Machine Learning and Signal Processing for Wideband Spectrum Knowledge Acquisition in Cognitive Radios
(duration: 3.5 hours)

Professor Sudharman K. Jayaweera
Associate Chair and Director of the Graduate Program
Department of Electrical and Computer Engineering
University of New Mexico, Albuquerque, NM
USA
E-mail: jayaweera@ece.unm.edu
 

Abstract: Future wireless communication systems will undoubtedly be based on Cognitive radios (CR) in some form or fashion. While software-defined radios (SDRs) can be intelligent radios, what sets cognitive radios apart from SDR is their ability to learn and be self-aware. Thus, a cognitive radio architecture must necessarily consist of modules that support these functionalities.
Self-awareness is achieved through spectrum knowledge acquisition, in part called the spectrum sensing. The role of signal processing and machine learning in spectrum knowledge acquisition and subsequent use of such acquired knowledge in decision-making and radio reconfiguration cannot be over-emphasized. In fact, one may argue that it is signal processing and machine learning that give rise to cognition and intelligence in a radio. These algorithms form the brain and brain functions of a cognitive radio while an SDR platform acts as the body of the radio. This tutorial is motivated by the timeliness of emphasizing this aspect of cognitive radios.
The objective of this half-day tutorial is to discuss cognitive radios from the perspectives of signal processing and machine learning. This emphasize will be used to highlight the potential of cognitive radio technology to go far beyond what has perhaps been considered so far in literature. In particular, the focus is on future autonomous cognitive radios aimed at broader applications rather than simply dynamic spectrum sharing (DSS).
First, a functional architecture of a wideband cognitive radio that emphasizes the role of signal processing and machine learning will be identified. The tutorial will then focus on particular signal processing problems that are unique to cognitive radios, in particular those that are associated with the problem of spectrum knowledge acquisition which can be divided in to three sub-problems: Wideband spectrum scanning, spectral activity detection and signal classification and identification. The tutorial will develop a complete suite of signal processing in detail covering all three sub-problems by combining classical statistical signal processing with novel machine learning algorithms.
Overall, the tutorial will discuss limitations of existing techniques to meet the challenges posed by cognitive radios and then lay out a research plan in signal processing and machine learning for realizing the full potential of cognitive radios to be autonomous, intelligent and self-aware radios.

Short biography: Born in Matara, Sri Lanka, Sudharman K. Jayaweera completed his high school education at the Rahula College, Matara, and was a science journalist at the Associated Newspapers Ceylon Limited (ANCL) till 1993. In 1997, he received the B.E. degree in Electrical and Electronic Engineering with First Class Honors from the University of Melbourne, Australia. He obtained his M.A. and PhD degrees in Electrical Engineering from Princeton University, USA in 2001 and 2003, respectively. A senior member of the IEEE, Dr. Jayaweera is currently an Associate Professor at the Department of Electrical and Computer Engineering at University of New Mexico, Albuquerque, NM where he is the Associate Chair and the Director of the Graduate Program. He has held fellowships at the Kirtland Air Force Research Laboratory Space Vehicles Directorate (summers of 2009-2011) and at the Naval Postgraduate School in Monterey, CA (part of 2013).
He was the keynote speaker at The 8th International conference on telecommunication system, services, and application (TSSA�14) held in Kuta Bali, Indonesia (Oct. 2014) and was the plenary speaker at IEEE Radio and Wireless Week held in Phoenix, AZ (January, 2011) and at the 6th IEEE International Conference in Industrial and Information Systems (ICIIS'2011) held in Kandy, Sri Lanka (Aug. 2011). His current research interests include cognitive and cooperative communications, machine learning, information theory of networked-control systems, statistical signal processing and control and optimization in smart-grid. His research has won 3 best paper awards at IEEE conferences. Dr. Jayaweera is the author of the Wiley book titled Signal Processing for Cognitive Radios published in 2014.
An editor of IEEE Transactions on Vehicular Technology, Dr. Jayaweera has served on organization and Technical Program Committees of numerous IEEE conferences. Most recently, he served as the Tutorial and Workshop Chair of the 2013 Fall IEEE Vehicular Technology Conference, Co-Chair of the Cognitive Radio and Networks Symposium at the IEEE Globecom 2015, Co-Chair of the Cognitive Radio Track at the 2015 Fall IEEE VTC, General Chair of the First Workshop on Wideband Mobile Cognitive Radios (WMCR) at the IEEE VTC Fall 2013 and the Publicity Chair of the 10th IEEE Broadband Wireless Access Workshop (BWA) at 2014 IEEE Globecom conference.