.11e DLS or 802.11z tunneled DLS (TDLS). A WLAN using an Independent BSS (IBSS) mode may not have an AP, and the STAs within or using the IBSS (e.g., all of the STAs) can communicate directly with each other. The IBSS communication mode is sometimes referred to herein as an "ad hoc" communication mode.

[0045] When using the 802.11ac infrastructure mode of operation or a similar mode of operation, an AP can transmit beacons on a fixed channel, such as a primary channel. The primary channel can be a fixed width (e.g., a 20 MHz wide bandwidth) or a width dynamically set via signaling. The primary channel can be the operating channel of the BSS and can be used by STAs to establish a connection with the AP. In some representative embodiments, carrier sense multiple access with collision avoidance (CSMA / CA) can be implemented, for example, in an 802.11 system. In CSMA / CA, STAs (e.g., every STA), including the AP, can sense the primary channel. If the primary channel is sensed / detected and / or determined to be busy by a particular STA, the particular STA can back off. One STA (e.g., only one station) can transmit at any given time in a given BSS.

[0046] High-throughput (HT) STAs may use, for example, a 40 MHz wide channel for communication via a combination of a primary 20 MHz channel with adjacent or non-adjacent 20 MHz channels to form a 40 MHz wide channel.

[0047] A very high throughput (VHT) STA can support 20 MHz, 40 MHz, 80 MHz, and / or 160 MHz wide channels. A 40 MHz channel and / or an 80 MHz channel can be formed by combining contiguous 20 MHz channels. A 160 MHz channel can be formed by combining eight contiguous 20 MHz channels or by combining two non-contiguous 80 MHz channels, sometimes referred to as an 80+80 configuration. For the 80+80 configuration, after channel encoding, the data can be passed through a segment parser that can split the data into two streams. Inverse fast Fourier transform (IFFT) processing and time-domain processing can be performed separately for each stream. The streams can be mapped onto two 80 MHz channels, and the data can be transmitted by the transmitting STA. At the receiver of the receiving STA, the operations described above for the 80+80 configuration can be reversed, and the combined data can be sent to a medium access control (MAC) layer, entity, etc.

[0048] Sub-1 GHz operating modes are supported by 802.11af and 802.11ah. Channel operating bandwidths and carriers are reduced in 802.11af and 802.11ah relative to those used in 802.11n and 802.11ac. 802.11af supports 5 MHz, 10 MHz, and 20 MHz bandwidths in the TV white space (TVWS) spectrum, while 802.11ah supports 1 MHz, 2 MHz, 4 MHz, 8 MHz, and 16 MHz bandwidths using non-TVWS spectrum. According to a representative embodiment, 802.11ah can support meter-type control / machine-type communication (MTC), such as MTC devices in macro coverage areas. MTC devices can have limited capabilities, including, for example, support for some and / or limited bandwidths (e.g., only support for that). MTC devices can include batteries with above-threshold battery life (e.g., to maintain very long battery life).

[0049] WLAN systems that can support multiple channels and channel bandwidths, such as 802.11n, 802.11ac, 802.11af, and 802.11ah, include a channel that can be designated as a primary channel. The primary channel can have a bandwidth equal to the largest common operating bandwidth supported by all STAs in the BSS. The bandwidth of the primary channel can be set and / or limited by a STA, from among all STAs operating in the BSS, that supports the smallest bandwidth operating mode. In an 802.11ah example, the primary channel may be 1 MHz wide for a STA (e.g., an MTC-type device) that supports (e.g., only supports) the 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. For example, if a STA (that only supports 1 MHz operating mode) transmits to an AP such that the primary channel is busy, the entire available frequency band may be considered busy, even though most of the frequency band may remain idle and available for use.

[0050] In the United States, the available frequency bands that can be used by 802.11ah are 902 MHz to 928 MHz. In South Korea, the available frequency bands are 917.5 MHz to 923.5 MHz. In Japan, the available frequency bands are 916.5 MHz to 927.5 MHz. The total available bandwidth for 802.11ah is 6 MHz to 26 MHz, depending on the country code.

[0051] 1D is a system diagram illustrating the RAN 113 and the CN 115, according to one embodiment. As mentioned above, the RAN 113 may employ NR radio technology to communicate with the WTRUs 102a, 102b, 102c over the air interface 116. The RAN 113 may also be in communication with the CN 115.

[0052] The RAN 113 may include gNBs 180a, 180b, and 180c, although it will be understood that the RAN 113 may include any number of gNBs while remaining consistent with an embodiment. The gNBs 180a, 180b, and 180c may each include one or more transceivers for communicating with the WTRUs 102a, 102b, and 102c over the air interface 116. In an embodiment, the gNBs 180a, 180b, and 180c may implement MIMO technology. For example, the gNBs 180a and 180b may utilize beamforming to transmit signals to and / or receive signals from the WTRUs 102a, 102b, and 102c. Thus, the gNB 180a may, for example, use multiple antennas to transmit wireless signals to and / or receive wireless signals from the WTRU 102a. In one 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 an unlicensed spectrum, while the remaining component carriers may be on a licensed spectrum. In one embodiment, the gNBs 180a, 180b, 180c may implement coordinated multi-point (CoMP) technology. For example, the WTRU 102a may receive coordinated transmissions from the gNBs 180a and 180b (and / or 180c).

[0053] The WTRUs 102a, 102b, 102c may communicate with the gNBs 180a, 180b, 180c using transmissions associated with scalable numerologies. For example, the OFDM symbol spacing and / or OFDM subcarrier spacing may be different for different transmissions, different cells, and / or different portions of the wireless transmission spectrum. The WTRUs 102a, 102b, 102c may communicate with the gNBs 180a, 180b, 180c using subframes or transmission time intervals (TTIs) of varying or scalable lengths (e.g., including varying numbers of OFDM symbols and / or varying lengths of absolute time duration).

[0054] 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 a standalone configuration, the WTRUs 102a, 102b, 102c can communicate with the gNBs 180a, 180b, 180c without accessing any other RAN (e.g., eNodeBs 160a, 160b, 160c). In a standalone configuration, the WTRUs 102a, 102b, 102c can utilize one or more of the gNBs 180a, 180b, 180c as mobility anchor points. In a standalone configuration, the WTRUs 102a, 102b, 102c can communicate with the gNBs 180a, 180b, 180c using signals in unlicensed bands. In a non-standalone configuration, the WTRUs 102a, 102b, 102c may communicate / connect with a gNB 180a, 180b, 180c while also communicating / connecting with another RAN, such as an eNodeB 160a, 160b, 160c. For example, the WTRUs 102a, 102b, 102c may implement the DC principle to communicate with one or more gNBs 180a, 180b, 180c and one or more eNodeBs 160a, 160b, 160c substantially simultaneously. In a non-standalone configuration, the eNodeBs 160a, 160b, 160c may act as mobility anchors for the WTRUs 102a, 102b, 102c, and the gNBs 180a, 180b, 180c may provide additional coverage and / or throughput for serving the WTRUs 102a, 102b, 102c.

[0055] 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 for network slicing, dual connectivity, interworking between NR and E-UTRA, routing of user plane data to user plane functions (UPFs) 184a, 184b, routing of control plane information to access and mobility management functions (AMFs) 182a, 182b, etc. As shown in FIG. 1D, the gNBs 180a, 180b, 180c may communicate with one another via an Xn interface.

[0056] 1D may include at least one AMF 182a, 182b, at least one UPF 184a, 184b, at least one Session Management Function (SMF) 183a, 183b, and at least one Data Network (DN) 185a, 185b. While each of the above elements is shown as part of the CN 115, it will be understood that any of these elements may be owned and / or operated by an entity other than the CN operator.

[0057] The AMF 182a, 182b may be connected to one or more of the gNBs 180a, 180b, 180c in the RAN 113 via an N2 interface and may act as a control node. For example, the AMF 182a, 182b may be responsible for authenticating users of the WTRUs 102a, 102b, 102c, supporting network slicing (e.g., handling different protocol data unit (PDU) sessions with different requirements), selecting a particular SMF 183a, 183b, managing registration areas, terminating NAS signaling, mobility management, etc. Network slicing may be used by the AMF 182a, 182b to customize CN support for the WTRUs 102a, 102b, 102c, for example, based on the type of service being utilized by the 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, etc. The AMFs 182a, 182b may provide a control plane function for switching between the RAN 113 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 Wi-Fi.

[0058] The SMFs 183a and 183b may be connected to the AMFs 182a and 182b in the CN 115 via an N11 interface. The SMFs 183a and 183b may also be connected to the UPFs 184a and 184b in the CN 115 via an N4 interface. The SMFs 183a and 183b may select and control the UPFs 184a and 184b and configure the routing of traffic through the UPFs 184a and 184b. The SMFs 183a and 183b may perform other functions, such as managing and allocating UE IP addresses, managing PDU sessions, controlling policy enforcement and QoS, providing downlink data notification, etc. The PDU session type may be IP-based, non-IP-based, Ethernet-based, etc.

[0059] The UPFs 184a, 184b may be connected to one or more of the gNBs 180a, 180b, 180c in the RAN 113 via an N3 interface, which may provide the WTRUs 102a, 102b, 102c with access to packet-switched networks such as the Internet 110, for example, to facilitate communications between the WTRUs 102a, 102b, 102c and IP-enabled devices. The UPFs 184a, 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 downlink packets, providing mobility anchoring, etc.

[0060] The CN 115 may facilitate communication with other networks. For example, the CN 115 may include or communicate with an IP gateway (e.g., an IP Multimedia Subsystem (IMS) server) that acts as an interface between the CN 115 and the PSTN 108. Additionally, the CN 115 may provide the WTRUs 102a, 102b, 102c with access to other networks 112, which may include other wired and / or wireless networks owned and / or operated by other service providers. In one embodiment, the WTRUs 102a, 102b, 102c may be connected to the local data networks (DNs) 185a, 185b through the UPFs 184a, 184b via an N3 interface to the UPFs 184a, 184b and an N6 interface between the UPFs 184a, 184b and the DNs 185a, 185b.

[0061] 1A-1D and the corresponding description thereof, one or more, or all, of the functions described herein with respect to any of the WTRUs 102a-d, base stations 114a-b, eNodeBs 160a-c, MME 162, SGW 164, PGW 166, gNBs 180a-c, AMFs 182a-b, UPFs 184a-b, SMFs 183a-b, DNs 185a-b, and / or any other element / device(s) described herein may be performed by one or more emulation elements / 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 functionality.

[0062] The emulation device may be designed to implement one or more tests of other devices in a laboratory environment and / or in a carrier network environment. For example, one or more emulation devices may perform one or more, or all, functions while fully or partially implemented and / or deployed as part of a wired and / or wireless communication network to test other devices in the communication network. One or more emulation devices may perform one or more, or all, functions while temporarily implemented / deployed as part of a wired and / or wireless communication network. The emulation device may be directly coupled to another device for testing purposes and / or may perform testing using over-the-air wireless communication.

[0063] The one or more emulation devices may perform one or more functions, inclusive, while not being implemented / deployed as part of a wired and / or wireless communication network. For example, the emulation devices may be utilized in test labs and / or test scenarios in non-deployed (e.g., test) wired and / or wireless communication networks to implement testing of one or more components. The one or more emulation devices may be test equipment. Direct RF coupling and / or wireless communication via RF circuitry (which may, for example, include one or more antennas) may be used by the emulation devices to transmit and / or receive data.

[0064] Information compression methods such as channel state information (CSI) compression may be researched for future wireless communication systems. CSI data for multiple-input multiple-output (MIMO) communications used in transmission to the receiving end may cause a significant amount of overhead unless efficient compression techniques are used. CSI compression techniques using the Rel-18 SI scope for artificial intelligence / machine learning (AI / ML) for the air interface may focus on autoencoder-based compression, where the WTRU and / or gNB implement the encoder and / or decoder section(s) of an autoencoder.

[0065] An autoencoder (AE) is a type of neural network (e.g., a deep neural network (DNN)) used for data compression applications. The AE may be a two-sided model including an encoder model that can transform input data into a smaller-dimensional latent space (thus performing a compression operation) and / or a decoder model that can perform the inverse operation and / or reconstruct data based on the smaller-dimensional latent space. For one or more CSI compression operations, the encoder portion of the AE may reside in the WTRU, and the decoder portion of the AE may reside on the network side. The decoder portion may be responsible for reconstructing (e.g., recovering) the CSI from the received compressed CSI. The output of the WTRU-side encoder may be real-valued and / or may include further quantization, for example, to fit limited feedback overhead. In an example, quantizing the output values ​​may be uniform quantization. Other methods for quantization may include non-uniform quantization and / or vector quantization. Training the AE with a quantizer can achieve quantization of the compressed CSI. The quantizer may be inserted independently after training the AE. Efficient methods for quantization of compressed CSI may be described herein.

[0066] Training the AE with a learned quantizer may limit the flexibility of the compressed output and / or may require a separately trained neural network (NN) for a given overhead. For example, switching NN structures based on available feedback overhead (e.g., indicated by a base station (BS) to a WTRU) may not be feasible and / or impractical.

[0067] The values ​​of the encoder output of a CSI compression AE may be more frequent at low amplitudes (e.g., may be unevenly distributed). A uniform quantizer may not be optimal for AE-based CSI compression output. The distribution of the output may differ between outputs and / or may change over time. The distribution of the output may depend on the AE.

[0068] Vector quantization can be computationally intensive. For example, in vector quantization, a codeword may be compared to one or more (e.g., all) vectors in a codeblock. When implementing vector quantization, updating the codeblock online may incur overhead.

[0069] Methods for configuring an adaptive (e.g., primary) quantizer are provided herein and may include procedures for initial configuration of the adaptive (e.g., primary) quantizer for indication to the WTRU. These procedures may include, but are not limited to, the type of method for building the quantizer, the number of quantizers per output, quantization performance monitoring metrics and / or the type of secondary quantization, etc. The procedures herein may further include procedures for indication of self-configuration of the adaptive (e.g., primary) quantizer.

[0070] A method for determining quantizer parameters is provided herein and may include a procedure for determining quantizer parameters. Based on an initial configuration, the procedure may include, but is not limited to, determining characteristics of the distribution of encoder output values ​​for a distribution-based quantizer, cluster centroids for a cluster-based quantizer, quantizer parameters for all encoder outputs at once and / or separately, and / or adapting the quantizer according to the determined parameters.

[0071] Quantizer parameter feedback is provided herein and may include procedures for WTRU feedback (e.g., explicitly and / or implicitly, dynamically and / or semi-statically and / or statically), which may, for example, send quantizer parameters and / or configurations to the gNB for approval and / or implementation of an inverse quantizer.

[0072] Methods for monitoring and / or reporting quantizer performance are provided herein and may include procedures for dynamically monitoring quantization performance and / or periodically and / or dynamically monitoring quantizer performance and / or reporting quantizer performance and / or redetermining parameters based on an initial configuration and / or an indication from the gNB.

[0073] Methods for quantizer fallback are provided herein and may include procedures for determining quantizer fallback, including defining error threshold events, events that trigger the fallback, and / or mechanisms for fallback and reporting of the fallback.

[0074] The CSI may include one or more of a channel quality index (CQI), a rank indicator (RI), a precoding matrix index (PMI), an L1 channel measurement (e.g., a received signal received power (RSRP), e.g., L1-RSRP, or a signal-to-interference-and-noise ratio (SINR)), a CSI reference signal (CSI-RS) resource indicator (CRI), a synchronization signal / physical broadcast channel (SS / PBCH) block resource indicator (SSBRI), a layer indicator (LI), and / or any other measurement quantity measured by the WTRU from configured reference signals (e.g., CSI-RS and / or SS / PBCH blocks and / or any other reference signals).

[0075] The criteria for selecting from multiple CSI quantization configurations may be based on one or more of the RSRP measurement, the CSI measurement, the type of feedback report, the resource used to send the feedback report, and / or a metric associated with the encoded one or more CSI measurements.

[0076] CSI reporting may be performed using one or more frameworks. For example, the WTRU may report CSI through an uplink control channel on a physical uplink control channel (PUCCH) and / or per request of the gNB on an uplink (UL) physical uplink shared channel (PUSCH) grant. The CSI-RS may cover the full bandwidth or a portion of the bandwidth of the bandwidth portion (BWP), depending on the configuration.

[0077] The CSI-RS may be configured, for example, in each physical resource block (PRB) and / or every other PRB within the CSI-RS bandwidth. The CSI-RS resources may be configured periodic, semi-persistent, and / or aperiodic in the time domain. Semi-persistent CSI-RS may be similar to periodic CSI-RS. However, the resources may be deactivated by a medium access control element (MAC CE), and / or the WTRU may report associated measurements when (e.g., only when) the resources are activated. With aperiodic CSI-RS, the WTRU may trigger and / or report the measured CSI-RS on the PUSCH (e.g., by a request in the downlink control information (DCI)). Periodic reporting may be carried via the PUCCH. Either the PUCCH and / or the PUSCH can carry semi-persistent reporting. The scheduler may use the reported CSI when allocating optimal resource blocks (e.g., based on determining the time-frequency selectivity of the channel, a precoding matrix, a beam, a transmission mode, and / or selecting a preferred modulation and coding scheme (MCS). The reliability, accuracy, and / or timeliness of the WTRU CSI reporting may be included to meet ultra reliable low liability communications (URLLC) service requirements.

[0078] The WTRU may be configured with a CSI measurement configuration, which may include one or more CSI reporting configurations, resource configurations, and / or links between the one or more CSI reporting configurations and / or one or more resource configurations.

[0079] 2 shows an example of a configuration 200 for CSI reporting configurations, resource configurations, and / or links. In a CSI measurement configuration, one or more of the following configuration parameters may be provided: A CSI measurement configuration may include N≧1 CSI reporting configurations, M≧1 resource configurations, and / or a CSI measurement configuration linking N CSI reporting configurations with M resource configurations.

[0080] The CSI measurement configuration may include a CSI reporting configuration including one or more of: time domain behavior (e.g., aperiodic, periodic, and / or semi / persistent), frequency granularity (e.g., for PMI and / or CQI, etc.), CSI reporting type (e.g., PMI, CQI, RI, and / or CRI, etc.), and / or PMI type (e.g., Type I and / or Type II) and / or codebook configuration (e.g., if PMI is reported).

[0081] The CSI measurement configuration may include a resource configuration including one or more of a time-domain behavior (e.g., aperiodic, periodic, and / or semi-persistent), an RS type (e.g., for channel measurement and / or interference measurement), and / or S≧1 resource sets, and / or each resource set may include K resources.

[0082] The CSI measurement configuration may include one or more of: one or more CSI reporting configurations, one or more resource configurations, and / or a reference transmission scheme configuration (eg, for CQI). CSI reporting for a component carrier may support one or more of the following frequency granularities: wideband CSI, partial-band CSI, and / or sub-band CSI.

[0083] The CSI reporting framework can include codebook-based precoding. Figure 3 illustrates the concept of codebook-based precoding 300 with feedback information 302. The feedback information 302 can include a precoding matrix index (PMI), which may be referred to as a codeword index in a codebook, as shown in Figure 3.

[0084] As shown in FIG. 3, the codebook 304 may include a set of precoding vectors / matrices 306 for each rank and / or number of antenna ports. In an example, each precoding vector / matrix may include an index so that the receiver can notify the transmitter of a preferred precoding vector / matrix index. Codebook-based precoding may have performance degradation due to including a finite number of precoding vectors / matrices (e.g., compared to non-codebook-based precoding). However, an advantage of codebook-based precoding may be lower control signaling / feedback overhead. Table 1 shows an example of a codebook for 2Tx.

[0085] [Table 1]

[0086] The indication of the selected CSI quantization setting may include an index of a codebook associated with the multiple CSI quantization settings.

[0087] The CSI reporting framework may include CSI processing criteria. A CSI processing unit (CPU) may be referred to as a minimum CSI processing unit, and / or a WTRU may support one or more CPUs (e.g., N CPUs). A WTRU with N CPUs may estimate N CSI feedback calculations in parallel, where N may be a WTRU capability. The WTRU may implement (e.g., implement only) N high-priority CSI feedbacks. If the WTRU is required to estimate more than N CSI feedbacks simultaneously, the rest may not be estimated.

[0088] The start and / or end of the CPU may be determined based on the CSI reporting type (e.g., aperiodic, periodic, and / or semi-persistent) as described herein. For aperiodic CSI reporting, the CPU may be occupied starting from the first OFDM symbol after the PDCCH trigger and until the last OFDM symbol of the PUSCH carrying the CSI report. For periodic and / or semi-persistent CSI reporting, the CPU may be occupied starting from the first OFDM symbol of one or more associated measurement resources (e.g., not earlier than the CSI reference resource) and until the last OFDM symbol of the CSI report. The number of occupied CPUs may differ based on the CSI measurement type (e.g., beam-based and / or non-beam-based) as described herein.

[0089] A non-beam-associated report may include Ks CPUs when there are Ks CSI-RS resources in the CSI-RS resource set for channel measurement. A beam-associated report (e.g., “cri-RSRP,” “ssb-Index-RSRP,” and / or “none”) may include one CPU regardless of the number of CSI-RS resources in the CSI-RS resource set for channel measurement (e.g., due to low CSI calculation complexity). A beam-associated report including “none” may be used for P3 operation and / or aperiodic TRS transmission. Aperiodic CSI reporting using a single CSI-RS resource may occupy one CPU. A CSI reporting of Ks CSI-RS resources may occupy Ks CPUs because the WTRU can perform CSI measurements for each CSI-RS resource.

[0090] Number of unoccupied CPUs (N u ) is the CPU required for CSI reporting (N r ), the WTRU may drop the N_r-N_u CSI reports based on priority in the case of uplink control information (UCI) on the PUSCH without data and / or hybrid automatic repeat request (HARQ). Additionally, the WTRU may report dummy information in the Nr-Nu CSI reports based on priority to avoid rate matching handling of the PUSCH in other cases.

[0091] Artificial intelligence (AI) may be implemented as described herein. Artificial intelligence may be defined as behavior exhibited by a machine. Such behavior may mimic cognitive functions to sense, reason, adapt, and / or perform.

[0092] Machine learning (ML) and ML principles may be implemented as described herein. Machine learning may refer to a type of algorithm that solves problems based on learning through experience (e.g., data) without being explicitly programmed (e.g., by configuring a set(s) of rules). ML may be considered a subset of AI.

[0093] Different ML paradigms can be envisioned based on the nature of the data and / or feedback available to the learning algorithm. For example, supervised learning techniques can involve learning a function that maps inputs to outputs based on labeled training examples, where each training example can be a pair of input and / or corresponding output. For example, unsupervised learning techniques can involve detecting patterns in data without pre-existing labels. Reinforcement learning techniques can involve performing a sequence of actions in an environment to maximize a cumulative reward.

[0094] It may be possible to apply ML algorithms using a combination and / or interpolation of the techniques described herein. For example, semi-supervised learning techniques can use a combination of small amounts of labeled data with large amounts of unlabeled data during training. Semi-supervised learning can be somewhere between unsupervised learning (e.g., without labeled training data) and supervised learning (e.g., with only labeled training data).

[0095] Deep learning (DL) may be implemented as described herein. DL may refer to a class of ML algorithms that employ artificial neural networks loosely inspired by biological systems (e.g., specifically deep neural networks (DNNs)). DNNs may be a special class of machine learning models inspired by the human brain, in which inputs may be linearly transformed and / or passed multiple times through nonlinear activation functions. DNNs may consist of multiple layers, each consisting of a linear transformation and / or a given nonlinear activation function.

[0096] DNNs can be trained using training data via a backpropagation algorithm. DNNs can be implemented in various domains (e.g., speech, vision, and / or natural language, etc.) and / or for various machine learning settings (e.g., supervised, unsupervised, and / or semi-supervised). The term AI / ML-based methods and / or processes can refer to achieving behavior and / or adapting to requirements by learning based on data without explicit configuration of a sequence of action steps. Such methods can make it possible to learn complex behaviors that may be difficult to specify and / or implement when using legacy methods.

[0097] CSI compression may be implemented as described herein. CSI reports to indicate a precoder for a MIMO system may be compressed using a system such as that shown in FIG. 4. Specifically, FIG. 4 shows an example of a block diagram 400 of a CSI compression framework. The CSI compression framework may consist of an AE trained to compress a channel matrix 402 and / or a preprocessed channel matrix 404. The channel matrix 402 at the WTRU 450 may be preprocessed and / or compressed using an encoder section 406 of an autoencoder. The encoder output 408 may be quantized at 410 and / or fed back 412 to the gNB 460. At the gNB 460 side, inverse quantization 414 may be applied, and / or the output 416 of the inverse quantizer may be provided to a decoder section 418 of the autoencoder. The output 420 of the decoder 418 may be post-processed at 422 to obtain a channel matrix 424 of the precoder.

[0098] The end-to-end performance of an autoencoder can be measured using the normalized mean squared error (NMSE) metric. NMSE is

[0099]

number

[0100] where:

[0101]

number

[0102] is the reconstructed channel matrix 424 at the gNB, and E{.} denotes the averaging operation over multiple samples.

[0103] Quantization may be implemented as described herein. Quantization may be used to convert continuous values ​​into a set of discrete values. Various quantization methods may include uniform and / or non-uniform, as shown in FIG. 5, as well as vector quantization, as shown in FIG. 6.

[0104] 5 shows examples of scalar uniform 500 and / or non-uniform 550 quantizers. A non-uniform quantizer 550 may be used when input values ​​occur more frequently around a particular value.

[0105] 6 shows an example diagram 600 of vector quantization. In vector quantization, an input vector 602 may be compared to one or more (e.g., all) vectors 604 in a code block 606, and / or an index of the closest vector 608 may be sent to the receiving end at 610. Characteristics of the input vector 602 may determine the vector 604 in the code block 606.

[0106] The performance of a quantizer can be measured using quantization noise (QN). The average QN associated with a quantizer is given by:

[0107]

number

[0108] where z in and z q denote the input and output of the selected quantizer, respectively.

[0109] Systems and / or methods proposed herein for adaptive (e.g., first-order) CSI quantization may address one or more problems as described herein. Channel knowledge (e.g., accurate knowledge) may be included for downlink scheduling and / or link adaptation purposes for both SU-MIMO and / or MU-MIMO. The gNB may obtain CSI from the WTRU for scheduling and / or link adaptation purposes. The gNB may enable channel estimation at the WTRU using a DL CSI reference signal (CSI-RS) and / or may enable it by feeding back estimated CSI (e.g., implicit CSI, CQI, PMI, RI, and / or LI) in (one or more) WTRU CSI reports.

[0110] However, because NR supports several antenna ports, there may be overhead associated with CSI feedback reporting. There may be overhead for CSI Type II codebooks (e.g., Rel-15 CSI Type II and / or Rel-16 / 17 Type II / eType II CSI codebooks). The overhead may be expected to increase further as the system bandwidth and / or the number of antennas increase in B5G massive MIMO systems.

[0111] AI, including ML, may be applied to reduce CSI feedback overhead and / or improve CSI feedback accuracy with similar (e.g., the same) overhead through AI / ML-based compression, prediction, and / or both compression and / or prediction.

[0112] However, there can be various challenges associated with supporting artificial intelligence in 3GPP communication systems and / or within the 3GPP communication protocol stack. In the case of AE-based techniques, the encoder-decoder pair can be matched for correct end-to-end processing.

[0113] A CSI framework using an AE architecture can include a quantization process to reduce the number of feedback bits. There may be two approaches to quantization step processing. In a first approach, part of the training of the AE architecture can include quantizer selection (e.g., quantizer-aware training). In a second approach, the AE training and / or the quantization process(es) may be operated separately (e.g., quantizer-unaware training). In the first approach, including a quantizer in the training of the AE may limit the flexibility of the compressed output by requiring separately trained AEs for a given overhead. If the first approach is used, frequent switching between NN structures may be required in both the WTRU and / or the gNB depending on the available feedback overhead.

[0114] Vector quantization may be another potential method for addressing uneven encoder distribution. The values ​​of the encoder output of a CSI-compressed AE may be more frequent at low amplitudes (e.g., unevenly distributed). Therefore, one or more (e.g., conventional) uniform quantizers may not be optimal for AE-based compressed output. Furthermore, each encoder output of an AE may exhibit different characteristics for how the values ​​are distributed. Each trained AE may also exhibit different characteristics for the distribution of encoder output values.

[0115] However, vector quantization can be computationally intensive due to the comparison of an input vector with one or more (e.g., all) vectors in a codeblock. Furthermore, updating a codeblock in a vector quantizer can incur overhead.

[0116] Methods and / or enablers for adaptive quantization are separate from AE training but may be built based on the trained AE. Exemplary methods can address issues related to quantizer-aware AE training. The proposed adaptive (e.g., first-order) quantizer can adapt to one or more characteristics of the values ​​of the encoder output of the AE. The quantizer resolution can be readily changed to fit the compressed information to varying UCI overhead. Thus, the proposed system and / or method can provide an adaptive and / or flexible framework for determining quantizer(s) that can be applied to one or more (e.g., arbitrary) information compressions (e.g., using AEs).

[0117] Systems and methods for adaptive CSI quantization are described herein. For example, quantization of compressed information at the output of an encoder can affect the quality of the reconstructed information at the receiving end. In wireless communications, CSI compression may be an example of information compression achieved using AE. Designing an adaptive (e.g., first-order) quantizer constructed based on the characteristics of the encoder output values ​​can increase the efficiency of quantization.

[0118] Figure 7 shows a high-level diagram of the adaptive quantization process for CSI compression. A quantizer 702 may be constructed based on the decision(s) of a quantizer selection and / or configuration block 704. The quantizer selection and / or configuration block 704 may determine the quantizer parameter(s) 706 based on the output values ​​708 of an encoder 710. The quantizer parameter(s) 706 may be fed back 712 via UCI to the gNB 760 for implementation of an inverse quantizer 714. The quantized compressed CSI 716 is fed back to the gNB via 718.

[0119] 7 may not be limited to communications from the WTRU 750 to the gNB 760. The block diagram of the autoencoder 710 and / or the quantizer 702 may apply to communications from the gNB 760 to the WTRU 750, as well as WTRU-to-WTRU communications. Additionally or alternatively, the systems and / or methods described herein may not be limited to CSI compression. Rather, the systems and / or methods described herein may apply to one or more (e.g., any) information compression processes, including quantization.

[0120] Adaptive CSI quantization is also described herein, including, for example, that the WTRU may have the ability to select and / or determine parameter(s) of an adaptive (e.g., first-order) quantizer and / or monitor their performance. Enabling an adaptive (e.g., first-order) quantizer may include processes including initial configuration, quantizer parameter determination, parameter feedback, and quantization performance monitoring and / or reporting.

[0121] Configuration of adaptive quantization, including, for example, non-codebook-based methods and / or procedures for configuration of an adaptive (e.g., first-order) CSI quantizer, is also described herein. The systems, configurations, and / or methods may be applicable to one or more (e.g., any) types of quantization, including uniform, non-uniform, and / or vector-based. The WTRU may be configured, indicated, and / or requested to determine one or more (e.g., multiple) CSI quantizers via one or more of RRC, MAC CE, and / or DCI.

[0122] The WTRU's configuration for performing adaptive quantization may include one or more of the criteria described herein. For example, the approach the WTRU uses for CSI quantization may include determining a quantizer as part of an AI / ML-based CSI framework (e.g., compression, prediction, and / or joint compression and / or prediction). The WTRU may determine the quantizer separately from the AI / ML-based CSI framework.

[0123] The WTRU may determine the number of CSI quantizers. The WTRU may determine the number of CSI quantizers using a single-quantizer method, in which one quantizer may be determined for one or more (e.g., all) encoder outputs, and / or may determine it using a multi-quantizer method, in which multiple quantizers may be determined, one for each encoder output.

[0124] The WTRU may use one or more types of methods to determine the one or more (e.g., multiple) CSI quantizers by the WTRU. For example, the method may include a distribution method in which the WTRU calculates a distribution of encoder output values ​​using history samples. In one example, the gNB may indicate to the WTRU a maximum, minimum, and / or range of the number of one or more history samples to use. The WTRU may then select one or more samples within the range. In one example, the WTRU may use a particular statistical distribution for the determination of the distribution of the encoder output values. The type of method the WTRU uses to determine the one or more (e.g., multiple) CSI quantizers may include a clustering-based method in which the WTRU calculates a centroid of the encoder output values ​​using one or more history samples. In an example, the gNB may indicate to the WTRU how many centroids to determine for the CSI quantizer(s).

[0125] The WTRU may use performance metrics (e.g., KPIs) to monitor the performance of the adaptive (e.g., first-order) quantizer determined by the WTRU. In an example, the KPIs used to monitor the performance of the adaptive (e.g., first-order) quantizer determined by the WTRU may include quantization noise. The quantization noise may include a configuration including one or more (e.g., multiple) thresholds for the CSI quantization noise. The KPIs used to monitor the performance of the adaptive (e.g., first-order) quantizer determined by the WTRU may include NMSE. In an example, the NMSE may include a configuration including one or more (e.g., multiple) thresholds for the NMSE.

[0126] The WTRU may receive configuration information from a network node (e.g., a gNB). The configuration information may indicate multiple channel state information (CSI) quantization settings and / or criteria for selecting from multiple CSI quantization settings. The WTRU may perform one or more CSI measurements on one or more reference signals. The WTRU may encode the one or more CSI measurements. The WTRU may select a CSI quantization setting from the multiple quantization settings based on the criteria indicated in the configuration message. The WTRU may quantize the encoded one or more CSI measurements using the selected CSI quantization setting. The WTRU may send a feedback report to the network node including the quantized one or more CSI measurements and / or an indication of the selected CSI quantization setting.

[0127] The configuration of the WTRU to perform adaptive quantization may include configuring fallback CSI quantization. The fallback CSI quantization may include configuring a second-order quantizer and / or a set of quantizers as a fallback. For example, the WTRU may use a legacy (e.g., uniform) CSI quantizer as a fallback.

[0128] The WTRU may signal the determined quantizer and / or set of quantizers to the gNB using a feedback mechanism. The feedback mechanism may include an explicit signaling method. In an example, the WTRU may signal the selected quantizer and / or set of quantizers as part of a CSI report (e.g., using the PUSCH and / or PUCCH depending on the CSI report configuration). In an example, the WTRU may signal the selected quantizer and / or set of quantizers using uplink control signaling (e.g., using the UCI and / or PUCCH). In an example, the WTRU may be configured to signal the selected quantizer and / or set of quantizers using a MAC CE. The feedback mechanism by which the WTRU signals the determined quantizer and / or set of quantizers to the gNB may include an implicit signaling method. In an example, the WTRU may implicitly signal a selected quantizer and / or set of quantizers by selecting some UL resources (e.g., PUSCH, PUCCH, and / or sounding reference signal (SRS)).

[0129] The WTRU may determine the quantizers in a self-configured manner (e.g., including criteria regarding type, number, method, and / or KPIs, etc.). The configuration of adaptive quantization by the WTRU may utilize implicit methods to determine parameters. Such parameters may include methods that are implicitly determined based on parameters, training conditions, and / or use cases (e.g., predefined conditions).

[0130] Based on a configuration and / or set of configurations, the WTRU may receive and / or request one or more RSs and / or sets of RSs to determine a quantizer.

[0131] The quantization parameters may be determined using the gNB configuration as described herein. The procedure by which the WTRU determines the quantization parameters using the indicated configuration is described herein.

[0132] The WTRU may be configured with a type of method for determining the parameters of the CSI quantizer. For example, the method may be distribution-based and / or clustering-based. Additionally or alternatively, the WTRU may be configured with a single quantizer option, in which the same quantizer may be used for one or more (e.g., all) encoder outputs, and / or a multiple quantizer option, in which different quantizers may be used for the encoder outputs.

[0133] A distribution-based quantizer method may be described herein. The values ​​of the encoder output of the AE may be unevenly distributed between the minimum and / or maximum values ​​of the encoder output. The distribution may be characterized as a Gaussian distribution and / or any other distribution. Figure 8 shows an example of a distribution of encoder output values ​​based on a Gaussian distribution, including a probability density function (PDF) and / or a cumulative distribution function (CDF).

[0134] A non-uniform quantizer can quantize the encoder output values. The non-uniform quantizer can be constructed, for example, by one or more (e.g., arbitrary) methods that are implicit to both ends given the parameters of the distribution. For Gaussian-distributed encoder output values, the mean and / or variance of the distribution can be sufficient to determine the parameters of the distribution and / or construct the quantizer.

[0135] To calculate the PDF and / or CDF of the encoder output values, the WTRU may collect historical sample values ​​for the output of the AE's encoder. The number of samples may be predefined and / or implicit depending on the gNB indication.

[0136] An exemplary method for constructing a non-uniform quantizer may be a CDF transform followed by a uniform quantizer, given the distribution and / or quantizer parameters. The transform F(x) may be uniformly distributed between 0 and 1, where F represents the CDF of x, for example, assuming x represents a non-uniformly distributed value to be quantized.

[0137] 9 illustrates a non-uniform quantization process 900, which includes determining the parameters of the distribution. The encoder output values ​​Y 902 may be applied to a CDF transform 904. The transformed values ​​Y F 906 may be applied to a uniform quantizer 908. The output of the quantizer Y q 910 may contain a particular number of bits depending on the resolution of the quantizer 908 .

[0138] To determine the number of steps (e.g., resolution) of the quantizer, in examples, the UCI feedback overhead and / or the number of encoder outputs may be used. For example, if the number of encoder outputs is N=10 and / or the UCI overhead is U=60 bits, each output value may be represented by U / N=6 bits, resulting in 64 steps in the quantizer.

[0139] If the distribution is implicit and / or indicated by the WTRU, the WTRU may feed back parameters of the distribution (e.g., mean and / or variance, assuming Gaussian distributed output values) to the gNB. If the distribution cannot be identified as an explicit distribution, a CDF obtained from a histogram of the output values ​​may be used for the CDF conversion operation. In that case, the CDF may be fed back to the gNB.

[0140] Cluster-based quantizer methods are also described herein. For example, the WTRU may build one or more clustering-based quantizers. History samples from the encoder's output may be collected. The number of history samples may be implicit and / or indicated by the gNB. A clustering method, such as k-means, may partition the history samples into one or more clusters (e.g., a specific number).

[0141] 10 shows an example of clustering 1000 for encoder output values, assuming the outputs are Gaussian distributed with mean 0.1 and / or variance 0.1. For illustrative purposes, 100 historical samples may be collected and / or partitioned into 8 clusters using the k-means method. FIG. 10 shows dots 1010, which may represent sample values, and / or x marks 1020, which may represent cluster means (e.g., centroids).

[0142] For example, the centroids may be non-uniformly spaced depending on the history samples, as shown in Figure 10. Following the calculation of the cluster centroids, the centroid values ​​may be fed back to the gNB for the inverse quantization process.

[0143] 11 shows the quantization process in the WTRU after the calculation of the centroids. For each value Y to be quantized 1102, a non-uniform quantizer 1150 can search for the closest centroid in the centroid set C calculated in the clustering stage at 1104. The quantizer output Y q 1106 may represent an index of the centroid and / or include a specific number of bits (e.g., depending on the number of centroids). The number of clusters and / or centroids may be determined based on the UCI overhead and / or the number of encoder outputs. For example, if the encoder has N=10 outputs and / or the UCI overhead is U=60 bits, U / N=6 bits may be used to represent the cluster centroid, which means that 64 clusters may be calculated.

[0144] The modification may be implemented based on the number of quantizers, as described herein. For example, the WTRU may be shown to construct a single quantizer for use on one or more (e.g., all) encoder outputs and / or multiple quantizers for separately quantizing the encoder outputs. If the WTRU is shown to construct a single quantizer, history samples from one or more (e.g., all) outputs of the encoder may be combined. If a distribution-based method is shown, the combined samples may be used to calculate a single histogram for the encoder output values, which is used to calculate the PDF and / or CDF. If a clustering-based method is shown, the combined samples may be used to calculate a single set of centroids.

[0145] If the WTRU is shown to implement multiple quantizers, samples from each encoder output may be collected separately. For example, assuming there are 10 encoder outputs, 10 sets of history samples may be collected. If a distribution-based method is shown, a separate histogram may be calculated for each encoder output. If a clustering-based method is shown, a separate set of centroids may be calculated for each encoder output. Figure 12 shows an example of a quantization structure 1200 for multiple quantizers, in which each encoder output 1202a-c may be applied to a separate quantizer 1204a-c.

[0146] A dequantization process using a distribution-based method is described herein. For example, when a receiving end, such as another WTRU and / or gNB, receives feedback regarding the quantization parameters, an inverse quantizer may be constructed. In the case of distribution-based quantization, the receiving end may have already received feedback regarding the parameters of the distribution. Figure 13 shows an example non-uniform inverse quantizer 1350 for the distribution-based method 1300, which calculates the received quantized encoder output values.

[0147]

number

[0148] 1302 may be dequantized using a uniform dequantizer 1304,

[0149]

number

[0150] 1302 is a scalar value

[0151]

number

[0152] Mapped to 1306. Scalar value

[0153]

number

[0154] 1306 performs the inverse CDF transform F to obtain the dequantized values ​​Y 1310. -1 (·)1308 may be applied.

[0155] Described herein is an inverse quantization process using a clustering-based method. For example, in a clustering-based approach, a receiving end (e.g., another WTRU and / or a gNB) may receive an indication of the centroid of a cluster. An inverse quantizer at the receiving end may map the received index of the centroid to the actual value of the centroid.

[0156] The parameters of the quantizer may be determined using WTRU self-configuration as described herein. The procedure by which the WTRU determines the quantization parameters using self-configuration is described herein.

[0157] For example, when a WTRU is configured for self-configuration, the WTRU may include determining one or more quantization performance metrics, a type of method for determining a quantizer, a number of quantizers, and / or a secondary quantizer. One or more (e.g., all) determined configurations may be fed back to the gNB.

[0158] The WTRU may determine the quantization performance metric. The selection of the quantization performance metric may be implicit based on the current state of the WTRU (e.g., speed and / or received signal strength indicator (RSSI), etc.). The WTRU may determine the quantization performance metric using a lookup table approach. The WTRU may determine the quantizer based on an index associated with a predefined lookup table. Example quantization performance metrics may include QN and / or NMSE.

[0159] In an example, the selection of the quantization performance metric may depend on the capability(s) of the WTRU. For example, if the UE has access to the decoder portion of the AE, the WTRU may select NMSE. Otherwise, the WTRU may select QN as the performance metric.

[0160] The WTRU may determine the type of method for determining the parameters of the quantizer. The selection of the quantizer type may be implicit, for example, based on the current state of the WTRU (e.g., speed and / or RSSI, etc.). The WTRU may determine the quantization type using a look-up table approach. Exemplary quantization types may be uniform, distribution-based, and / or clustering-based, etc.

[0161] In an example, the WTRU may construct one or more quantizers (e.g., distribution-based and / or clustering-based) and / or compare their performance, e.g., based on a selected quantization performance metric. The WTRU may use historical samples to construct the quantizers and / or use one or more (e.g., any) pre-defined quantizers. The performance comparison may be based on an average of performance metrics calculated using one or more (e.g., all) historical samples. The WTRU may select the quantizer with the most dominant performance metric.

[0162] The indication of the selected CSI quantization setting may include an indication of a quantization type, a quantization parameter, a quantizer granularity, and / or a quantizer resolution.

[0163] The WTRU may determine the parameters of the quantizer after the WTRU selects the type of quantizer. For example, in the case of a distribution-based method, the distribution parameters (e.g., mean and / or variance) and / or CDF may be fed back to the gNB.

[0164] The WTRU may determine the number of quantizers. The selection of the number of quantizers (e.g., a single quantizer for one or more (e.g., all) outputs, and / or one per output, etc.) may be implicit based on the WTRU's current state (e.g., speed and / or RSSI, etc.). In an example, the WTRU may use a look-up table approach to determine the number of quantizers. The WTRU may construct one or more (e.g., all) options for the number of quantizers and / or compare their performance based on a selected quantization performance metric. The WTRU may use history samples to construct the quantizers. The performance comparison may be based, for example, on an average of a performance metric calculated using one or more (e.g., all) history samples. The WTRU may select the option with the most dominant performance metric.

[0165] The WTRU may determine the second-order quantizer. The selection of the second-order quantizer may be implicit based on the WTRU's current state (e.g., speed, RSSI, etc.). The WTRU may determine the number of quantizers using a look-up table approach. In an example, the WTRU may build one or more (e.g., all) second-order quantizer options and / or compare their performance. The WTRU may select the option with the most promising performance. The WTRU may select the second-order quantizer based on a second criterion, which includes the WTRU's current state speed or the current state of the received signal strength indicator (RSSI). The performance of the first and second-order quantizers may be measured via quantization noise (QN), cosine similarity, and squared generalized cosine similarity (SGCS).

[0166] The WTRU may feed back quantizer parameters. One or more (e.g., all) determined configurations, such as a quantization performance metric, a quantizer type, quantizer parameters, a number of quantizers, and / or a second-order quantizer, may be fed back to the gNB. The WTRU may feed back information for the gNB to construct an inverse quantizer.

[0167] The feedback of the quantization parameters may be performed as described herein. One or more methods in the context of quantization performed by the WTRU and / or dequantization performed in the gNB may be considered examples and / or may be applicable (e.g., equally applicable) in any case where quantization and / or dequantization may be performed by the WTRU, the gNB, and / or any other network node.

[0168] One or more systems and / or methods described herein in the context of a WTRU-to-gNB transmission may be considered examples. These systems and / or methods may be applied (e.g., equally applicable) to a WTRU-to-WTRU transmission, a gNB-to-WTRU transmission, and / or a transmission between one or more (e.g., any) WTRUs.

[0169] One or more systems and / or methods described herein in the context of quantizing CSI feedback may be considered examples. The systems and / or methods may be applied to (e.g., are equally applicable to) quantizing any type of feedback and / or to any transmission.

[0170] Quantization parameter feedback may include exemplary quantizer feedback content, as described herein. In examples, the term quantization feedback may refer to an indication of a recommended quantization type and / or set of quantization types (e.g., uniform quantization, non-uniform quantization, scalar quantization, and / or vector quantization, etc.), a method for constructing a quantizer (e.g., distribution-based and / or clustering-based, etc.) and / or its parameterization (e.g., mean and / or variance of the distribution, cluster centroid index, centroid value, and / or number of clusters, etc.), the granularity and / or number of applicable quantizers (e.g., a quantization specific to each encoder output, etc.), and / or a parameterization of the quantizer (e.g., a quantization specific to each encoder output, etc.). The quantization feedback may include one or more of: a quantizer, a common quantizer for each encoder output, and / or a single quantizer for all encoder outputs, an indication of the quantizer resolution (e.g., quantization range, quantization error, quantization overhead, and / or number of bits, etc.), an indication of a fallback quantizer and / or set thereof, an indication of operating conditions under which the quantizer may be implicitly determined (e.g., use case, RSSI, SINR, WTRU speed, performance metric associated with the quantizer, etc.), and / or their parameterization. In an example, the WTRU may be configured to send quantization feedback to indicate a recommended quantization parameter and / or a set of recommended quantization parameters to indicate a selected quantizer and / or a subset of quantizers.

[0171] The feedback of the quantization parameter may include transmitting quantizer feedback in the MAC-CE. In an example, the WTRU may be configured to transmit the quantization feedback semi-statically. The WTRU may receive an activation command from the gNB, for example, in the DL MAC CE, to activate semi-persistent reporting of the quantization feedback. The WTRU may transmit the quantization feedback in the UL MAC CE.

[0172] The WTRU may receive a quantization feedback configuration at the DL MAC CE that activates quantization feedback reporting. For example, the configuration may include a type of quantization feedback, a content of the quantization feedback, a size of the quantization feedback, and / or a periodicity of the quantization feedback, etc. The WTRU may receive a configuration associated with the quantization feedback transmission in an RRC message. In an example, the WTRU may receive one or more (e.g., multiple) sets of configurations associated with the quantization feedback transmission in the RRC message. The WTRU may determine an appropriate quantization feedback configuration, for example, based on an indication in the DL MAC CE. In an example, the WTRU may be configured with a set of quantization parameters and / or their ranges. The WTRU may transmit quantization feedback, for example, such that the reported parameters are within the configured set or range. The WTRU may stop / pause quantization feedback transmission upon receiving a deactivation command from the gNB, for example, possibly at the DL MAC CE.

[0173] The feedback of the quantization parameters may include transmitting joint quantizer feedback and / or CSI feedback in UCI. In an example, the WTRU may be configured to dynamically transmit quantization feedback. For example, the WTRU may transmit quantization feedback in UCI and / or L1 report / feedback. In an example, the WTRU may transmit quantization feedback as part of CSI feedback. For example, the WTRU may receive one or more aspects of a quantization feedback configuration in CSI-MeasConfig. The WTRU may transmit the quantization feedback as a new reporting quantity. The WTRU may transmit the quantization feedback if the reporting quantity is configured as cri-PMI-QNF, where the QNF may indicate that the WTRU should include the quantization feedback with CSI feedback carrying a CRS resource indication and / or a precoding matrix indication. The WTRU may determine the content of the quantization feedback based on an indication received, for example, in an aperiodic CSI report request message.

[0174] The feedback of the quantization parameters may include transmitting independent quantizer feedback and / or CSI feedback in UCI. In an example, the WTRU may be configured to transmit quantization feedback independently from CSI feedback. The term "independently" herein may refer to transmissions performed on different time scales and / or initiated by different triggers, conditions, and / or on control messages dedicated to quantization feedback. In an example, the WTRU may be configured with dedicated PUCCH resources for quantization feedback.

[0175] The WTRU may determine the type and / or content of the quantization feedback as a function of the allocation of PUCCH resources. For example, the WTRU may be configured with a first set of PUCCH resources associated with a first set of quantizer configurations / feedback and / or a second set of PUCCH resources associated with a second set of quantizer configurations / feedback. The WTRU may, for example, select PUCCH resources based on the determined quantizer feedback. The WTRU may be configured with PUSCH resources and / or PUCCH resources for quantizer feedback transmission. The WTRU may determine whether PUSCH and / or PUCCH resources should be selected for transmission based on the content of the quantization feedback.

[0176] The quantization parameter feedback may include an implicit indication of quantizer feedback. In an example, a WTRU may be configured with multiple PUCCH resources, and each PUCCH resource may be associated with a quantizer type and / or quantizer parameterization. The WTRU may implicitly indicate the quantizer parameterization, for example, based on a selection of a PUCCH resource. The WTRU may indicate a first portion of the quantizer feedback based on the selection of a PUCCH resource and / or indicate a second portion of the quantizer feedback in the content of a UCI transmission on the selected PUCCH resource. In an example, the WTRU may indicate a quantizer type via the selection of a PUCCH resource and / or indicate a recommended quantizer parameterization via quantization feedback transmitted on the selected PUCCH resource. The WTRU may indicate the quantizer feedback based on the selection of a PUCCH resource and / or CSI feedback in the content of a UCI transmission on the selected PUCCH resource. Based on selecting the pre-configured PUCCH resource, the WTRU may indicate that the WTRU applies a particular quantization and / or may transmit quantized CSI feedback on the selected PUCCH resource.

[0177] The WTRU may transmit quantizer feedback based on one or more preconfigured conditions and / or triggers. For example, the WTRU may be configured to trigger quantizer feedback transmission when the performance of the current active quantizer falls below a threshold. For example, the WTRU may be configured to trigger quantizer feedback transmission when a change in quantizer performance exceeds the performance of the current active quantizer by a preconfigured threshold, under the hypothesis that the gNB applies the WTRU-recommended quantizer feedback. The WTRU may trigger quantizer feedback transmission upon a change in the CSI feedback configuration and / or active bandwidth portion. For example, the WTRU may trigger quantizer feedback transmission based on preconfigured WTRU measurements. Furthermore, the WTRU may trigger quantizer feedback when the SINR, RSRP, etc. change by a preconfigured threshold.

[0178] The WTRU may apply updates to the quantizer parameters based on an indication received from the gNB. For example, the WTRU may receive an indication from the gNB (e.g., in a DCI) acknowledging that the gNB has applied the recommended quantizer parameters reported by the WTRU in the quantizer feedback. The indication may be implicit, and the WTRU may update the quantizer parameters according to the most recent transmission of the quantizer feedback. The WTRU may receive an explicit indication from the gNB indicating a logical identity associated with the quantizer feedback transmitted by the WTRU. The WTRU may apply the quantizer update at a preconfigured time offset from the DCI carrying the indication. The WTRU may receive a complete (e.g., full) configuration of the quantizer parameterization in a MAC CE and / or RRC message.

[0179] Quantization performance may be monitored and / or reported as described herein. The WTRU may monitor the quantization performance based on one or more pre-configured metrics (e.g., NMSE, QN, cosine similarity, SGCS, and / or PDSCH performance). The WTRU may monitor the performance of the selected / designed / primary quantizer. In an example, the primary quantizer may indicate the quantizer used at the current moment, and / or the (one or more) secondary quantizers may represent one or more (e.g., all) of the other configured quantizers that are not used at the current moment. The WTRU may monitor the quantization performance through comparing the performance of the primary quantizer with the performance of the secondary quantizer (e.g., if configured).

[0180] The performance may be based on one or more pre-configured metrics (e.g., NMSE, QN, cosine similarity, SGCS, physical downlink shared and / or channel (PDSCH) performance). The WTRU may use the QN metric to monitor the primary quantizer performance. The QN associated with any of the pre-configured quantizers may be

[0181]

number

[0182] where z in and z q and denote the output of the AI ​​encoder model and the output of the selected quantizer, respectively. The WTRU may, for example, determine whether the QN of the first quantizer is equal to the pre-configured QN threshold (ε th), the WTRU may still use the first quantizer in the next transmission to quantize the compressed CSI. For example, if the QN of the first quantizer exceeds a pre-configured QN value, the WTRU may calculate the QN of the second quantizer and / or select the one that results in the smallest QN. If the QN of the first quantizer is still smaller than the smallest QN across all second quantizers, the WTRU may still use the first quantizer in the next transmission and / or report that the configured QN threshold is not met.

[0183] The WTRU may transmit quantization information (e.g., a selected quantizer and / or its parameters) as part of a CSI report, and / or the WTRU may transmit quantization performance either periodically and / or aperiodically (e.g., when triggered to report one or more metrics associated with the quantization performance).

[0184] The WTRU may report quantization information as part of the ML-based CSI report. The quantization information may include first-order quantizer performance collected over a specific period of time. The quantization information may also include selected first-order quantizer parameters and / or portions thereof.

[0185] For performance monitoring, the WTRU may also send one or more of the following parameters: The WTRU may send QuantizerPerformanceThreshold for performance monitoring, which includes a binary value indicating whether the quantizer performance meets a preconfigured threshold.

[0186] The WTRU may send QuantizerSwitch for performance monitoring. QuantizerSwitch may contain a binary value that indicates whether the WTRU has switched to one of the second-order quantizers and / or whether it is still using the first-order quantizer that was used in the previous transmission.

[0187] The WTRU may send QuantizerPerformanceMetric for performance monitoring. QuantizerPerformanceMetric may take several values ​​(e.g., four), with each value indicating which metric is used to monitor quantization performance. The gNB may also configure QuantizerPerformanceMetric.

[0188] The WTRU may send QuantizerPerformance for performance monitoring. QuantizerPerformance may include parameters indicating the first-order quantizer performance and / or may take several values ​​depending on the QuantizerPerformanceMetric value. For example, QuantizerPerformance may take up to 16 NMSE levels with 1 dB resolution if the QuantizerPerformanceMetric value is set to NMSE.

[0189] The WTRU may be configured to indicate one or more of the parameters described herein (e.g., periodically, semi-periodically, and / or aperiodically) and / or some parameters may be configured to report periodically and / or other parameters aperiodically. For example, the WTRU may report quantizer performance every N slots whenever the performance falls below some preconfigured threshold and / or whenever the WTRU detects the use of a second-order quantizer. The WTRU may report the quantization performance periodically. The gNB may indicate whether the WTRU should switch to a second-order quantizer. The WTRU may explicitly send parameters of a preferred second-order quantizer and / or the gNB may configure the WTRU to switch to any second-order quantizer if the gNB indicates that the WTRU should switch to a second-order quantizer.

[0190] The quantizer fallback may be implemented as described herein. The quantizer fallback may include a quantizer error event configuration. The WTRU may monitor the performance of a quantizer and / or a set of quantizers (e.g., a primary quantizer and / or a secondary quantizer) to detect when a quantizer error event occurs. The WTRU may be configured to detect the error event using a QuantizerErrorThreshold. For example, the QuantizerErrorThreshold may represent the maximum tolerable quantization noise when the WTRU is configured with QN for the QuantizerPerformanceMetric. RRC signaling may configure the WTRU with the QuantizerErrorThreshold and / or the QuantizerPerformanceMetric.

[0191] Quantizer fallback may be implemented using several triggers. For example, quantizer fallback can be triggered upon detection of a quantizer error event. The trigger for quantizer fallback can include measured performance of a configured quantizer (e.g., or set of quantizers) meeting a configured threshold. The trigger for quantizer fallback can include measured performance of a configured primary quantizer (e.g., or set of primary quantizers) and / or, if configured, measured performance of a secondary quantizer (e.g., or set of secondary quantizers) meeting a configured error threshold. An error event (e.g., a trigger for quantizer fallback) can be detected when a measured QuantizerPerformance (e.g., NMSE and / or QN) exceeds a configured error threshold (e.g., NMSE and / or QN). In an example, if a WTRU is configured with NMSE and / or quantization noise (QN) as the QuantizerPerformanceMetric, the threshold can be QuantizerErrorThreshold.

[0192] The quantizer fallback may be implemented while the WTRU collects history samples (e.g., compressed values ​​at the ML encoder output) to determine the quantizer parameters. The number of history samples may be predefined and / or configured by the gNB, for example, by RRC signaling (e.g., in CSI-MeasConfig). The WTRU may start counting the number of compressed CSI samples (e.g., for history sample collection) upon a change in the CSI-RS configuration. The WTRU may start counting the number of compressed CSI samples (e.g., for history sample collection) when the WTRU uses a new quantizer.

[0193] The quantizer fallback mechanism may be implemented as described herein. For example, when quantizer fallback triggers, the WTRU may switch to a second-order one of the quantizers (e.g., a second-order set). The WTRU may switch if the performance of the second-order quantizer is better than a configured quantization error threshold. When quantizer fallback triggers, the WTRU may switch to a default quantizer if the second-order quantizer and / or set of quantizers is not configured. When quantizer fallback triggers, the WTRU may continue to use the current quantizer and / or send an indication to the gNB requesting a switch to a fallback quantizer. Furthermore, the WTRU may switch legacy CSI reporting if the performance of the default quantizer is worse than a configured quantization error threshold.

[0194] Examples of default quantizers may include a uniform quantizer with a predefined number of bits and / or a non-uniform quantizer with predefined parameters.

[0195] Quantizer fallback event reporting may be implemented as described herein. When the WTRU detects that a quantizer error event has occurred, the WTRU may report a quantizer fallback. The report may include QuantizerPerformance, which may include measured quantizer performance. The report may include QuantizerSwitch, which may indicate a flag indicating whether the WTRU switched quantizers. The report may include a quantizer ID, which may indicate the ID of a new quantizer, for example, when the WTRU autonomously switched quantizers upon detection of an error event. In an example, the WTRU may send a quantizer switch request to the gNB upon detecting a quantization error event.

[0196] Quantization codebook-based adaptive CSI quantization may be implemented as described herein. The WTRU may have the capability to select and / or determine parameters of an adaptive (e.g., first-order) quantizer and / or monitor performance. The WTRU may determine one or more parameters of a codebook for a pre-defined CSI quantizer. Enabling the adaptive (e.g., first-order) quantizer may include steps of initial configuration, quantizer parameter determination, parameter feedback, and quantization performance monitoring and / or reporting.

[0197] Adaptive quantization configuration is described herein. Codebook-based methods and / or procedures for adaptive (e.g., first-order) CSI quantizer configuration are also described herein. The systems and / or methods may be applicable to one or more (e.g., any) types of codebooks of quantizers, including uniform, non-uniform, and / or vector-based. The WTRU may be configured, indicated, and / or requested to select one or more (e.g., multiple) CSI quantizers via one or more RRC, MAC CE, and / or DCI (e.g., both dynamic and / or semi-static options are possible) to align the quantizers at the WTRU and / or gNB. For example, the WTRU may receive configuration information from a network node. As discussed below, the configuration information may include a CSI quantization setting and / or criteria for selecting from CSI quantization settings.

[0198] The WTRU's configuration for performing codebook-based adaptive CSI quantization may include one or more criteria (e.g., for selecting from CSI quantization settings). The WTRU may use different approaches for CSI quantization. In an example, the approaches may include a quantizer determined as part of an AI / ML-based CSI framework (e.g., compression, prediction, and / or joint compression and / or prediction). Furthermore, the WTRU may approach CSI quantization by determining a quantizer separately from the AI / ML-based CSI framework. The WTRU may determine one or more parameters of a codebook for a pre-defined CSI quantizer.

[0199] The WTRU may determine the number of CSI quantizers from a predetermined codebook. The WTRU may determine the number of CSI quantizers from a predetermined codebook using a single-quantizer method, where the codebook may determine one quantizer for one or more (e.g., all) encoder outputs. In an example, the WTRU may determine the number of CSI quantizers from a predetermined codebook using a multiple-quantizer method, where the codebook may determine multiple quantizers, one for each encoder output.

[0200] Different types of methods may be implemented by the WTRU to determine one or more CSI codebook quantizers. The method implemented by the WTRU to determine one or more (e.g., multiple) CSI codebook-based quantizers may include one or more predefined codebooks and / or sets of codebooks known to the WTRU and / or gNB. The WTRU may select the respective quantizer and / or set of quantizers based on an index that references a predefined lookup table. In an example, the WTRU may determine one and / or set of codebook parameters associated with a predefined CSI quantizer.

[0201] The performance metric (e.g., KPI) may monitor the performance of the adaptive (e.g., first-order) quantizer determined by the WTRU. The KPI may include QN. The configuration may include one or more (e.g., multiple) thresholds for CSI quantization noise. The KPI used to monitor the performance of the adaptive (e.g., first-order) quantizer may include NMSE. In an example, the configuration may include one or more (e.g., multiple) thresholds for NMSE.

[0202] The configuration of the WTRU may include configuring fallback CSI quantization. The fallback CSI quantization may include configuring a second-order quantizer and / or a set of quantizers as a fallback. In an example, the WTRU may be configured to use a legacy (e.g., uniform) CSI quantizer as a fallback.

[0203] The WTRU can use different feedback mechanisms to signal the determined codebook-based quantizer and / or set of quantizers to the gNB to align the quantizers at the WTRU and / or the gNB. The feedback may be provided using explicit signaling methods. In an example, the WTRU may signal the selected quantizer and / or set of quantizers as part of a CSI report (e.g., using the PUSCH and / or PUCCH depending on the CSI report configuration). In an example, the WTRU may signal the selected quantizer and / or set of quantizers using uplink control signaling (e.g., using the UCI and / or PUCCH). The feedback may be provided using implicit signaling methods. The WTRU may implicitly signal the selected quantizer and / or set of quantizers by selection of some UL resources (e.g., the PUSCH, the PUCCH, and / or the SRS).

[0204] The WTRU may configure adaptive quantization to utilize an implicit method using a codebook to determine parameters. Such parameters may include those that are implicitly determined based on, for example, parameters, training conditions, and / or use cases (e.g., predefined conditions). The WTRU may receive and / or request one or more RSs and / or sets of RSs to determine a quantizer based on a configuration and / or set of configurations.

[0205] The quantizer may be selected as described herein. The WTRU may be configured with and / or maintain and / or design multiple quantizers and / or multiple sets of quantizers. For any one feedback, the WTRU may use a set of quantizers. The set of quantizers may define multiple quantizers used for different outputs of the AI / ML model. The set of quantizers may include fewer quantizers than the number of AI / ML model outputs. One or more quantizers may be applied to one or more outputs of the AI / ML model. The set of quantizers may include a single quantizer applied to one or more (e.g., all) outputs of the AI / ML model at a given time.

[0206] As described herein, a trigger may select a set of quantizers and / or a selection criterion may be implemented. As mentioned above, the WTRU may receive configuration information indicating multiple CSI quantization settings and / or criteria for selecting from multiple CSI quantization settings to align quantizers at the WTRU and / or gNB. As described herein, the WTRU may select one or more CSI quantization settings based on the criteria indicated in the configuration message.

[0207] The WTRU may consider a set of quantizers to be active if the WTRU can use the set to generate the payload of an UL transmission (e.g., as a feedback report). The WTRU may consider one set of quantizers as a fallback set of quantizers. The fallback set of quantizers may be active (e.g., always active). The WTRU may select a quantizer set from the group of the active set of quantizers. The WTRU may make a quantizer set selection per feedback report instance. The selection may be valid for multiple feedback report instances. An event may trigger the selection of a quantizer set.

[0208] The WTRU may trigger and / or (re)select a set of quantizers and / or perform the selection of the set of quantizers based on several factors (e.g., based on a configuration message). For example, the WTRU may trigger and (re)select a set of quantizers and / or perform the selection of the set of quantizers based on the performance of one or more sets of quantizers. The WTRU may test the performance of one or more sets of quantizers and / or select the set of quantizers that maximizes the performance. The performance may be determined based on a metric. The metric may include one or more of the quantization error (e.g., mean / average error and / or max / min error within the set of quantizers), compression rate, and / or minimum compression error (e.g., overall error of the AI / ML model and / or set of quantizers). A metric may be determined for each quantizer in the set of quantizers, for each output of the AI / ML model, averaged across all quantizers, and / or averaged across one or more (e.g., all) outputs of the AI / ML model.

[0209] The WTRU may trigger and / or (re)select a set of quantizers and / or perform quantizer set selection based on performance changes with a previously used set of quantizers. For example, the WTRU may select a new set of quantizers if the performance of the new set of quantizers is above a threshold and / or the performance of the previously used set of quantizers is below a threshold. The WTRU may select a new set of quantizers if the performance of the new set of quantizers is more favorable (e.g., possibly with an offset) than the performance of the previously used set of quantizers.

[0210] The performance of a set of quantizers may be defined as described herein. The performance of a previous set of quantizers may be determined based on the performance achieved for a previous transmission and / or feedback reporting instance. The performance of a previous set of quantizers and / or a new set of quantizers may be determined based on the expected performance for an upcoming transmission and / or feedback reporting instance.

[0211] The WTRU may trigger and (re)select a set of quantizers and / or perform quantizer set selection based on measurements. The feedback report may include measurements (e.g., inputs of an AI / ML model). The measurements may include one or more of RSRP, reference signal received quality (RSRQ), RSSI, CO, constant bit rate (CBR), SINR, CQI, RI, PMI, angle of arrival (AoA), angle of departure (AoD), Doppler spread, delay spread, Doppler shift, average delay, and / or WTRU speed. For example, the WTRU may select a first set of quantizers when experiencing a first set of channel characteristics and / or select a second set of quantizers when experiencing a second set of channel characteristics.

[0212] The WTRU may trigger and / or (re)select a set of quantizers and / or perform the selection of the set of quantizers based on parameters of the transmission and / or feedback report. For example, the WTRU may select a set of quantizers based on a priority of the feedback report. The priority of the feedback report may be configurable and / or may be determined by the WTRU as a function of the associated transmission. In an example, the WTRU may determine a threshold and / or an offset and / or a performance metric to select a set of quantizers as a function of a parameter (e.g., priority) of the feedback report.

[0213] The WTRU may trigger and / or (re)select a set of quantizers and / or perform the selection of the set of quantizers based on whether the transmission and / or feedback report collides with another transmission from the WTRU. For example, the WTRU may determine that resources of the feedback report overlap with another transmission from the WTRU. The WTRU may trigger and / or select a set of quantizers and / or select a set of quantizers based on whether the overlap will lead to multiplexing of multiple transmissions and / or dropping of lower priority transmissions.

[0214] The WTRU may trigger and / or (re)select a set of quantizers and / or perform quantizer set selection based on transmit power. For example, the WTRU may trigger and / or select a quantizer set based on transmit power, available transmit power, and / or power headroom.

[0215] The WTRU may trigger and / or (re)select a set of quantizers and / or perform quantizer set selection based on a transmission resource. For example, the WTRU may trigger and / or select a set of quantizers based on a feedback report resource (e.g., or parameters thereof) and / or a feedback resource type (e.g., PUCCH format). In an example, the WTRU may select a set of quantizers depending on the channel on which the transmission and / or feedback report occurs. For example, the WTRU may select a quantizer based on whether the feedback report is transmitted on the PUCCH and / or simultaneous PUCCH / PUSCH and / or as UCI on the PUSCH.

[0216] The WTRU may trigger and / or (re)select a quantizer set and / or perform the quantizer set selection based on channel availability. For example, the WTRU may select a quantizer set based on whether the WTRU can initiate a channel occupation time (COT) and / or share the COT for transmission. The WTRU may select a quantizer set in response to a listen-before-talk (LBT) procedure and / or based on parameters of the LBT procedure.

[0217] The WTRU may trigger and / or (re)select a set of quantizers and / or perform quantizer set selection based on a time to live. For example, the WTRU may select a quantizer set based on a remaining time to live if the time to live is expiring and / or has expired.

[0218] The WTRU may trigger and / or (re)select the quantizer set and / or perform the quantizer set selection based on the intended receiver. For example, the WTRU may select the quantizer set based on the gNB, cell, and / or transmission / reception point (TRP).

[0219] The WTRU may trigger and / or (re)select the set of quantizers and / or perform the selection of the set of quantizers based on the performance of the associated transmissions. For example, the WTRU may determine the set of quantizers based on the performance of the associated UL and / or DL ​​transmissions (e.g., transmissions of the same priority). For example, based on the HARQ-ACK and / or HARQ-NACK rate, the WTRU may select the set of quantizers.

[0220] The WTRU may trigger and / or (re)select a set of quantizers and / or perform the selection of the set of quantizers based on an associated transmission and / or a transmission type. For example, parameters of the associated transmission may include one or more of a logical channel, a data radio bearer (DRB), a signaling radio bearer (SRB), and / or a function of the transmission (e.g., measurement reporting, and / or radio access (RA)).

[0221] The WTRU may trigger and / or (re)select a set of quantizers and / or perform the selection of the set of quantizers based on the AI / ML encoder used. For example, the WTRU may select a set of quantizers based on the index, type, and / or parameters of the AI / ML encoder used.

[0222] The WTRU may trigger and / or (re)select a quantizer set and / or perform quantizer set selection based on receiving an indication from the gNB. For example, the WTRU may use a quantizer set. The WTRU may receive a configuration via a quantizer set index. In an example, the WTRU may receive an indication to trigger quantizer set (re)selection. The WTRU may receive the indication via RRC and / or DCI and / or MAC CE transmission.

[0223] The WTRU may trigger and / or (re)select a set of quantizers and / or perform quantizer set selection based on the timing of the transmission. For example, the WTRU may trigger and / or select a set of quantizers based on the timing of the transmission of a feedback report. The WTRU may select a set of quantizers at periodic time instances and / or a set of quantizers for every n feedback report instances.

[0224] The WTRU may trigger and / or (re)select a quantizer set and / or perform quantizer set selection based on a timer. The WTRU may (re)start a timer when the WTRU selects a quantizer set, a new set of quantizers, and / or trigger and / or select a new set of quantizers when the timer elapses. The timer may be modeled as several time instances, slots, subframes, and / or symbols. The WTRU may count the time instances, slots, subframes, and / or symbols and / or select a new set of quantizers when the counter reaches a configurable value. The WTRU may restart the counter when the WTRU selects a new set of quantizers. In an example, the rules for a WTRU to select a quantizer set may also apply to the rules for a WTRU that is triggered to select a quantizer set.

[0225] The quantizer fallback set may be implemented as described herein. In an example, a WTRU's selection of a quantizer set may fail. For example, the WTRU may be configured with a set of quantizers. However, that set of quantizers may not meet another criterion (e.g., a performance criterion). The WTRU may select and / or use a default and / or fallback set of quantizers if the selection fails. The quantizer fallback and / or default set may be configured. The WTRU may determine the quantizer fallback and / or default set. The WTRU may indicate the quantizer fallback and / or default set to the gNB to align the quantizers at the WTRU and / or the gNB. For example, the WTRU may override the gNB's indication of the quantizer set and / or the WTRU may report use of the quantizer fallback and / or default set, e.g., in a transmission using the quantizer fallback and / or default set.

[0226] The WTRU may send feedback of the selected quantizer to the network (e.g., in a feedback report) to align quantizers at the WTRU and / or the gNB. For example, the feedback of the selected quantizer may include a WTRU report of an identification of the set of quantizers. The WTRU may report the identification of the set of quantizers. The identification of the set of quantizers may include an index assigned to the set of quantizers, parameters of the set of quantizers (e.g., the number of quantizers in the set and / or the association between the quantizer and the AI / ML output), and / or parameters of the quantizers in the set.

[0227] The WTRU may report the quantizer type (e.g., uniform and / or non-uniform), quantization level, quantization scale, number of quantization levels, etc. The identification of the quantizer set may include an associated AI / ML model and / or capability to use the quantizer set. For example, the WTRU may report that the quantizer set should be used for CSI feedback reporting.

[0228] Different triggers may be implemented for reporting the identification information. The WTRU may trigger and / or report the identification information of the quantizer set by several different means. For example, the WTRU may use time as a trigger for reporting the identification information of the quantizer set. The WTRU may be configured with periodic instances when it may report the identification information of the currently used and / or active set of quantizers. The WTRU may trigger and / or report the identification information of the quantizer set by use of the quantizer set. For example, the WTRU may report the identification information of the used set of quantizers whenever the WTRU performs a transmission using the quantizer set. The WTRU may include the identification information of the quantizer set used to generate the CSI feedback report in one or more (e.g., all) CSI feedback reports.

[0229] A WTRU may trigger and / or report the identity of the quantizer set upon a change or (re)selection of the quantizer set. For example, when a WTRU triggers and / or (re)selects a quantizer set, the WTRU may report the identity of the newly selected set of quantizers. The reports may be independent and / or multiplexed in a transmission (e.g., a first transmission) that uses the selected set of quantizers. The WTRU may report the identity of the quantizer set whenever a selection is performed and / or when (e.g., only when) the newly selected set of quantizers is different from a previously used and / or indicated set of quantizers. The WTRU may trigger and / or report the identity of the quantizer set upon receipt of a reporting trigger. For example, the WTRU may receive a trigger to report the currently used set of quantizers. The trigger may be received in a DCI and / or a MAC CE.

[0230] The WTRU may report feedback of the selected quantizer using a reporting resource. The WTRU may report the identity of the quantizer set using one or more reporting resources. For example, the WTRU may report the identity of the quantizer set using a PUCCH (e.g., a PUCCH resource allocated for such reporting). In an example, the WTRU may multiplex the identity of the quantizer set in the PUCCH resource used to transmit UCI, SR, and / or HARQ-ACK. The WTRU may report the identity of the quantizer set using a PSCCH. The WTRU may report the identity of the quantizer set using a PUSCH (e.g., in UCI on the PUSCH), MAC CE, and / or RRC.

[0231] Receipt of the new set of quantizers may be acknowledged as described herein. When the WTRU is indicated a set of quantizers (e.g., by the gNB), the WTRU may acknowledge receipt of the indication and / or change to the indicated set of quantizers to align the quantizers at the WTRU and / or the gNB. The acknowledgement may be sent via means similar (e.g., the same) as those described herein for reporting the identity of the quantizer set.

[0232] The WTRU may report the performance of a set of quantizers as described herein. The WTRU may report the performance of one or more sets of quantizers. The WTRU may trigger and / or report the performance with the same trigger and / or reporting resources as described herein for reporting the identity of the set of quantizers.

[0233] A WTRU may receive an indication of a set of quantizers used by another node as described herein. The WTRU may receive an indication from another node (e.g., from a gNB for a DL transmission and / or from another WTRU for a SL transmission) that the other node used a set of quantizers for transmission to the WTRU. Based on the indicated set of quantizers used by the other node, the WTRU may perform several tasks. For example, based on the indicated set of quantizers used by the other node, the WTRU may select a set of inverse quantizers associated with the indicated set of quantizers. Based on the indicated set of quantizers used by the other node, the WTRU may train and / or retrain a set of inverse quantizers and / or an AI / ML model, send an acknowledgment that the WTRU received the indication of the set of quantizers, and / or send an indication of the identity of the selected set of inverse quantizers.

[0234] The identity of the set of inverse quantizers may be defined similarly to the identity of the set of quantizers. Based on the indicated set of quantizers used by the other node, the WTRU may select a new set of quantizers (e.g., for transmission from the WTRU to the other node) associated with the indicated set of quantizers. Based on the indicated set of quantizers used by the other node, the WTRU may report performance metrics. The performance metrics may be related to reception of transmissions using the new set of quantizers, and / or related to one or more sets of inverse quantizers, and / or related to the AI / ML model.

[0235] A WTRU may request that another node change its set of quantizers. A WTRU may report a request that another node change another node's selected set of quantizers. The request may indicate (e.g., include) a preferred set of quantizers and / or a cause for the request. The cause of the request may include one or more of a performance metric satisfying a condition, a change of an AI / ML model at the WTRU, and / or (re)training of an inverse quantizer and / or an AI / ML model at the WTRU. A request that another node change another node's selected set of quantizers may be transmitted via means similar to an indication of the identity of the selected quantizer set, as described herein. A request that another node change another node's selected set of quantizers may be transmitted via a HARQ-ACK report. For example, if the HARQ-ACK rate is below a threshold, the WTRU may include a request to change the selected set of quantizers.

[0236] CSI quantization may be implemented using non-matched quantizers. One or more types of quantizers may be implemented, and the quantizer types and / or methods for constructing the quantizers may include a uniform quantizer, a non-uniform quantizer, a distribution-based quantizer, and / or a clustering-based quantizer. One or more quantization parameters may be used to determine the properties of the quantizer. The properties of the quantizer may include quantization granularity, quantization overhead, quantization error, quantization range, quantization dynamic range, quantization complexity, quantization amplitude granularity, quantization phase granularity, and / or quantization precision, etc.

[0237] The quantization parameters may include one or more of the number of bits, the amplitude range, the phase range, the number of bits for the amplitude, the number of bits for the phase, and / or the quantization gap.

[0238] A quantizer may be characterized based on the quantization type and / or configuration, determination, and / or one or more quantization parameters used. Quantizer and inverse quantizer may be used interchangeably. For example, a quantizer at a transmitter may be, for example, a quantizer that converts values ​​(e.g., integer and / or complex values) into sets of bits. In an example, a quantizer at a receiver may be, for example, an inverse quantizer that converts sets of bits into values ​​(e.g., integer and / or complex values).

[0239] The quantizer in the WTRU may be used interchangeably with the quantizer in the encoder portion of the autoencoder (AE) in the AI / ML, and / or the quantizer in the gNB may be used interchangeably with the quantizer in the decoder portion of the AE in the AI / ML.

[0240] In a two-sided model (e.g., AE), the WTRU side model may be the encoder portion of the AE, and / or the gNB side model may be the decoder portion of the AE.

[0241] Independent quantizer determination may be performed in the WTRU and / or the gNB. A first quantizer may be determined for the WTRU side model, and / or a second quantizer may be determined for the gNB side model, where the first quantizer and / or the second quantizer may be determined based on several factors as described herein. In other words, the WTRU may determine a quantizer for the output of the encoder (e.g., the output of the WTRU side model), and the WTRU may determine a quantizer (e.g., quantizer type and associated parameters) based on several factors as described herein.

[0242] In an example, the WTRU may determine a quantizer for the output of the encoder based on the distribution of input and / or output data of the WTRU-side model. The data may be a data set used for training, testing, and / or inference. The WTRU may determine the quantizer for the output of the encoder based on uplink resources configured, determined, and / or indicated for reporting (e.g., reporting the output of the WTRU-side model). The uplink resources may include PUSCH, PUCCH, PRACH, and / or SRS resources. The WTRU may determine the quantizer for the output of the encoder based on the number of available bits for reporting the output data of the WTRU-side model. The number of available bits may be uncoded bits before the channel encoder and / or the total number of bits to be reported. The WTRU may determine the quantizer for the output of the encoder based on the WTRU type (e.g., WTRU category, etc.), the AI / ML model used (e.g., model ID), and / or the number of antenna ports associated with the CSI report (e.g., the number of CSI-RS ports).

[0243] The gNB may determine a quantizer for the decoder's input (e.g., input of the gNB-side model). The gNB may determine the quantizer (e.g., quantizer type and / or associated parameters) based on one or more factors, e.g., the gNB may determine the quantizer for the decoder's input based on a distribution of input and / or output data of the WTRU-side model, where the data may be a dataset used for training, testing, and / or inference. The gNB may determine the quantizer for the decoder's input based on an AI / ML model (e.g., model ID) used in the WTRU-side model (e.g., or both the WTRU-side model and / or the gNB-side model).

[0244] Quantizer set association may be implemented using a gNB-side model. For example, when a two-sided model is used, one or more quantizers may be used for the first side model (e.g., the WTRU) and / or the second side model (e.g., the gNB). A subset of quantizers in the first side model may be determined based on one or more factors, e.g., the subset of quantizers in the first side model may be determined based on an AI / ML model in the second side model. For example, the WTRU may determine the subset of quantizers based on information of the gNB-side model when a two-sided model is used. The information of the gNB-side model may include one or more of model identification information (e.g., model ID) and / or model functionality (e.g., CSI compression in the frequency domain, CSI compression in the time / frequency domain, and / or CSI compression in the time / frequency / spatial domain).

[0245] The subset of quantizers in the first side model may be determined based on the quantizer type and / or parameters used in the second side model. In an example, for example, if a uniform quantizer is used in the second side model, a quantizer with uniform quantization may be considered as a candidate quantizer for the first side model. If a non-uniform quantizer (e.g., distribution-based, clustering-based) is used in the second side model, a quantizer with a non-uniform quantizer type may be considered as a candidate quantizer in the first side model. The first side model may be in the WTRU and / or the gNB, and / or the second side model may be in the gNB and / or the WTRU.

[0246] The WTRU and / or gNB may determine a quantizer within the determined subset of quantizers to perform reporting of the output of the first side model (e.g., encoder) in the bilateral model. The WTRU and / or gNB may determine a quantizer within the determined subset of quantizers to perform determining a first side model and / or a second side model associated with the determined quantizer (e.g., when the quantizer is part of an AI / ML model). The WTRU and / or gNB may determine a quantizer within the determined subset of quantizers to perform reconstruction of compressed data using the second side model and / or fine-tuning of the AI / ML model using additional collected data sets.

[0247] Information regarding the determined quantizer may be signaled to the node (e.g., gNB and / or UE) where the other side of the AE resides, and the quantizer information may be one or more of the following. For example, the quantizer information may include a quantizer identification. One or more quantizers may be used, and each quantizer may be indexed with an identifier. For example, the quantizer information may include quantizer characteristics. The quantizer characteristics may be associated with a quantizer identification of a method for constructing the quantizer. The quantizer characteristics may be associated with a quantizer identification of a quantizer parameter. The quantizer characteristics may be associated with a quantizer identification of a quantizer overhead and / or a bitwidth.

[0248] In an example, a subset of quantizers in a first side model (e.g., a WTRU) may be determined based on quantizer characteristics indicated from a second side model (e.g., a gNB). For example, the gNB may indicate and / or configure a set of quantizer characteristics to the WTRU to be used in the WTRU side model. In an example, the WTRU may determine a quantizer that satisfies the set of quantizer characteristics indicated and / or configured by the gNB.

[0249] The quantizer characteristics may include a method for constructing the quantizer (e.g., uniform, distribution-based, and / or clustering-based). In examples, the quantizer characteristics may include quantizer parameters (e.g., range, amplitude granularity, phase granularity, and / or quantization bits, etc.), use cases (e.g., CSI compression, CSI prediction, and / or beam prediction, etc.), channel conditions (e.g., high Doppler and / or low Doppler, etc.), and / or deployment scenarios (e.g., Umi, Uma, and / or InH, etc.).

[0250] The quantizers determined at the WTRU side may be unknown to and / or reported to the gNB. The subset of quantizers for the first side model and / or the second side model may be implicitly determined based on several factors. For example, the subset of quantizers for the first side model and / or the second side model may be implicitly determined based on training conditions for the AI / ML model. The training conditions may include at least one of a training dataset size, a training latency, and a training type of both side models (e.g., Type 1: joint training, Type 2: joint training over the air interface, and / or Type 3: separate training). The subset of quantizers for the first side model and / or the second side model may be implicitly determined based on a use case (e.g., CSI compression, CSI prediction, and / or beam prediction, etc.), a channel condition (e.g., high Doppler and / or low Doppler, etc.), and / or a deployment scenario (e.g., Umi, Uma, and / or InH, etc.).

[0251] Example systems and / or methods for distribution-based adaptive CSI quantization are described herein. For example, a WTRU may receive a configuration for an adaptive (e.g., first-order) quantizer, e.g., to align quantizers at the WTRU and the gNB. The gNB may configure the WTRU to include a type of method (e.g., distribution-based and / or Gaussian) for constructing the quantizer. The gNB may configure the WTRU to include several quantizers, including multiple quantizers. The gNB may configure the WTRU to include a quantization performance metric including QN. The gNB may configure the WTRU to include a second-order quantizer. The second-order quantizer may be uniform. The gNB may configure the WTRU to include a feedback type, which may be a quantization parameter in the MAC CE.

[0252] The WTRU may use a second-order quantizer during the process(es) to determine parameters and / or construct an adaptive (e.g., first-order) quantizer. During the process(es), a second-order quantizer may be used to quantize the CSI compression.

[0253] The WTRU may determine parameters of an adaptive (e.g., first-order) quantizer and / or adapt the quantization block. To determine the quantizer parameters and / or adapt the quantization block, the WTRU may implement as described herein. For example, the WTRU may separately collect samples of encoder output values ​​to be used to construct one quantizer for each output. The WTRU may calculate the mean and / or variance of the samples for each encoder output. The WTRU may adapt the CDF transform of the non-uniform quantizer block for each output based on the calculated parameters of the distribution. The WTRU may then determine the resolution (e.g., or number of steps) of the uniform quantizer in the non-uniform quantizer block. For example, assuming the number of encoder outputs is 10 and / or the UCI size is 60 bits, a 64-step uniform quantizer may be constructed to obtain 6 bits of resolution.

[0254] The WTRU may feed back parameters of the distribution to the gNB. The parameters may include two scalar values ​​for the mean and / or variance of the distribution for each output (e.g., two 32-bit long values ​​for each of the 10 encoder outputs). The WTRU may semi-statically send feedback regarding the quantization parameters at the UL MAC CE. The WTRU may send feedback to the gNB for the gNB to build an inverse quantizer.

[0255] The WTRU and / or the gNB may start using a distribution-based first-order quantizer. In an example, the WTRU may receive an indication from the gNB to switch to the distribution-based quantizer. The WTRU may, for example, switch to the distribution-based first-order quantizer after the indication. Each output of the encoder may be quantized using a distribution-based quantizer.

[0256] The WTRU may monitor the performance of the distribution-based first-order quantizer in each compressed CSI. The WTRU may, for example, calculate the quantization noise of the distribution-based first-order quantizer and / or the second-order quantizer in each compressed CSI transmission. The WTRU may use the distribution-based quantizer, for example, if the QN of the distribution-based quantizer is lower than that of the second-order quantizer. Otherwise, the WTRU may use the second-order quantizer. The WTRU may feedback whether the distribution-based quantizer and / or the second-order quantizer are used with the compressed CSI in the UCI field (e.g., using a single bit).

[0257] The WTRU may report the performance of the distribution-based quantizer to the gNB. The WTRU may, for example, periodically (e.g., every N slots) report the quantizer performance metrics to the gNB. The WTRU may receive an indication to update the quantizer parameters if the quantizer performance falls below a certain threshold.

[0258] Described herein are noteworthy systems and / or methods for cluster-based adaptive CSI quantization. For example, a WTRU may receive a configuration for an adaptive (e.g., first-order) quantizer. A gNB may configure the WTRU to include a type of method for building the quantizer, including a cluster-based configuration, a number of quantizers, including a single quantizer, a quantization performance metric, including NMSE, and / or a second-order quantizer. The second-order quantizer may be uniform. The gNB may configure the WTRU to include a feedback type. The feedback type may include a quantization parameter in a MAC CE.

[0259] The WTRU may use the second-order quantizer during the process to determine parameters and / or construct an adaptive (e.g., first-order) quantizer, during which the second-order quantizer may be used to quantize the CSI compression.

[0260] The WTRU may calculate the centroids of the clusters and / or adapt the quantization block. The WTRU may collect combined samples of encoder output values. The WTRU may determine the number of clusters based on the UCI overhead and / or the number of encoder outputs. Assuming the UCI overhead is 60 bits and / or the number of encoder outputs is 10, the number of clusters may be 64, where each cluster is identified by 6 bits. The WTRU may partition the historical samples into 64 clusters and / or calculate the centroids using, for example, a k-means method. The WTRU may adapt the quantizer to the calculated centroids.

[0261] The WTRU can feed back the cluster centroids to the gNB. The centroids can be scalar values, for example, represented by 32 bits for each of 64 clusters. The WTRU can semi-statically send feedback regarding quantization parameters in the UL MAC CE. The WTRU can send feedback to the gNB for the gNB to build an inverse quantizer.

[0262] The WTRU and / or the gNB may start using a cluster-based primary quantizer. The WTRU may receive an indication from the gNB to switch to the cluster-based quantizer. The WTRU may switch to the cluster-based quantizer after the indication. Each output value of the encoder may be represented by a cluster index having a centroid closest to the encoder output value.

[0263] The WTRU may, for example, monitor the performance of the cluster-based quantizer in each compressed CSI. The WTRU may calculate the NMSE of the cluster-based quantizer and / or the second-order quantizer in each compressed CSI transmission. The WTRU may use the cluster-based quantizer if the NMSE of the cluster-based quantizer is lower than that of the second-order quantizer. Otherwise, the WTRU may use the second-order quantizer. The WTRU may feedback whether the cluster-based quantizer and / or the second-order quantizer are used with the compressed CSI in the UCI field (e.g., using a single bit).

[0264] The WTRU may report the performance of the cluster-based quantizer to the gNB. In an example, the WTRU may periodically (e.g., every N slots) report the quantizer performance metrics to the gNB. The WTRU may receive an indication to update the quantizer parameters if the quantizer performance falls below a certain threshold.

[0265] Although features and / or elements have been described above in particular combinations, those skilled in the art will appreciate that each feature or element may be used alone or in any combination with the other features and elements. Furthermore, the methods described herein may be implemented in a computer program, software, or firmware embodied in a computer-readable medium for execution by a computer or processor. Examples of computer-readable media include electronic signals (transmitted via wired or wireless connections) and computer-readable storage media. Examples of computer-readable storage media include, but are not limited to, read-only memory (ROM), random-access memory (RAM), registers, cache memory, semiconductor memory devices, magnetic media such as internal hard disks and removable disks, magneto-optical media, and optical media such as CD-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

1. Processor and memory 1. A wireless transmit / receive unit (WTRU) comprising: The processor: receiving configuration information from a network node for alignment of one or more compressed channel state information (CSI) quantizers at the WTRU and the network node, the configuration information indicating a plurality of CSI quantization settings and criteria for selecting from the plurality of CSI quantization settings; performing one or more CSI measurements on one or more reference signals; encoding the one or more CSI measurements; selecting a CSI quantization setting from the plurality of quantization settings based on the criteria indicated in the configuration information; quantizing the encoded one or more CSI measurements using the selected CSI quantization setting; sending a feedback report to the network node, the feedback report including the quantized one or more CSI measurements and an indication of the selected CSI quantization setting for the alignment of the one or more compressed CSI quantizers at the WTRU and the network node. A wireless transmit / receive unit (WTRU) configured to:

2. 2. The WTRU of claim 1, wherein the criteria is based on one or more of a reference signal received power (RSRP) measurement, a CSI measurement, a type of the feedback report, a resource used to send the feedback report, or a metric associated with the encoded one or more CSI measurements.

3. The WTRU of claim 1 , wherein the indication of the selected CSI quantization setting includes an index associated with the plurality of CSI quantization settings.

4. The WTRU of claim 1 , wherein the indication of the selected CSI quantization setting includes an indication of a quantization type, a quantization parameter, a quantizer granularity, or a quantizer resolution.

5. The WTRU of claim 4 , wherein the quantization type includes a single quantizer method, a multiple quantizer method, a distribution-based method, a cluster-based method, or uniform, non-uniform, or vector-based quantization.

6. The WTRU of claim 4 , wherein the quantization parameter comprises a centroid of a distribution or cluster of encoder outputs.

7. 10. The WTRU of claim 1, wherein the processor is configured with one or more fallback CSI quantizers, and wherein the processor is configured to select one or more compressed CSI fallback quantizers based on a second criterion.

8. 8. The WTRU of claim 7, wherein the second criteria includes a current state of the WTRU's current speed, RSSI, or the performance of the one or more compressed CSI quantizers, and the one or more fallback quantizers are measured via cosine similarity or quality noise.

9. The processor: The WTRU of claim 1 , configured to determine a quantizer parameter based on an index associated with a predefined lookup table.

10. The processor: The WTRU of claim 1 , configured to determine one or more parameters of a codebook for a predefined CSI quantizer.

11. 1. A method implemented by a wireless transmit / receive unit (WTRU), comprising: receiving configuration information from a network node for alignment of one or more compressed channel state information (CSI) quantizers at the WTRU and the network node, the configuration information indicating a plurality of CSI quantization settings and criteria for selecting from the plurality of CSI quantization settings; performing one or more CSI measurements on one or more reference signals; encoding the one or more CSI measurements; selecting a CSI quantization setting from the plurality of quantization settings based on the criteria indicated in the configuration information; quantizing the encoded one or more CSI measurements using the selected CSI quantization setting; sending a feedback report to the network node, the feedback report including the quantized one or more CSI measurements and an indication of the selected CSI quantization setting for the alignment of the one or more compressed CSI quantizers at the WTRU and the network node; A method for providing the above.

12. 12. The method of claim 11, wherein the criterion is based on one or more of a reference signal received power (RSRP) measurement, a CSI measurement, a type of the feedback report, a resource used to send the feedback report, or a metric associated with the encoded one or more CSI measurements.

13. The method of claim 11 , wherein the indication of the selected CSI quantization setting comprises an index associated with the plurality of CSI quantization settings.

14. The method of claim 11 , wherein the indication of the selected CSI quantization setting includes an indication of a quantization type, a quantization parameter, a quantizer granularity, or a quantizer resolution.

15. The method of claim 14 , wherein the quantization type includes a single quantizer method, a multiple quantizer method, a distribution-based method, a cluster-based method, or uniform, non-uniform, or vector-based quantization.

16. The method of claim 14 , wherein the quantization parameter comprises a centroid of a distribution or cluster of encoder outputs.

17. The method of claim 11 , further comprising one or more fallback quantizers, wherein the one or more fallback quantizers are selected based on a second criterion.

18. 18. The method of claim 17, wherein the second criteria includes a current state of the WTRU's current speed, RSSI, or performance of the one or more compressed CSI quantizers, and the one or more fallback quantizers are measured via cosine similarity or quality noise.

19. determining the quantizer parameters based on an index associated with a predefined lookup table; The method of claim 11 further comprising:

20. determining one or more parameters of a codebook for a predefined CSI quantizer; The method of claim 11 further comprising:

Images