Qualcomm Incorporated

Pavan Kumar Vitthaladevuni

Tingfang Ji

Krishna Kiran Mukkavilli

Naga Bhushan

Taesang Yoo

June Namgoong

Jay Kumar Sundararajan

Joseph Binamira Soriaga

Alexandros Manolakos

Danlu Zhang

Bo Chen

Ruifeng Ma

Yu Zhang

Jilei Hou

Sharad Deepak Sambhwani

Muhammad Sayed Khairy Abdelghaffar

Hwan Joon Kwon

Weimin Duan

Wanshi Chen

Jing Jiang

Ravi Agarwal

Junyi Li

Haitong Sun

Bilal Sadiq

Bibhu P Mohanty

Ozcan Ozturk

Tianyang Bai

Vasanthan Raghavan

Juergen Cezanne

John Edward Smee

Aziz Gholmieh

Hao Xu

Chenxi Hao

Arjun Bharadwaj

Yi Huang

Hung Dinh Ly

Muhammad Nazmul Islam

Chao Wei

Arnaud Meylan

Mehmet Yavuz

Yeliz Tokgoz

Josef J Blanz

Stefan Brueck

Runxin Wang

Xiaoxia Zhang

Sony Akkarakaran

Wei Zeng

Yongbin Wei

Seyed Ali Akbar Fakoorian

Peter Pui Lok Ang

Renqiu Wang

Vinay Chande

Liangming Wu

Qiaoyu Li

Amir Aminzadeh Gohari

Jong Hyeon Park

Yuwei Ren

Siddharth Mohan

Sanaz Barghi

Chih Ping Li

Yibo Jiang

Seyed Ali Ahmadzadeh

Weiliang Zeng

Lu Yuan

Sridhar Ramanujam

Sooryanarayanan Gopalakrishnan

Brandon Allen Burdge

Can Karakus

Krishna Kiran Mukkavili

Bibhu Prasad Mohanty

Methods, systems, and devices for wireless communications are described. A user equipment (UE) may transmit a request for sets of neural network weights, each set corresponding to a neural network, to one or more network entities. The one or more network entities may transmit multiple sets of neural network weights to the UE in response to the request. The UE may run a classifier neural network on the sets of neural network weights to select a set of neural network weights corresponding to a neural network. The UE may transmit a neural network identifier to the one or more network entities that transmitted neural network weights.

Online training and augmentation of neural networks for channel state feedback

Various aspects of the present disclosure relate to wireless communication. In some aspects, a client may receive an observed environmental vector feedback configuration associated with a reporting procedure for reporting updates corresponding to at least one observed environmental vector that is based at least in part on one or more features associated with an environment of the client, wherein the observed environmental vector comprises input for a first machine learning component that determines client specific parameters for use by a second machine learning component. The client may transmit an update corresponding to the at least one observed environmental vector based at least in part on the observed environmental vector feedback configuration. Numerous other aspects are provided.

Observed environmental vector feedback for wireless communication

Certain aspects of the present disclosure provide techniques for fast adaptation of transmission properties of sounding reference signal (SRS) resource sets. A method that may be performed by a user equipment (UE) includes receiving sounding reference signal (SRS) configuration information, configuring the UE with one or more SRS resource sets, wherein each of the one or more SRS resource sets comprises one or more SRS resources; receiving signaling comprising a field indicating which of the one or more SRS resources in the one or more SRS resource sets are active and which of the one or more SRS resources in the one or more SRS resources sets are inactive; and transmitting one or more SRSs using only the one or more SRS resources that are active.

Fast adaptation of transmission properties of SRS resource sets

Methods, systems, and devices for wireless communications are described. A user equipment (UE) may receive an indication of one or more encoding operations to use for encoding a compressed dataset, the one or more encoding operations including a differential encoding operation, or an entropy encoding operation, or both. In some examples, using a neural network, the UE may first encode a dataset based on an additional encoding operation to generate a compressed dataset and then quantize the compressed dataset encoded based on the additional encoding operation. Subsequently, after the dataset has been initially encoded and then quantized, the UE may use the indication of the one or more encoding operations to further encode and compress the dataset. The UE may then transmit the dataset to a second device based on the one or more encoding operations.

Encoding techniques for neural network architectures

Various aspects of the present disclosure generally relate to wireless communication. In some aspects, a user equipment (UE) may transmit a sounding reference signal (SRS) on a first set of sub-bands (). The UE may report channel state information (CSI) associated with a second set of sub-bands (). The first set of sub-bands may be different from the second set of sub-bands. The CSI may include information that identifies one or more frequency domain correlations among the first set of sub-bands and the second set of sub-bands. Numerous other aspects are described.

Frequency domain correlation feedback for enhanced frequency domain partial reciprocity downlink

Various aspects of the present disclosure generally relate to wireless communication. In some aspects, a user equipment (UE) may receive a configuration that indicates a sounding reference signal (SRS) resource set to be used for analog channel state feedback (CSF) and an association between the SRS resource set and a set of downlink channel state information reference signals (CSI-RSs). The UE may measure the set of downlink CSI-RSs associated with the SRS resource set. The UE may transmit one or more SRSs for analog CSF using the SRS resource set and based at least in part on measuring the set of downlink CSI-RSs. Numerous other aspects are provided.

Sounding reference signal resource set configuration for analog channel state feedback

Methods, systems, and devices for wireless communications are described. A first device and a second device may communicate via a channel. The first device may generate and transmit a reference signal, which may be a distortion probing reference signal with a high peak to average power ratio. In one implementation, the first device may use the reference signal as an input for a neural network model to learn a nonlinear response of the second device transmission components. In another implementation, the second device may sample the generated reference signal, and use the samples as inputs for a neural network model to learn the nonlinear response. The first device and the second device may exchange signaling based on learning the nonlinear response, and each device may compensate for the nonlinear response when communicating via the channel.

Distortion probing reference signals

Various aspects of the present disclosure generally relate to wireless communication. In some aspects, a client may select, based at least in part on a classifier, an autoencoder of a set of autoencoders to be used for encoding an observed wireless communication vector to generate a latent vector. The client may transmit the latent vector and an indication of the autoencoder. Numerous other aspects are provided.

Federated learning for classifiers and autoencoders for wireless communication

This disclosure provides methods, devices and systems for channel sounding for wireless communications. Some implementations more specifically relate to scheduling sounding reference signal (SRS) resource sets for wireless devices having more than 4 receive (RX) antenna ports. In some implementations, a base station may determine an antenna switching capability of a user equipment (UE). The antenna switching capability indicates a number of RX antenna ports of the UE. The base station schedules a number of SRS resource sets for the UE based at least in part on the number of RX antenna ports in excess of four. For example, the number of RX antenna ports may be equal to 8. As another example, the number of RX antenna ports may be equal to 6. The base station further receives, from the UE, uplink transmissions of one or more SRS resources for each of the scheduled SRS resource sets.

SRS antenna switching for multiple receive antennas

Methods, systems, and devices for wireless communications are described. A user equipment (UE) may receive a configuration identifying a first set of frequency resources for transmitting a sounding reference signal, where the sounding reference signal may be used at least for downlink channel state information acquisition. The UE may receive an indication of an association between the configuration for the sounding reference signal and a second set of frequency resources and determine, based on the association, values for one or more sounding reference signal transmission parameters for the first set of frequency resources. The UE may transmit the sounding reference signal on the first set of frequency resources based on the determined one or more values.

Reference signal resource association options for enhanced sounding

A user equipment (UE) receives a first channel state information reference signal (CSI-RS) on a first set of beams and receives a second CSI-RS on a second set of beams. The apparatus determines a channel quality indicator (CQI), a rank indicator (RI), and a precoding matrix indicator (PMI) based on the first CSI-RS. The apparatus reports the CQI and RI using an uplink control channel and indicates the PMI using a precoded SRS transmission. The precoded SRS transmission may be precoded based on a dissimilarity between the first CSI-RS and the second CSI-RS. The first CSI-RS may include a defined precoder, and the second CSI-RS may include a precoder based on an uplink channel estimate of a non-precoded SRS from the UE.

Analog CSF for FDD partial reciprocity

Various aspects of the present disclosure generally relate to wireless communication. In some aspects, a first device may encode a data set using one or more extraction operations and compression operations associated with a neural network, the one or more extraction operations and compression operations being based at least in part on a set of features of the data set to produce a compressed data set. The first device may transmit the compressed data set to a second device. Numerous other aspects are provided.

Encoding a data set using a neural network for uplink communication

In an aspect, a UE obtains information (e.g., UE-specific information) associated with a set of triggering criteria (e.g., from a server, a serving network, e.g., in conjunction with or separate from a set of neural network functions) for a set of neural network functions, the set of neural network functions configured to facilitate positioning measurement feature processing at the UE, the set of neural network functions being generated dynamically based on machine-learning associated with one or more historical measurement procedures. The UE obtains positioning measurement data associated with a location of the UE, and processes the positioning measurement data into a respective set of positioning measurement features based at least in part upon the positioning measurement data and at least one neural network function from the set of neural network functions that is triggered by at least one triggering criterion from the set of triggering criteria.

Selective triggering of neural network functions for positioning measurement feature processing at a user equipment

A method of wireless communication by a user equipment (UE) includes receiving, from a base station, a configuration to train a neural network for multiple different signal to noise ratios (SNRs) of a channel estimate for a wireless communication channel. The method also includes determining a current SNR of the channel estimate is above a first threshold value. The method further includes training the neural network based on the channel estimate, to obtain a first trained neural network. The method still further includes perturbing the channel estimate to obtain a perturbed channel estimate, and training the neural network based on the perturbed channel estimate, to obtain a second trained neural network. The method includes reporting, to the base station, parameters of the first trained neural network along with the channel estimate, and parameters of the second trained neural network.

Signaling for additional training of neural networks for multiple channel conditions

A base station may precode a first reference signal using a first precoder (P1) and at least one second reference signal using at least one second precoder (P2). At least two UEs may measure the first reference signal and the at least one second reference signal, and transmit channel state feedback based on the measuring (m11 and m12 for a first UE, and m21 and m22 for a second UE). The channel state feedback may comprise channel state information compressed by at least one neural network. The base station may derive a higher-level precoder (W) for multiple-input multiple-output (MIMO) downlink transmission to the at least two UEs.

Low overhead and high accuracy channel state feedback using neural networks

Growing community of inventors

San Diego, CA, United States of America

Pavan Kumar Vitthaladevuni