WO2022000365A1

WO2022000365A1 - Machine learning based downlink channel estimation and prediction

Info

Publication number: WO2022000365A1
Application number: PCT/CN2020/099710
Authority: WO
Inventors: Yuwei REN; Liangming WU; Tianyang BAI
Original assignee: Qualcomm Incorporated
Priority date: 2020-07-01
Filing date: 2020-07-01
Publication date: 2022-01-06

Abstract

A method of wireless communication by a user equipment (UE) includes receiving a machine learning model from a base station. The method also includes inputting channel measurements obtained from reference signals to the machine learning model. The method further includes inputting parameters to the machine learning model. Based on the parameters and the channel measurements input to the machine learning model, the machine learning model infers a neural network output. Based on the neural network output, the method estimates characteristics of a current downlink data channel and/or predicts characteristics of a future downlink data channel.

Description

MACHINE LEARNING BASED DOWNLINK CHANNEL ESTIMATION AND PREDICTION

FIELD OF THE DISCLOSURE

Aspects of the present disclosure generally relate to wireless communications, and more particularly to techniques and apparatuses for machine learning based downlink channel estimation and prediction.

BACKGROUND

Wireless communications systems are widely deployed to provide various telecommunications services such as telephony, video, data, messaging, and broadcasts. Typical wireless communications systems may employ multiple-access technologies capable of supporting communications with multiple users by sharing available system resources (e.g., bandwidth, transmit power, and/or the like) . Examples of such multiple-access technologies include code division multiple access (CDMA) systems, time division multiple access (TDMA) systems, frequency-division multiple access (FDMA) systems, orthogonal frequency-division multiple access (OFDMA) systems, single-carrier frequency-division multiple access (SC-FDMA) systems, time division synchronous code division multiple access (TD-SCDMA) systems, and long term evolution (LTE) . LTE/LTE-Advanced is a set of enhancements to the universal mobile telecommunications system (UMTS) mobile standard promulgated by the Third Generation Partnership Project (3GPP) .

A wireless communications network may include a number of base stations (BSs) that can support communications for a number of user equipment (UEs) . A user equipment (UE) may communicate with a base station (BS) via the downlink and uplink. The downlink (or forward link) refers to the communications link from the BS to the UE, and the uplink (or reverse link) refers to the communications link from the UE to the BS. As will be described in more detail, a BS may be referred to as a Node B, a gNB, an access point (AP) , a radio head, a transmit receive point (TRP) , a new radio (NR) BS, a 5G Node B, and/or the like.

The above multiple access technologies have been adopted in various telecommunications standards to provide a common protocol that enables different user equipment to communicate on a municipal, national, regional, and even global level. New Radio (NR) , which may also be referred to as 5G, is a set of enhancements to the LTE mobile standard promulgated by the Third Generation Partnership Project (3GPP) . NR is designed to better support mobile broadband Internet access by improving spectral efficiency, lowering costs, improving services, making use of new spectrum, and better integrating with other open standards using orthogonal frequency division multiplexing (OFDM) with a cyclic prefix (CP) (CP-OFDM) on the downlink (DL) , using CP-OFDM and/or SC-FDM (e.g., also known as discrete Fourier transform spread OFDM (DFT-s-OFDM) ) on the uplink (UL) , as well as supporting beamforming, multiple-input multiple-output (MIMO) antenna technology, and carrier aggregation.

Artificial neural networks may comprise interconnected groups of artificial neurons (e.g., neuron models) . The artificial neural network may be a computational device or represented as a method to be performed by a computational device. Convolutional neural networks, such as deep convolutional neural networks, are a type of feed-forward artificial neural network. Convolutional neural networks may include layers of neurons that may be configured in a tiled receptive field. It would be desirable to apply neural network processing to wireless communications to achieve greater efficiencies.

SUMMARY

According to aspects of the present disclosure, a method of wireless communication by a user equipment (UE) receives a machine learning model from a base station. The method also inputs channel measurements obtained from reference signals to the machine learning model. The method further includes inputting parameters to the machine learning model. Based on the parameters and the channel measurements input to the machine learning model, the machine learning model infers a neural network output. Based on the neural network output, the method estimates characteristics of a current downlink data channel and/or predicts characteristics of a future downlink data channel.

Another aspect of the present disclosure is directed to an apparatus including means for receiving, by a user equipment (UE) , a machine learning model from a base station. The apparatus also includes means for inputting channel measurements obtained from reference signals to the machine learning model. The apparatus further includes means for inputting parameters to the machine learning model. The apparatus also includes means for inferring a neural network output based on the parameters and the channel measurements input to the machine learning model. Finally, the apparatus includes means for estimating characteristics of a current downlink data channel and/or predicting characteristics of a future downlink data channel based on the neural network output of the machine learning model.

In another aspect, an apparatus for wireless communication at a user equipment (UE) includes a processor and a memory coupled to the processor. Instructions stored in the memory, when executed by the processor, cause the apparatus to receive a machine learning model from a base station. The instructions also cause the apparatus to input channel measurements obtained from reference signals to the machine learning model. The instructions further cause the apparatus to input parameters to the machine learning model. The instructions also cause the apparatus to infer a neural network output based on the parameters and the channel measurements input to the machine learning model. Finally, the instructions cause the apparatus to estimate characteristics of a current downlink data channel and/or predict characteristics of a future downlink data channel based on the neural network output of the machine learning model.

In yet another aspect, a non-transitory computer-readable medium records program code. The program code is executed by a user equipment (UE) and comprises program code to receive a machine learning model from a base station. The program code also includes program code to input channel measurements obtained from reference signals to the machine learning model. The program code further includes program code to input parameters to the machine learning model. The program code also includes program code to infer a neural network output based on the parameters and the channel measurements input to the machine learning model. The program code further includes program code to estimate characteristics of a current downlink data channel and/or predict characteristics of a future downlink data channel based on the neural network output of the machine learning model.

According to an aspect of the present disclosure, a method of wireless communication by a base station includes estimating characteristics of an uplink channel based on reference signal measurements received from a user equipment (UE) . The method also includes receiving, from the UE, downlink channel measurement feedback. The method also includes training a machine learning model for the UE based on the downlink channel measurement feedback and the estimated characteristics of the uplink channel. The method further includes updating network weights and input parameters based on the training. The method also includes transmitting the machine learning model, the network weights, and the input parameters to the UE.

Another aspect of the present disclosure is directed to an apparatus for wireless communication in a base station, including means for estimating characteristics of an uplink channel based on reference signal measurements received from a user equipment (UE) . The apparatus also includes means for receiving, from the UE, downlink channel measurement feedback. The apparatus further includes means for training a machine learning model for the UE based on the downlink channel measurement feedback and the estimated characteristics of the uplink channel. The apparatus also includes means for updating network weights and input parameters based on the training. The apparatus further includes means for transmitting the machine learning model, the network weights, and the input parameters to the UE.

In another aspect, an apparatus for wireless communication at a base station includes a processor and a memory coupled to the processor. Instructions stored in the memory, when executed by the processor, cause the apparatus to estimate characteristics of an uplink channel based on reference signal measurements received from a user equipment (UE) . The instructions also cause the apparatus to receive, from the UE, downlink channel measurement feedback. The instructions further cause the apparatus to train a machine learning model for the UE based on the downlink channel measurement feedback and the estimated characteristics of the uplink channel. The instructions also cause the apparatus to update network weights and input parameters based on the training. The instructions further cause the apparatus to transmit the machine learning model, the network weights, and the input parameters to the UE.

In yet another aspect, a non-transitory computer-readable medium records program code. The program code is executed by a base station and comprises program code to estimate characteristics of an uplink channel based on reference signal measurements received from a user equipment (UE) . The program code also includes program code to receive, from the UE, downlink channel measurement feedback. The program code further includes program code to train a machine learning model for the UE based on the downlink channel measurement feedback and the estimated characteristics of the uplink channel. The program code also includes program code to update network weights and input parameters based on the training. The program code further includes program code to transmit the machine learning model, the network weights, and the input parameters to the UE.

Aspects generally include a method, apparatus, system, computer program product, non-transitory computer-readable medium, user equipment, base station, wireless communication device, and processing system as substantially described with reference to and as illustrated by the accompanying drawings and specification.

The foregoing has outlined rather broadly the features and technical advantages of examples according to the disclosure in order that the detailed description that follows may be better understood. Additional features and advantages will be described. The conception and specific examples disclosed may be readily utilized as a basis for modifying or designing other structures for carrying out the same purposes of the present disclosure. Such equivalent constructions do not depart from the scope of the appended claims. Characteristics of the concepts disclosed, both their organization and method of operation, together with associated advantages will be better understood from the following description when considered in connection with the accompanying figures. Each of the figures is provided for the purposes of illustration and description, and not as a definition of the limits of the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

So that features of the present disclosure can be understood in detail, a particular description, may be had by reference to aspects, some of which are illustrated in the appended drawings. It is to be noted, however, that the appended drawings illustrate only certain aspects of this disclosure and are therefore not to be considered limiting of its scope, for the description may admit to other equally effective aspects. The same reference numbers in different drawings may identify the same or similar elements.

FIGURE 1 is a block diagram conceptually illustrating an example of a wireless communications network, in accordance with various aspects of the present disclosure.

FIGURE 2 is a block diagram conceptually illustrating an example of a base station in communication with a user equipment (UE) in a wireless communications network, in accordance with various aspects of the present disclosure.

FIGURE 3 illustrates an example implementation of designing a neural network using a system-on-a-chip (SOC) , including a general-purpose processor, in accordance with certain aspects of the present disclosure.

FIGURES 4A, 4B, and 4C are diagrams illustrating a neural network, in accordance with aspects of the present disclosure.

FIGURE 4D is a diagram illustrating an exemplary deep convolutional network (DCN) , in accordance with aspects of the present disclosure.

FIGURE 5 is a block diagram illustrating an exemplary deep convolutional network (DCN) , in accordance with aspects of the present disclosure.

FIGURE 6 is a timing diagram illustrating downlink channel measurement.

FIGURE 7 is a block diagram illustrating reference signal patterns.

FIGURE 8 is a block diagram illustrating estimation based on a reference signal pattern.

FIGURE 9 is a timing diagram illustrating machine learning based channel estimation and prediction, in accordance with aspects of the present disclosure.

FIGURE 10 is a block diagram illustrating machine learning based channel estimation and prediction, in accordance with aspects of the present disclosure.

FIGURE 11 is a block diagram illustrating an example implementation of a model for machine learning based channel estimation and prediction, in accordance with aspects of the present disclosure.

FIGURE 12 is a diagram illustrating an example process performed, for example, by a user equipment (UE) , in accordance with various aspects of the present disclosure.

FIGURE 13 is a diagram illustrating an example process performed, for example, by a base station, in accordance with various aspects of the present disclosure.

DETAILED DESCRIPTION

Various aspects of the disclosure are described more fully below with reference to the accompanying drawings. This disclosure may, however, be embodied in many different forms and should not be construed as limited to any specific structure or function presented throughout this disclosure. Rather, these aspects are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art. Based on the teachings herein one skilled in the art should appreciate that the scope of the disclosure is intended to cover any aspect of the disclosure, whether implemented independently of or combined with any other aspect of the disclosure. For example, an apparatus may be implemented or a method may be practiced using any number of the aspects set forth. In addition, the scope of the disclosure is intended to cover such an apparatus or method which is practiced using other structure, functionality, or structure and functionality in addition to or other than the various aspects of the disclosure set forth. It should be understood that any aspect of the disclosure disclosed may be embodied by one or more elements of a claim.

Several aspects of telecommunications systems will now be presented with reference to various apparatuses and techniques. These apparatuses and techniques will be described in the following detailed description and illustrated in the accompanying drawings by various blocks, modules, components, circuits, steps, processes, algorithms, and/or the like (collectively referred to as “elements” ) . These elements may be implemented using hardware, software, or combinations thereof. Whether such elements are implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system.

It should be noted that while aspects may be described using terminology commonly associated with 5G and later wireless technologies, aspects of the present disclosure can be applied in other generation-based communications systems, such as and including 3G and/or 4G technologies.

Machine learning has shown a capability to address complicated issues, such as nonlinear approximations. An example application in a real world environment is a channel between a base station and a UE modeled with a nonlinear relationship. Traditional channel estimates are based on designs with sometimes unrealistic assumptions, such as a linear channel. Thus, traditional channel estimates may have poor performance in certain situations. In some of these situations, machine learning methods may infer the embedded nonlinear feature, to obtain better channel estimation, while reducing resources.

In aspects of the present disclosure, a single neural network (NN) model, is defined on the UE side. The input to the neural network is the downlink (DL) channel measurement and/or uplink (UL) channel measurement. The output of the neural network is the estimated channel for real-time processing and/or a predicted channel for future applications. The model itself, as well as the neural network weights, are transmitted from the network to the UE. The model and weights may be specific to specific scenarios.

The network provides parameters to the UE. The parameters may be related to specific scenarios. The parameters may also indicate an effectiveness of the model with respect to time and frequency duration. There may also be other optimization parameters.

The network derives the neural network model, network weights, and input parameters in a variety of ways. For example, based on CSI-RS feedback from the UE, the network obtains the downlink channel. Based on uplink signals, such as a sounding reference signal (SRS) and/or DMRS, the network also obtains the uplink channel. That is, the network may receive more channel information in comparison to channel information received at the UE. Thus, the network receives more information and can create a better model than a UE. Based on the channel information, the network extracts scenario information, (e.g., mobility scenario, bandwidth or other information) . The network may also fine tune the network in real-time to update the network weights.

FIGURE 1 is a diagram illustrating a network 100 in which aspects of the present disclosure may be practiced. The network 100 may be a 5G or NR network or some other wireless network, such as an LTE network. The wireless network 100 may include a number of BSs 110 (shown as BS 110a, BS 110b, BS 110c, and BS 110d) and other network entities. A BS is an entity that communicates with user equipment (UEs) and may also be referred to as a base station, a NR BS, a Node B, a gNB, a 5G node B (NB) , an access point, a transmit receive point (TRP) , and/or the like. Each BS may provide communications coverage for a particular geographic area. In 3GPP, the term “cell” can refer to a coverage area of a BS and/or a BS subsystem serving this coverage area, depending on the context in which the term is used.

A BS may provide communications coverage for a macro cell, a pico cell, a femto cell, and/or another type of cell. A macro cell may cover a relatively large geographic area (e.g., several kilometers in radius) and may allow unrestricted access by UEs with service subscription. A pico cell may cover a relatively small geographic area and may allow unrestricted access by UEs with service subscription. A femto cell may cover a relatively small geographic area (e.g., a home) and may allow restricted access by UEs having association with the femto cell (e.g., UEs in a closed subscriber group (CSG) ) . A BS for a macro cell may be referred to as a macro BS. A BS for a pico cell may be referred to as a pico BS. A BS for a femto cell may be referred to as a femto BS or a home BS. In the example shown in FIGURE 1, a BS 110a may be a macro BS for a macro cell 102a, a BS 110b may be a pico BS for a pico cell 102b, and a BS 110c may be a femto BS for a femto cell 102c. A BS may support one or multiple (e.g., three) cells. The terms “eNB” , “base station” , “NR BS” , “gNB” , “TRP” , “AP” , “node B” , “5G NB” , and “cell” may be used interchangeably herein.

In some aspects, a cell may not necessarily be stationary, and the geographic area of the cell may move according to the location of a mobile BS. In some aspects, the BSs may be interconnected to one another and/or to one or more other BSs or network nodes (not shown) in the wireless network 100 through various types of backhaul interfaces such as a direct physical connection, a virtual network, and/or the like using any suitable transport network.

The wireless network 100 may also include relay stations. A relay station is an entity that can receive a transmission of data from an upstream station (e.g., a BS or a UE) and send a transmission of the data to a downstream station (e.g., a UE or a BS) . A relay station may also be a UE that can relay transmissions for other UEs. In the example shown in FIGURE 1, a relay station 110d may communicate with macro BS 110a and a UE 120d in order to facilitate communications between the BS 110a and UE 120d. A relay station may also be referred to as a relay BS, a relay base station, a relay, and/or the like.

The wireless network 100 may be a heterogeneous network that includes BSs of different types, e.g., macro BSs, pico BSs, femto BSs, relay BSs, and/or the like. These different types of BSs may have different transmit power levels, different coverage areas, and different impact on interference in the wireless network 100. For example, macro BSs may have a high transmit power level (e.g., 5 to 40 Watts) whereas pico BSs, femto BSs, and relay BSs may have lower transmit power levels (e.g., 0.1 to 2 Watts) .

A network controller 130 may couple to a set of BSs and may provide coordination and control for these BSs. The network controller 130 may communicate with the BSs via a backhaul. The BSs may also communicate with one another, e.g., directly or indirectly via a wireless or wireline backhaul.

UEs 120 (e.g., 120a, 120b, 120c) may be dispersed throughout the wireless network 100, and each UE may be stationary or mobile. A UE may also be referred to as an access terminal, a terminal, a mobile station, a subscriber unit, a station, and/or the like. A UE may be a cellular phone (e.g., a smart phone) , a personal digital assistant (PDA) , a wireless modem, a wireless communications device, a handheld device, a laptop computer, a cordless phone, a wireless local loop (WLL) station, a tablet, a camera, a gaming device, a netbook, a smartbook, an ultrabook, a medical device or equipment, biometric sensors/devices, wearable devices (smart watches, smart clothing, smart glasses, smart wrist bands, smart jewelry (e.g., smart ring, smart bracelet) ) , an entertainment device (e.g., a music or video device, or a satellite radio) , a vehicular component or sensor, smart meters/sensors, industrial manufacturing equipment, a global positioning system device, or any other suitable device that is configured to communicate via a wireless or wired medium.

Some UEs may be considered machine-type communications (MTC) or evolved or enhanced machine-type communications (eMTC) UEs. MTC and eMTC UEs include, for example, robots, drones, remote devices, sensors, meters, monitors, location tags, and/or the like, that may communicate with a base station, another device (e.g., remote device) , or some other entity. A wireless node may provide, for example, connectivity for or to a network (e.g., a wide area network such as Internet or a cellular network) via a wired or wireless communications link. Some UEs may be considered Internet-of-Things (IoT) devices, and/or may be implemented as NB-IoT (narrowband internet of things) devices. Some UEs may be considered a customer premises equipment (CPE) . UE 120 may be included inside a housing that houses components of UE 120, such as processor components, memory components, and/or the like.

In general, any number of wireless networks may be deployed in a given geographic area. Each wireless network may support a particular RAT and may operate on one or more frequencies. A RAT may also be referred to as a radio technology, an air interface, and/or the like. A frequency may also be referred to as a carrier, a frequency channel, and/or the like. Each frequency may support a single RAT in a given geographic area in order to avoid interference between wireless networks of different RATs. In some cases, NR or 5G RAT networks may be deployed.

In some aspects, two or more UEs 120 (e.g., shown as UE 120a and UE 120e) may communicate directly using one or more sidelink channels (e.g., without using a base station 110 as an intermediary to communicate with one another) . For example, the UEs 120 may communicate using peer-to-peer (P2P) communications, device-to-device (D2D) communications, a vehicle-to-everything (V2X) protocol (e.g., which may include a vehicle-to-vehicle (V2V) protocol, a vehicle-to-infrastructure (V2I) protocol, and/or the like) , a mesh network, and/or the like. In this case, the UE 120 may perform scheduling operations, resource selection operations, and/or other operations described elsewhere herein as being performed by the base station 110. For example, the base station 110 may configure a UE 120 via downlink control information (DCI) , radio resource control (RRC) signaling, a media access control-control element (MAC-CE) or via system information (e.g., a system information block (SIB) ) .

As indicated above, FIGURE 1 is provided merely as an example. Other examples may differ from what is described with regard to FIGURE 1.

FIGURE 2 shows a block diagram of a design 200 of the base station 110 and UE 120, which may be one of the base stations and one of the UEs in FIGURE 1. The base station 110 may be equipped with T antennas 234a through 234t, and UE 120 may be equipped with R antennas 252a through 252r, where in general T ≥ 1 and R ≥ 1.

At the base station 110, a transmit processor 220 may receive data from a data source 212 for one or more UEs, select one or more modulation and coding schemes (MCS) for each UE based at least in part on channel quality indicators (CQIs) received from the UE, process (e.g., encode and modulate) the data for each UE based at least in part on the MCS (s) selected for the UE, and provide data symbols for all UEs. Decreasing the MCS lowers throughput but increases reliability of the transmission. The transmit processor 220 may also process system information (e.g., for semi-static resource partitioning information (SRPI) and/or the like) and control information (e.g., CQI requests, grants, upper layer signaling, and/or the like) and provide overhead symbols and control symbols. The transmit processor 220 may also generate reference symbols for reference signals (e.g., the cell-specific reference signal (CRS) ) and synchronization signals (e.g., the primary synchronization signal (PSS) and secondary synchronization signal (SSS) ) . A transmit (TX) multiple-input multiple-output (MIMO) processor 230 may perform spatial processing (e.g., precoding) on the data symbols, the control symbols, the overhead symbols, and/or the reference symbols, if applicable, and may provide T output symbol streams to T modulators (MODs) 232a through 232t. Each modulator 232 may process a respective output symbol stream (e.g., for OFDM and/or the like) to obtain an output sample stream. Each modulator 232 may further process (e.g., convert to analog, amplify, filter, and upconvert) the output sample stream to obtain a downlink signal. T downlink signals from modulators 232a through 232t may be transmitted via T antennas 234a through 234t, respectively. According to various aspects described in more detail below, the synchronization signals can be generated with location encoding to convey additional information.

At the UE 120, antennas 252a through 252r may receive the downlink signals from the base station 110 and/or other base stations and may provide received signals to demodulators (DEMODs) 254a through 254r, respectively. Each demodulator 254 may condition (e.g., filter, amplify, downconvert, and digitize) a received signal to obtain input samples. Each demodulator 254 may further process the input samples (e.g., for OFDM and/or the like) to obtain received symbols. A MIMO detector 256 may obtain received symbols from all R demodulators 254a through 254r, perform MIMO detection on the received symbols if applicable, and provide detected symbols. A receive processor 258 may process (e.g., demodulate and decode) the detected symbols, provide decoded data for the UE 120 to a data sink 260, and provide decoded control information and system information to a controller/processor 280. A channel processor may determine reference signal received power (RSRP) , received signal strength indicator (RSSI) , reference signal received quality (RSRQ) , channel quality indicator (CQI) , and/or the like. In some aspects, one or more components of the UE 120 may be included in a housing.

On the uplink, at the UE 120, a transmit processor 264 may receive and process data from a data source 262 and control information (e.g., for reports comprising RSRP, RSSI, RSRQ, CQI, and/or the like) from the controller/processor 280. Transmit processor 264 may also generate reference symbols for one or more reference signals. The symbols from the transmit processor 264 may be precoded by a TX MIMO processor 266 if applicable, further processed by modulators 254a through 254r (e.g., for DFT-s-OFDM, CP-OFDM, and/or the like) , and transmitted to the base station 110. At the base station 110, the uplink signals from the UE 120 and other UEs may be received by the antennas 234, processed by the demodulators 254, detected by a MIMO detector 236 if applicable, and further processed by a receive processor 238 to obtain decoded data and control information sent by the UE 120. The receive processor 238 may provide the decoded data to a data sink 239 and the decoded control information to a controller/processor 240. The base station 110 may include communications unit 244 and communicate to the network controller 130 via the communications unit 244. The network controller 130 may include a communications unit 294, a controller/processor 290, and a memory 292.

The controller/processor 240 of the base station 110, the controller/processor 280 of the UE 120, and/or any other component (s) of FIGURE 2 may perform one or more techniques associated with machine learning for channel estimation, as described in more detail elsewhere. For example, the controller/processor 240 of the base station 110, the controller/processor 280 of the UE 120, and/or any other component (s) of FIGURE 2 may perform or direct operations of, for example, the processes of FIGURES 9, 12, and 13 and/or other processes as described.

Memories

242 and 282 may store data and program codes for the base station 110 and UE 120, respectively. A scheduler 246 may schedule UEs for data transmission on the downlink and/or uplink.

In some aspects, the UE 120 and/or base station 110 may include means for receiving, means for inputting, means for inferring, means for estimating, means for training, means for updating, means for transmitting, means for determining, means for triggering, means for requesting, and/or means for reporting. Such means may include one or more components of the UE 120 and/or base station 110 described in connection with FIGURE 2.

As indicated above, FIGURE 2 is provided merely as an example. Other examples may differ from what is described with regard to FIGURE 2.

In some cases, different types of devices supporting different types of applications and/or services may coexist in a cell. Examples of different types of devices include UE handsets, customer premises equipment (CPEs) , vehicles, Internet of Things (IoT) devices, and/or the like. Examples of different types of applications include ultra-reliable low-latency communications (URLLC) applications, massive machine-type communications (mMTC) applications, enhanced mobile broadband (eMBB) applications, vehicle-to-anything (V2X) applications, and/or the like. Furthermore, in some cases, a single device may support different applications or services simultaneously.

FIGURE 3 illustrates an example implementation of a system-on-a-chip (SOC) 300, which may include a central processing unit (CPU) 302 or a multi-core CPU configured for generating gradients for neural network training, in accordance with certain aspects of the present disclosure. The SOC 300 may be included in the base station 110 or UE 120. Variables (e.g., neural signals and synaptic weights) , system parameters associated with a computational device (e.g., neural network with weights) , delays, frequency bin information, and task information may be stored in a memory block associated with a neural processing unit (NPU) 308, in a memory block associated with a CPU 302, in a memory block associated with a graphics processing unit (GPU) 304, in a memory block associated with a digital signal processor (DSP) 306, in a memory block 318, or may be distributed across multiple blocks. Instructions executed at the CPU 302 may be loaded from a program memory associated with the CPU 302 or may be loaded from a memory block 318.

The SOC 300 may also include additional processing blocks tailored to specific functions, such as a GPU 304, a DSP 306, a connectivity block 310, which may include fifth generation (5G) connectivity, fourth generation long term evolution (4G LTE) connectivity, Wi-Fi connectivity, USB connectivity, Bluetooth connectivity, and the like, and a multimedia processor 312 that may, for example, detect and recognize gestures. In one implementation, the NPU is implemented in the CPU, DSP, and/or GPU. The SOC 300 may also include a sensor processor 314, image signal processors (ISPs) 316, and/or navigation module 320, which may include a global positioning system.

The SOC 300 may be based on an ARM instruction set. In an aspect of the present disclosure, the instructions loaded into the general-purpose processor 302 may comprise code to receive a machine learning model from a base station; code to input channel measurements obtained from reference signals to the machine learning model; code to input parameters to the machine learning model; code to infer a neural network output based on the parameters and channel measurements input to the machine learning model; and code to estimate a current downlink data channel and/or predicting a future downlink data channel based on the neural network output of the machine learning model. The instructions loaded into the general-purpose processor 302 may also comprise code to estimate an uplink channel based on measurements of reference signals received from a user equipment (UE) ; code to receive, from the UE, downlink channel measurement feedback; code to train a machine learning model for the UE based on the downlink channel measurement feedback and the uplink channel; code to update network weights and input parameters based on the training; and code to transmit the machine learning model, the network weights, and the input parameters to the UE.

Deep learning architectures may perform an object recognition task by learning to represent inputs at successively higher levels of abstraction in each layer, thereby building up a useful feature representation of the input data. In this way, deep learning addresses a major bottleneck of traditional machine learning. Prior to the advent of deep learning, a machine learning approach to an object recognition problem may have relied heavily on human engineered features, perhaps in combination with a shallow classifier. A shallow classifier may be a two-class linear classifier, for example, in which a weighted sum of the feature vector components may be compared with a threshold to predict to which class the input belongs. Human engineered features may be templates or kernels tailored to a specific problem domain by engineers with domain expertise. Deep learning architectures, in contrast, may learn to represent features that are similar to what a human engineer might design, but through training. Furthermore, a deep network may learn to represent and recognize new types of features that a human might not have considered.

A deep learning architecture may learn a hierarchy of features. If presented with visual data, for example, the first layer may learn to recognize relatively simple features, such as edges, in the input stream. In another example, if presented with auditory data, the first layer may learn to recognize spectral power in specific frequencies. The second layer, taking the output of the first layer as input, may learn to recognize combinations of features, such as simple shapes for visual data or combinations of sounds for auditory data. For instance, higher layers may learn to represent complex shapes in visual data or words in auditory data. Still higher layers may learn to recognize common visual objects or spoken phrases.

Deep learning architectures may perform especially well when applied to problems that have a natural hierarchical structure. For example, the classification of motorized vehicles may benefit from first learning to recognize wheels, windshields, and other features. These features may be combined at higher layers in different ways to recognize cars, trucks, and airplanes.

Neural networks may be designed with a variety of connectivity patterns. In feed-forward networks, information is passed from lower to higher layers, with each neuron in a given layer communicating to neurons in higher layers. A hierarchical representation may be built up in successive layers of a feed-forward network, as described above. Neural networks may also have recurrent or feedback (also called top-down) connections. In a recurrent connection, the output from a neuron in a given layer may be communicated to another neuron in the same layer. A recurrent architecture may be helpful in recognizing patterns that span more than one of the input data chunks that are delivered to the neural network in a sequence. A connection from a neuron in a given layer to a neuron in a lower layer is called a feedback (or top-down) connection. A network with many feedback connections may be helpful when the recognition of a high-level concept may aid in discriminating the particular low-level features of an input.

The connections between layers of a neural network may be fully connected or locally connected. FIGURE 4A illustrates an example of a fully connected neural network 402. In a fully connected neural network 402, a neuron in a first layer may communicate its output to every neuron in a second layer, so that each neuron in the second layer will receive input from every neuron in the first layer. FIGURE 4B illustrates an example of a locally connected neural network 404. In a locally connected neural network 404, a neuron in a first layer may be connected to a limited number of neurons in the second layer. More generally, a locally connected layer of the locally connected neural network 404 may be configured so that each neuron in a layer will have the same or a similar connectivity pattern, but with connections strengths that may have different values (e.g., 410, 412, 414, and 416) . The locally connected connectivity pattern may give rise to spatially distinct receptive fields in a higher layer, because the higher layer neurons in a given region may receive inputs that are tuned through training to the properties of a restricted portion of the total input to the network.

One example of a locally connected neural network is a convolutional neural network. FIGURE 4C illustrates an example of a convolutional neural network 406. The convolutional neural network 406 may be configured such that the connection strengths associated with the inputs for each neuron in the second layer are shared (e.g., 408) . Convolutional neural networks may be well suited to problems in which the spatial location of inputs is meaningful.

One type of convolutional neural network is a deep convolutional network (DCN) . FIGURE 4D illustrates a detailed example of a DCN 400 designed to recognize visual features from an image 426 input from an image capturing device 430, such as a car-mounted camera. The DCN 400 of the current example may be trained to identify traffic signs and a number provided on the traffic sign. Of course, the DCN 400 may be trained for other tasks, such as identifying lane markings or identifying traffic lights.

The DCN 400 may be trained with supervised learning. During training, the DCN 400 may be presented with an image, such as the image 426 of a speed limit sign, and a forward pass may then be computed to produce an output 422. The DCN 400 may include a feature extraction section and a classification section. Upon receiving the image 426, a convolutional layer 432 may apply convolutional kernels (not shown) to the image 426 to generate a first set of feature maps 418. As an example, the convolutional kernel for the convolutional layer 432 may be a 5x5 kernel that generates 28x28 feature maps. In the present example, because four different feature maps are generated in the first set of feature maps 418, four different convolutional kernels were applied to the image 426 at the convolutional layer 432. The convolutional kernels may also be referred to as filters or convolutional filters.

The first set of feature maps 418 may be subsampled by a max pooling layer (not shown) to generate a second set of feature maps 420. The max pooling layer reduces the size of the first set of feature maps 418. That is, a size of the second set of feature maps 420, such as 14x14, is less than the size of the first set of feature maps 418, such as 28x28. The reduced size provides similar information to a subsequent layer while reducing memory consumption. The second set of feature maps 420 may be further convolved via one or more subsequent convolutional layers (not shown) to generate one or more subsequent sets of feature maps (not shown) .

In the example of FIGURE 4D, the second set of feature maps 420 is convolved to generate a first feature vector 424. Furthermore, the first feature vector 424 is further convolved to generate a second feature vector 428. Each feature of the second feature vector 428 may include a number that corresponds to a possible feature of the image 426, such as “sign, ” “60, ” and “100. ” A softmax function (not shown) may convert the numbers in the second feature vector 428 to a probability. As such, an output 422 of the DCN 400 is a probability of the image 426 including one or more features.

In the present example, the probabilities in the output 422 for “sign” and “60” are higher than the probabilities of the others of the output 422, such as “30, ” “40, ” “50, ” “70, ” “80, ” “90, ” and “100” . Before training, the output 422 produced by the DCN 400 is likely to be incorrect. Thus, an error may be calculated between the output 422 and a target output. The target output is the ground truth of the image 426 (e.g., “sign” and “60” ) . The weights of the DCN 400 may then be adjusted so the output 422 of the DCN 400 is more closely aligned with the target output.

To adjust the weights, a learning algorithm may compute a gradient vector for the weights. The gradient may indicate an amount that an error would increase or decrease if the weight were adjusted. At the top layer, the gradient may correspond directly to the value of a weight connecting an activated neuron in the penultimate layer and a neuron in the output layer. In lower layers, the gradient may depend on the value of the weights and on the computed error gradients of the higher layers. The weights may then be adjusted to reduce the error. This manner of adjusting the weights may be referred to as “back propagation” as it involves a “backward pass” through the neural network.

In practice, the error gradient of weights may be calculated over a small number of examples, so that the calculated gradient approximates the true error gradient. This approximation method may be referred to as stochastic gradient descent. Stochastic gradient descent may be repeated until the achievable error rate of the entire system has stopped decreasing or until the error rate has reached a target level. After learning, the DCN may be presented with new images (e.g., the speed limit sign of the image 426) and a forward pass through the network may yield an output 422 that may be considered an inference or a prediction of the DCN.

Deep belief networks (DBNs) are probabilistic models comprising multiple layers of hidden nodes. DBNs may be used to extract a hierarchical representation of training data sets. A DBN may be obtained by stacking up layers of Restricted Boltzmann Machines (RBMs) . An RBM is a type of artificial neural network that can learn a probability distribution over a set of inputs. Because RBMs can learn a probability distribution in the absence of information about the class to which each input should be categorized, RBMs are often used in unsupervised learning. Using a hybrid unsupervised and supervised paradigm, the bottom RBMs of a DBN may be trained in an unsupervised manner and may serve as feature extractors, and the top RBM may be trained in a supervised manner (on a joint distribution of inputs from the previous layer and target classes) and may serve as a classifier.

Deep convolutional networks (DCNs) are networks of convolutional networks, configured with additional pooling and normalization layers. DCNs have achieved state-of-the-art performance on many tasks. DCNs can be trained using supervised learning in which both the input and output targets are known for many exemplars and are used to modify the weights of the network by use of gradient descent methods.

DCNs may be feed-forward networks. In addition, as described above, the connections from a neuron in a first layer of a DCN to a group of neurons in the next higher layer are shared across the neurons in the first layer. The feed-forward and shared connections of DCNs may be exploited for fast processing. The computational burden of a DCN may be much less, for example, than that of a similarly sized neural network that comprises recurrent or feedback connections.

The processing of each layer of a convolutional network may be considered a spatially invariant template or basis projection. If the input is first decomposed into multiple channels, such as the red, green, and blue channels of a color image, then the convolutional network trained on that input may be considered three-dimensional, with two spatial dimensions along the axes of the image and a third dimension capturing color information. The outputs of the convolutional connections may be considered to form a feature map in the subsequent layer, with each element of the feature map (e.g., 220) receiving input from a range of neurons in the previous layer (e.g., feature maps 218) and from each of the multiple channels. The values in the feature map may be further processed with a non-linearity, such as a rectification, max (0, x) . Values from adjacent neurons may be further pooled, which corresponds to down sampling, and may provide additional local invariance and dimensionality reduction. Normalization, which corresponds to whitening, may also be applied through lateral inhibition between neurons in the feature map.

The performance of deep learning architectures may increase as more labeled data points become available or as computational power increases. Modern deep neural networks are routinely trained with computing resources that are thousands of times greater than what was available to a typical researcher just fifteen years ago. New architectures and training paradigms may further boost the performance of deep learning. Rectified linear units may reduce a training issue known as vanishing gradients. New training techniques may reduce over-fitting and thus enable larger models to achieve better generalization. Encapsulation techniques may abstract data in a given receptive field and further boost overall performance.

FIGURE 5 is a block diagram illustrating a deep convolutional network 550. The deep convolutional network 550 may include multiple different types of layers based on connectivity and weight sharing. As shown in FIGURE 5, the deep convolutional network 550 includes the convolution blocks 554A, 554B. Each of the convolution blocks 554A, 554B may be configured with a convolution layer (CONV) 356, a normalization layer (LNorm) 558, and a max pooling layer (MAX POOL) 560.

The convolution layers 556 may include one or more convolutional filters, which may be applied to the input data to generate a feature map. Although only two of the convolution blocks 554A, 554B are shown, the present disclosure is not so limiting, and instead, any number of the convolution blocks 554A, 554B may be included in the deep convolutional network 550 according to design preference. The normalization layer 558 may normalize the output of the convolution filters. For example, the normalization layer 558 may provide whitening or lateral inhibition. The max pooling layer 560 may provide down sampling aggregation over space for local invariance and dimensionality reduction.

The parallel filter banks, for example, of a deep convolutional network may be loaded on a CPU 302 or GPU 304 of an SOC 300 to achieve high performance and low power consumption. In alternative embodiments, the parallel filter banks may be loaded on the DSP 306 or an ISP 316 of an SOC 300. In addition, the deep convolutional network 550 may access other processing blocks that may be present on the SOC 300, such as sensor processor 314 and navigation module 320, dedicated, respectively, to sensors and navigation.

The deep convolutional network 550 may also include one or more fully connected layers 562 (FC1 and FC2) . The deep convolutional network 550 may further include a logistic regression (LR) layer 564. Between each

layer

556, 558, 560, 562, 564 of the deep convolutional network 550 are weights (not shown) that are to be updated. The output of each of the layers (e.g., 556, 558, 560, 562, 564) may serve as an input of a succeeding one of the layers (e.g., 556, 558, 560, 562, 564) in the deep convolutional network 550 to learn hierarchical feature representations from input data 552 (e.g., images, audio, video, sensor data and/or other input data) supplied at the first of the convolution blocks 554A. The output of the deep convolutional network 550 is a classification score 566 for the input data 552. The classification score 566 may be a set of probabilities, where each probability is the probability of the input data, including a feature from a set of features.

As indicated above, FIGURES 3-5 are provided as examples. Other examples may differ from what is described with respect to FIGURES 3-5.

A wireless channel is highly complex. Channel estimation enables a receiver to understand the channel in order to recover transmitted information in a wireless communications system. Physical properties of the wireless channel may affect signals transmitted through it, resulting in attenuation, distortion, delays, and/or phase shift of the signals arriving at the receiver. Based on signals, such as known reference signals, a receiver can estimate properties or characteristics of a channel, which may be represented as a channel matrix H. By estimating an effect of the channel on the transmissions, the receive is able to recover the transmitted information.

In a new radio (NR) 5G system, there are two reference signals for downlink channel estimation. A demodulation reference signal (DMRS) is for channel estimation to demodulate a physical downlink control channel (PDCCH) and a physical downlink shared channel (PDSCH) . A channel state information reference signal (CSI-RS) is for channel estimation. The channel may be estimated for multiple purposes, including CSI (channel state information) acquisition, time and frequency tracking (e.g., to modify a time shift and frequency shift) , layer one reference signal received power (L1-RSRP) computation (e.g., for beam sweeping) , and mobility management.

FIGURE 6 is a timing diagram illustrating downlink channel measurement, in accordance with aspects of the present disclosure. For example, as shown in FIGURE 6, after the network configures resources for downlink reference signals, the network transmits the downlink (DL) reference signals (RSs) (e.g., CSI-RS and DMRS) at time t1. Upon receiving the reference signals at time t1, the UE measures the reference signals. Afterward, at times t2 and t3, respectively, the UE performs a DMRS based channel estimation and a CSI-RS based channel estimation. It is noted that channel estimations may occur in a different sequence from what is shown in FIGURE 6. Moreover, although time t1 represents the time when the reference signals are sent and received, it is noted that there is a propagation delay between the time when the reference signals are sent and received.

FIGURE 7 is a block diagram illustrating reference signal patterns, in accordance with aspects of the present disclosure. To match different scenarios, a network configures reference signal resources differently for channel estimation. In FIGURE 7, a demodulation reference signal (DMRS) is shown with different densities, that is, front-loaded DMRS patterns have different time-domain densities. A first pattern 1A has a baseline front-loaded DMRS with DMRS transmission on two symbols. A second pattern 1B has a front-loaded DMRS with twice the time-domain density as the first pattern 1A. A third pattern 1C has a front-loaded DMRS with four times the time-domain density as the first pattern 1A. Different densities have applications in different mobility environments. For example, with a high Doppler scenario, a DMRS density may be four times as high as a DMRS density configured for an environment with little movement or no movement at all (e.g., a static environment) .

The density of reference signals (RSs) may affect the accuracy of channel estimation and resource efficiency. For example, for a high speed scenario, the channel may vary quickly. Even with high density, the current estimated channel may be out-of-date for real-time applications. For a static environment, the channel may remain similar for a long duration. However, a fixed RS density may still consume significant resources. If the RS pattern is maintained in a limited channel variation environment, the user equipment (UE) may also suffer due to unnecessary computations. Without the configured RS, however, the UE may lose track of the channel.

As illustrated in FIGURE 7, discrete reference signal (RS) patterns may be predefined. Assuming channel correlation, the discrete patterns may represent the overall channel information. With discrete patterns, some of the remaining resources may be reserved for data transmission. However, with sparser reference signals, channel estimation may be less accurate.

To estimate a data channel, based on received reference signals, the UE attempts to extract the channel information in the reference signal resources. The extracted discrete channel information is then mapped to all of the resources, including the channels for data transmission. A non-machine learning or traditional method for obtaining a channel estimate for the data channel is interpolation, although other algorithms are also employed.

FIGURE 8 is a block diagram illustrating channel estimation based on a reference signal pattern, in accordance with aspects of the present disclosure. FIGURE 8 shows pattern 1B of FIGURE 7, with additional annotations for the data symbols and reference symbols. The channels of the data portion C1, C2 are interpolated based on the adjacent DMRS estimations H1, H2, H3. For example, the data channel C1 is interpolated based on the reference signals H1 and H2, whereas the data channel C2 is interpolated based on the reference signals H2 and H3. The interpolation may be used in the time and frequency domain. In one configuration the interpolation occurs as follows:

C1 = (H1 +H2) /2, C2 = (H2+H3) /2

Based on the discrete pattern, the UE performs the interpolation or applies other functions to map the entire set of resources, including the data channel. Still, the mapping may be limited. For example, channel variation in different portions can be nonlinear, whereas interpolation is a linear procedure to approximate a nonlinear channel variation. This linear procedure would inaccurately approximate the channel measurement of the overall resource.

Traditional methods may estimate the current channel and previous channel. Still, future channels should be predicted based on the current channel information. Traditional methods consider the current channel to approximate the future channel. However, the approximation may be inaccurate. For predicting a future channel, a traditional or non-machine learning solution may assume a current channel is a future channel, or the UE may predict a future channel based on a filter trend. Other traditional or non-machine learning techniques are also available.

Machine learning may be applied to solve complex problems, such as black box or nonlinear issues. According to aspects of the present disclosure, machine learning methods may extract an embedded nonlinear feature, which may be used to obtain a better channel estimation while consuming fewer resources. According to aspects of the present disclosure, a machine learning method is specified for channel estimation and also for channel prediction. The techniques of the present disclosure may estimate the current channel to map the overall resource. The techniques may also predict future channels. These techniques may be dynamic, improving accuracy, saving resources, and reducing unnecessary computations.

The network provides parameters to the UE. The parameters may be related to specific scenarios. For example, a high speed train scenario may be predefined. The predefined scenario may have a set of corresponding specific parameters, including available bandwidth, e.g., 20 MHz, effective time, e.g., 5 ms, beam index set, e.g., {04, 6, 7} , as well as other parameters. Once the high speed train scenario is indicated to the UE, the UE knows the corresponding parameter configuration. Thus, the UE does not receive the parameters one at a time. The parameters may also indicate an effectiveness of the model with respect to time and frequency duration. There may also be other optimization parameters.

The network derives the neural network model, network weights, and input parameters in a variety of ways. For example, based on CSI-RS feedback from the UE, the network obtains the downlink channel. Based on uplink signals, such as a sounding reference signal (SRS) and/or DMRS, the network also obtains the uplink channel. That is, the network may receive more channel information in comparison to channel information received at the UE. Thus, the network receives more information and can create a better model than a UE. Based on the channel information, the network extracts scenario information, (e.g., mobility scenario, bandwidth, or other information) . The network may also fine tune the network in real-time to update the network weights.

FIGURE 9 is a timing diagram illustrating machine learning based channel estimation and prediction, in accordance with aspects of the present disclosure. FIGURE 10 is a block diagram illustrating machine learning based channel estimation and prediction, in accordance with aspects of the present disclosure. As shown in FIGURE 9, at time t1 a network configures reference signal (RS) resources for channel estimation. For example, the network may configure a CSI-RS for downlink channel estimation and SRS for uplink channel estimation. The UE performs the DL channel estimation at time t2. At time t3, the UE transmits the uplink reference signal, which is also shown in FIGURE 10. At time t4, the UE sends measurement feedback to the network, which is also shown in FIGURE 10. The network receives the downlink channel and uplink channel at times t3 and t4. For a TDD structure, the UL channel is reciprocal with the downlink channel. That is, the downlink and uplink channels are correlated in the time domain. Although an exemplary sequence is shown and described with respect to FIGURE 9, such a sequence is merely exemplary. Other sequences are also contemplated.

The network trains and fine-tunes the machine learning model at time t5. As shown in FIGURE 10, the training and fine tuning are based on input, which may include the estimated downlink and uplink channel, bandwidth information, and scenario information. After training the network, at time t6, the network updates the model and network weights and transmits the weights to the UE. The UE receives the model and weights, and makes an inference at time t7, as also shown in FIGURE 10. That is, the UE estimates and/or predicts a channel.

FIGURE 11 is a block diagram illustrating an example implementation of a model for machine learning based channel estimation and prediction, in accordance with aspects of the present disclosure. As previously described, the UE and network share the same machine learning model and weights. The network side performs the training and validation, whereas the UE makes the inferences. In the example shown in FIGURE 11, the machine learning model is based on a sequence of long short-term memory (LSTM) cells. According to one aspect of the present disclosure, a length of the sequence (or buffer) is based on network or UE capability. In the example shown in FIGURE 11, the sequence has a length of five LSTM cells storing channel information C (t-4) , C (t-3) , X, C (t-1) , and C (t) , with cell C (t) corresponding to the most recent channel. The estimated or predicted channel corresponds to the cell C (t_1) . Each LSTM cell extracts channel features and estimates (or predicts) the channel. The most recent input is shown as C (t) . The other inputs C (t-1) , C (t-2) are previous features, buffered in the device. The input C (t) may be the estimated downlink channel fed back by the UE, the estimated UL channel measured in the gNodeB, bandwidth information of the input and the output, or a blank input for the estimated channel.

The cell designated by X corresponds to no channel information. Aspects of the present disclosure are directed to reducing the reference signal measurement cost, while ensuring the accuracy of the estimates and predictions. Thus, at some time stamps, the input may be blank. For example, when the UE is in a static environment (e.g., not moving) , the base station may not configure reference signals for channel estimation, saving the reference signal resources for other uses, such as data channels. Even when no reference signals are configured, the machine learning model should still estimate and predict channels, using a buffered channel state. In the example of FIGURE 11, the channel estimate (or predicted channel) not only considers the latest channel, but combines the previous buffered states while also accounting for a time period when no reference signal was measured.

According to aspects of the present disclosure, the machine learning model and weights are trained and optimized at the network side and used for the inference at the UE side. In an aspect of the present disclosure, the network may deliver the machine learning model to the UE. The content of the machine learning model includes the model structure and the corresponding weights. The model structure may be delivered to the UE in a variety of ways.

One option for delivery includes a predefined model structure. If updates to the model are desired, signaling may indicate an update is desired. For example, an LSTM structure may be predefined, and the LSTM cell length may be updated based on signaling. As described above, the model structure may be related to UE capability. For example, a UE with limited capability may support, at most, ten LSTM cells for inference. Thus, the UE may signal an update to the cell length based on the UE capability.

In a second delivery option, the model structure is directly communicated by signaling. The signaling may be radio resource control (RRC) signaling, a medium access control-control element (MAC-CE) or downlink control information (DCI) .

According to aspects of the present disclosure, the model weights may be periodically updated by signaling. In other aspects, the model weights are dynamically updated with signaling. For example, if low performance is observed by the UE and/or the base station, the base station may dynamically update the weights to match a new scenario. The signaling may be RRC signaling, a MAC-CE or DCI.

The network may also deliver other parameters to the UE. For example, the network may deliver scenario parameters (e.g., high Doppler, high speed, static movement, or other environments) . The network may also deliver configuration parameters, such as bandwidth information for the input or output, an effective time and/or frequency duration of the model, etc.

Bandwidth parameters are helpful because the UE estimates the reference signals to obtain the downlink channel information of these reference signals, and not the channel information across all resources. The estimated channel from the reference signals is transmitted to the network to obtain the output, which can be the channel information of all resources or only some resources in some bands. For example, the input channel information from the reference signals may be 20 MHz, and the output channel information may correspond to a 100 MHz resource. In this case, the bandwidth parameters (20 MHz, 100 MHz) may be input to the machine learning model.

The effective time duration and effective frequency duration parameters define a time and frequency duration where the machine learning model is expected to be effective. Outside of the duration, the machine learning model may not be accurate for channel estimation and prediction. For example, outside of the duration, the network and UE may switch back to a traditional, non-machine learning mode. In other aspects, the effective time duration and effective frequency parameters define a time and frequency, respectively, where the current weights or model structure is expected to be effective. Outside of the duration, the current weights or structure might not be accurate. In this case, the UE or network may update the weights or model structure.

In an aspect of the present disclosure, a network sends a message to control switching between a traditional mode and a machine learning mode. For example, the network may finish the machine learning model optimization, and update the model and weights to trigger the machine learning mode in the UE. The network may also transmit a message to instruct the UE switch to the machine learning mode.

The network may also switch the UE back to a traditional mode, instead of a machine learning mode. In this case, the network informs the UE that the machine learning model is not available. Alternatively, the network indicates the machine learning model is inaccurate, for example, when the network has optimized the machine learning model, and the model fails to track channel variations. The message may instruct the UE to switch back to the traditional mode. The message may also instruct the UE to switch back to the traditional mode and discard the machine learning model.

In a further aspect of the present disclosure, the UE reports failure of the machine learning model to the network. In this aspect, the UE uses the machine learning model to estimate and/or predict a channel, but with poor results. For example, the output of the model may not be accurate. That is, the channel estimate may not be used for downlink reception and demodulation. The results may be considered to be poor results when the output of the machine learning model is not better than the results from traditional methods. For example, conventional interpolation may be more accurate than machine learning model estimates.

The UE may report the failure message back to the network. In one aspect, the failure message may indicate failure of the machine learning model. In another aspect, the failure message indicates a difference between the machine learning model output and output from a traditional procedure. In response to receiving a failure report, the network may trigger channel measurement and feedback to enable the network to improve the machine learning model.

FIGURE 12 is a diagram illustrating an example process 1200 performed, for example, by a UE, in accordance with various aspects of the present disclosure. The example process 1200 is an example of machine learning based downlink channel estimation and prediction. As shown in FIGURE 12, in some aspects, the process 1200 may include receiving a machine learning model from a base station (block 1202) . For example, the user equipment (UE) (for example, using the antenna 252, DEMOD/MOD 254, MIMO detector 256, receive processor 258, controller processor 280, and/or memory 282) can receive a machine learning model. The process 1200 may include inputting channel measurements obtained from reference signals to the machine learning model (block 1204) . For example, the user equipment (UE) (for example, using the antenna 252, DEMOD/MOD 254, MIMO detector 256, receive processor 258, controller processor 280, and/or memory 282) can input channel measurements. As shown in FIGURE 12, in some aspects, the process 1200 may include inputting parameters to the machine learning model (block 1206) . For example, the user equipment (UE) (for example, using the antenna 252, DEMOD/MOD 254, MIMO detector 256, receive processor 258, controller processor 280, and/or memory 282) can input parameters.

As shown in FIGURE 12, in some aspects, the process 1200 may infer a neural network output based on the parameters and the channel measurements input to the machine learning model (block 1208) . For example, the user equipment (UE) (for example, using controller processor 280, and/or memory 282) can infer a neural network. The process 1200 may estimate characteristics of a current downlink data channel and/or predict characteristics of a future downlink data channel based on the neural network output of the machine learning model (block 1210) . For example, the user equipment (UE) (for example, using the controller processor 280, and/or memory 282) can estimate characteristics of a current downlink data channel.

FIGURE 13 is a diagram illustrating an example process 1300 performed, for example, by a base station, in accordance with various aspects of the present disclosure. The example process 1300 is an example of machine learning based downlink channel estimation and prediction. As shown in FIGURE 13, in some aspects, the process 1300 may include estimating characteristics of an uplink channel based on reference signal measurements received from a user equipment (UE) (block 1302) . For example, the base station (for example, using the antenna 234, MOD/DEMOD 232, TX MIMO 230, MIMO detector 236, transmit processor 220, receive processor 238, controller/processor 240, and/or memory 242) can estimate characteristics of an uplink channel. In some aspects, the process 1300 may include receiving, from the UE, downlink channel measurement feedback (block 1304) . For example, the base station (for example, using the antenna 234, MOD/DEMOD 232, MIMO detector 236, receive processor 238, controller/processor 240, and/or memory 242) can receive, from the UE, downlink channel measurement feedback.

As shown in FIGURE 13, in some aspects, the process 1300 may include training a machine learning model for the UE based on the downlink channel measurement feedback and the estimated characteristics of the uplink channel (block 1306) . For example, the base station (for example, using the controller/processor 240, and/or memory 242) can train a machine learning model. In some aspects, the process 1300 may include updating network weights and input parameters based on the training (block 1308) . For example, the base station (for example, using the controller/processor 240, and/or memory 242) can update network weights and input parameters. The process 1300 may include transmitting the machine learning model, the network weights, and the input parameters to the UE (block 1310) . For example, the base station (for example, using the antenna 234a, MOD/DEMOD 232a, TX MIMO 230, transmit processor 220, controller/processor 240, and/or memory 242) can transmit the machine learning model.

The foregoing disclosure provides illustration and description, but is not intended to be exhaustive or to limit the aspects to the precise form disclosed. Modifications and variations may be made in light of the above disclosure or may be acquired from practice of the aspects.

As used herein, the term “component” is intended to be broadly construed as hardware, firmware, and/or a combination of hardware and software. As used herein, a processor is implemented in hardware, firmware, and/or a combination of hardware and software.

Some aspects are described herein in connection with thresholds. As used herein, satisfying a threshold may, depending on the context, refer to a value being greater than the threshold, greater than or equal to the threshold, less than the threshold, less than or equal to the threshold, equal to the threshold, not equal to the threshold, and/or the like.

It will be apparent that systems and/or methods described herein may be implemented in different forms of hardware, firmware, and/or a combination of hardware and software. The actual specialized control hardware or software code used to implement these systems and/or methods is not limiting of the aspects. Thus, the operation and behavior of the systems and/or methods were described herein without reference to specific software code-it being understood that software and hardware can be designed to implement the systems and/or methods based, at least in part, on the description herein.

Even though particular combinations of features are recited in the claims and/or disclosed in the specification, these combinations are not intended to limit the disclosure of various aspects. In fact, many of these features may be combined in ways not specifically recited in the claims and/or disclosed in the specification. Although each dependent claim listed below may directly depend on only one claim, the disclosure of various aspects includes each dependent claim in combination with every other claim in the claim set. A phrase referring to “at least one of” a list of items refers to any combination of those items, including single members. As an example, “at least one of: a, b, or c” is intended to cover a, b, c, a-b, a-c, b-c, and a-b-c, as well as any combination with multiples of the same element (e.g., a-a, a-a-a, a-a-b, a-a-c, a-b-b, a-c-c, b-b, b-b-b, b-b-c, c-c, and c-c-c or any other ordering of a, b, and c) .

No element, act, or instruction used should be construed as critical or essential unless explicitly described as such. Also, as used, the articles “a” and “an” are intended to include one or more items, and may be used interchangeably with “one or more. ” Furthermore, as used, the terms “set” and “group” are intended to include one or more items (e.g., related items, unrelated items, a combination of related and unrelated items, and/or the like) , and may be used interchangeably with “one or more. ” Where only one item is intended, the phrase “only one” or similar language is used. Also, as used, the terms “has, ” “have, ” “having, ” and/or the like are intended to be open-ended terms. Further, the phrase “based on” is intended to mean “based, at least in part, on” unless explicitly stated otherwise.

Claims

A method of wireless communications, by a user equipment (UE) , comprising:

receiving a machine learning model from a base station;

inputting channel measurements obtained from reference signals to the machine learning model;

inputting parameters to the machine learning model;

inferring a neural network output based on the parameters and the channel measurements input to the machine learning model; and

estimating characteristics of a current downlink data channel and/or predicting characteristics of a future downlink data channel based on the neural network output of the machine learning model.
The method of claim 1, in which the parameters comprise bandwidth information corresponding to the channel measurements obtained from the reference signals, bandwidth information corresponding to the current downlink data channel or the future downlink data channel, and/or scenario parameters related to a mobility environment of the UE.
The method of claim 1, further comprising receiving, from the base station, the parameters, including bandwidth parameters, effective time parameters or scenario parameters for different mobility scenarios.
The method of claim 3, in which the scenario parameters are predefined, each scenario parameter corresponding to a specific set of parameters.
The method of claim 1, in which receiving the machine learning model comprises receiving different network weights and different network models for different scenarios.
The method of claim 1, further comprising receiving, from the base station, an effective time duration and/or frequency duration for employing:

the machine learning model; and/or

current network weights and a current network structure.
The method of claim 1, in which the machine learning model comprises a sequence of cells as input, the sequence of cells including: buffered input from a previous inference, and current input comprising the parameters and the channel measurements from the reference signals.
The method of claim 7, in which each cell of the sequence corresponds to a time stamp.
The method of claim 7, in which the buffered input includes a blank input of a previous time period.
The method of claim 7, in which a length of the sequence is based at least in part on a UE capability, or scenario parameters.
The method of claim 10, further comprising requesting the base station to change the length of the sequence.
The method of claim 1, further comprising receiving a message indicating whether to infer the neural network output or to operate in a non-machine learning mode.
The method of claim 12, in which the message indicates the machine learning model is not available or the machine learning model is performing below a threshold level.
The method of claim 1, further comprising reporting to the base station a failure of the machine learning model.
The method of claim 14, in which reporting the failure occurs in response to a performance level of the machine learning model falling below a threshold.
The method of claim 14, in which reporting the failure occurs in response to the neural network output being less accurate than a non-machine learning output.
The method of claim 16, in which reporting the failure includes indicating a difference between the neural network output and the non-machine learning output.
An apparatus for wireless communications at a user equipment (UE) , comprising:

means for receiving a machine learning model from a base station;

means for inputting channel measurements obtained from reference signals to the machine learning model;

means for inputting parameters to the machine learning model;

means for inferring a neural network output based on the parameters and the channel measurements input to the machine learning model; and

means for estimating characteristics of a current downlink data channel and/or predicting characteristics of a future downlink data channel based on the neural network output of the machine learning model.
The apparatus of claim 18, in which the parameters comprise bandwidth information corresponding to the channel measurements obtained from the reference signals, bandwidth information corresponding to the current downlink data channel or the future downlink data channel, and/or scenario parameters related to a mobility environment of the UE.
The apparatus of claim 18, further comprising means for receiving, from the base station, the parameters, including bandwidth parameters, effective time parameters or scenario parameters for different mobility scenarios.
The apparatus of claim 18, further comprising means for predefining scenario parameters, each scenario parameter corresponding to a specific set of parameters.
The apparatus of claim 18, further comprising means for receiving the machine learning model comprises receiving different network weights and different network models for different scenarios.
The apparatus of claim 18, further comprising means for receiving, from the base station, an effective time duration and/or frequency duration for employing:

the machine learning model; and/or

current network weights and a current network structure.
The apparatus of claim 18, in which the machine learning model further comprises a sequence of cells as input, the sequence of cells including: buffered input from a previous inference, and current input comprising the parameters and the channel measurements from the reference signals.
The apparatus of claim 24, in which each cell of the sequence corresponds to a time stamp.
The apparatus of claim 24, in which the buffered input includes a blank input of a previous time period.
The apparatus of claim 24, in which a length of the sequence is based at least in part on a UE capability, or scenario parameters.
The apparatus of claim 27, further comprising means for requesting the base station to change the length of the sequence.
The apparatus of claim 18, further comprising means for receiving a message indicating whether to infer the neural network output or to operate in a non-machine learning mode.
The apparatus of claim 29, in which the message indicates the machine learning model is not available or the machine learning model is performing below a threshold level.
The apparatus of claim 18, further comprising means for reporting to the base station a failure of the machine learning model.
The apparatus of claim 31, in which reporting the failure occurs in response to a performance level of the machine learning model falling below a threshold.
The apparatus of claim 31, in which reporting the failure occurs in response to the neural network output being less accurate than a non-machine learning output.
The apparatus of claim 33, in which reporting the failure occurs in response to a difference between the neural network output and the non-machine learning output.
An apparatus for wireless communications at a user equipment (UE) , comprising:

a processor;

a memory coupled with the processor; and

instructions stored in the memory and operable, when executed by the processor,

to cause the apparatus:

to receive a machine learning model from a base station;

to input channel measurements obtained from reference signals to the machine learning model;

to input parameters to the machine learning model;

to infer a neural network output based on the parameters and the channel measurements input to the machine learning model; and

to estimate characteristics of a current downlink data channel and/or predicting characteristics of a future downlink data channel based on the neural network output of the machine learning model.
The apparatus of claim 35, in which the parameters comprise bandwidth information corresponding to the channel measurements obtained from the reference signals, bandwidth information corresponding to the current downlink data channel or the future downlink data channel, and/or scenario parameters related to a mobility environment of the UE.
The apparatus of claim 35, in which the processor causes the apparatus to receive, from the base station, the parameters, including bandwidth parameters, effective time parameters or scenario parameters for different mobility scenarios.
The apparatus of claim 37, in which the scenario parameters are predefined, each scenario parameter corresponding to a specific set of parameters.
The apparatus of claim 35, in which the processor causes the apparatus to receive the machine learning model, which comprises receiving different network weights and different network models for different scenarios.
The apparatus of claim 35, in which the processor causes the apparatus to receive, from the base station, an effective time duration and/or frequency duration for employing:

the machine learning model; and/or

current network weights and a current network structure.
The apparatus of claim 35, in which the machine learning model comprises a sequence of cells as input, the sequence of cells including: buffered input from a previous inference, and current input comprising the parameters and the channel measurements from the reference signals.
The apparatus of claim 41, in which each cell of the sequence corresponds to a time stamp.
The apparatus of claim 41, in which the buffered input includes a blank input of a previous time period.
The apparatus of claim 41, in which a length of the sequence is based at least in part on a UE capability, or scenario parameters.
The apparatus of claim 44, in which the processor causes the apparatus to request the base station to change the length of the sequence.
The apparatus of claim 35, in which the processor causes the apparatus to receive a message indicating whether to infer the neural network output or to operate in a non-machine learning mode.
The apparatus of claim 46, in which the message indicates the machine learning model is not available or the machine learning model is performing below a threshold level.
The apparatus of claim 35, in which the processor causes the apparatus to report to the base station a failure of the machine learning model.
The apparatus of claim 48, in which reporting the failure occurs in response to a performance level of the machine learning model falling below a threshold.
The apparatus of claim 48, in which reporting the failure occurs in response to the neural network output being less accurate than a non-machine learning output.
The apparatus of claim 50, in which reporting the failure includes indicating a difference between the neural network output and the non-machine learning output.
A non-transitory computer-readable medium having program code recorded thereon, the program code executed by a user equipment (UE) and comprising:

program code to receive a machine learning model from a base station;

program code to input channel measurements obtained from reference signals to the machine learning model;

program code to input parameters to the machine learning model;

program code to infer a neural network output based on the parameters and the channel measurements input to the machine learning model; and

program code to estimate characteristics of a current downlink data channel and/or predicting characteristics of a future downlink data channel based on the neural network output of the machine learning model.
A method of wireless communications, by a base station, comprising:

estimating characteristics of an uplink channel based on reference signal measurements received from a user equipment (UE) ;

receiving, from the UE, downlink channel measurement feedback;

training a machine learning model for the UE based on the downlink channel measurement feedback and the estimated characteristics of the uplink channel;

updating network weights and input parameters based on the training; and

transmitting the machine learning model, the network weights, and the input parameters to the UE.
The method of claim 53, further comprising training the machine learning model based on scenario information and bandwidth information.
The method of claim 53, further comprising:

determining performance of the machine learning model is below a threshold;

updating the network weights in response to the performance falling below the threshold; and

transmitting the updated weights to the UE.
The method of claim 53, further comprising:

receiving from the UE a report of a failure of the machine learning model;

triggering uplink channel measurements, downlink channel measurements, and network feedback from the UE in response to receiving the report of the failure; and

updating the machine learning model, the network weights, and/or the input parameters based on the uplink channel measurements, downlink channel measurements, and the network feedback.
An apparatus for wireless communications at a base station, comprising:

means for estimating characteristics of an uplink channel based on reference signal measurements received from a user equipment (UE) ;

means for receiving, from the UE, downlink channel measurement feedback;

means for training a machine learning model for the UE based on the downlink channel measurement feedback and the estimated characteristics of the uplink channel;

means for updating network weights and input parameters based on the training; and

means for transmitting the machine learning model, the network weights, and the input parameters to the UE.
The apparatus of claim 57, further comprising means for training the machine learning model based on scenario information and bandwidth information.
The apparatus of claim 57, further comprising:

means for determining performance of the machine learning model is below a threshold;

means for updating the network weights in response to the performance falling below the threshold; and

means for transmitting the updated weights to the UE.
The apparatus of claim 57, further comprising:

means for receiving from the UE a report of a failure of the machine learning model;

means for triggering uplink channel measurements, downlink channel measurements, and network feedback from the UE in response to receiving the report of the failure; and

means for updating the machine learning model, the network weights, and/or the input parameters based on the uplink channel measurements, downlink channel measurements, and the network feedback.
An apparatus for wireless communications, by a base station, comprising:

a processor;

a memory coupled with the processor; and

instructions stored in the memory and operable, when executed by the processor,

to cause the apparatus:

to estimate characteristics of an uplink channel based on reference signal measurements received from a user equipment (UE) ;

to receive, from the UE, downlink channel measurement feedback;

to train a machine learning model for the UE based on the downlink channel measurement feedback and the estimated characteristics of the uplink channel;

to update network weights and input parameters based on the training; and

to transmit the machine learning model, the network weights, and the input parameters to the UE.
The apparatus of claim 61, in which the processor causes the apparatus to train the machine learning model based on scenario information and bandwidth information.
The apparatus of claim 61, in which the processor causes the apparatus:

to determine performance of the machine learning model is below a threshold;

to update the network weights in response to the performance falling below the threshold; and

to transmit the updated weights to the UE.
The apparatus of claim 61, in which the processor causes the apparatus:

to receive from the UE a report of a failure of the machine learning model;

to trigger uplink channel measurements, downlink channel measurements, and network feedback from the UE in response to receiving the report of the failure; and

to update the machine learning model, the network weights, and/or the input parameters based on the uplink channel measurements, downlink channel measurements, and the network feedback.
A non-transitory computer-readable medium having program code recorded thereon, the program code executed by a base station and comprising:

program code to estimate characteristics of an uplink channel based on reference signal measurements received from a user equipment (UE) ;

program code to receive, from the UE, downlink channel measurement feedback;

program code to train a machine learning model for the UE based on the downlink channel measurement feedback and the estimated characteristics of the uplink channel;

program code to update network weights and input parameters based on the training; and

program code to transmit the machine learning model, the network weights, and the input parameters to the UE.