Colloquium | Nuria González-Prelcic, University of Vigo & UT Austin

Making millimeter-wave MIMO a reality: Algorithms and Applications

All dates for this event occur in the past.

fotonuria.jpg
Making millimeter-wave MIMO a reality: Algorithms and Applications 

Dr. Nuria González-Prelcic, Ph.D.

University of Vigo (Spain) and University of Texas at Austin

 

Friday, June 14

10:00 – 11:00am

Dreese Labs 260

 

Abstract

A major challenge in millimeter wave (mmWave) communications is to configure the hybrid antenna arrays in the transceivers. Furthermore, the mobility of the users in a mmWave network makes it necessary to periodically reconfigure these arrays, according to variations in the  channel. A popular strategy for array configuration/reconfiguration is based on beam training and beam refinement. This approach tries to find or refine the best beam pair. It has been adopted in the last WiFi and cellular standards operating at millimeter wave frequencies, because it enables joint synchronization and beam training at low signal-to-noise ratios. It also introduces, however, high overhead in the communication system, that disables the use of millimeter wave frequencies for some of the promised used cases for new generation communication systems providing high data rates and low latency, like cellular-supported robotic applications, automated driving, or UAV operations.

This talk reviews the work performed by my research group on designing an alternative strategy to configure the hybrid antenna arrays based on channel estimation and tracking of the high dimensional mmWave MIMO channel. I will start describing our estimation strategy for frequency selective millimeter wave channels for a simplified setting, where we assume the received signal is already synchronized. Then, I will define a more practical setting, where channel variations, lack of synchronization, calibration errors or beam squint are considered. I will continue describing how we further reduce the overhead associated to our estimation framework by using out-of-band information coming from sensors or other communication systems operating in parallel, or by exploiting statistical priors about some channel parameters. I will describe in more detail our current solution to the joint synchronization and channel estimation problem, and will compare its performance and overhead to the current solution in the 5G New Radio standard. Finally, I will conclude with a description of the new signal processing challenges that we are currently trying to address to enable millimeter wave communication in the most challenging mobile scenarios. 

Biography

Nuria González-Prelcic is an Associate Professor in the Signal Theory and Communications Department, University of Vigo, Spain (on leave) and a visiting professor at the Electrical Computing Engineering Department, The University of Texas at Austin, since August 2017. Her main research interests include signal processing theory and signal processing for wireless communications, with a recent emphasis on signal processing for millimeter-wave MIMO communications, including vehicle-to-everything and air-to-everything communication. Over the last 4 years she has published around 70 papers (22 journal papers and around 50 conference papers) on the topics of compressed sensing theory and its applications to millimeter wave and massive MIMO communications. She is an Editor for the IEEE Transactions on Wireless Communications and an Area Editor for IEEE Signal Processing Magazine. She is an elected member of the IEEE Sensor Array and Multichannel Signal Processing Technical Committee. She was the founding director of the Atlantic Research Center for Information and Communication Technologies (atlanTTic) at the University of Vigo from July 2008 to January 2017. She is currently the assistant director of UT SAVES, a research center that addresses the challenges of wireless, networking, and sensing in vehicular systems.