Abstract: With the ever-growing demand for wireless-based applications, we face an enormous challenge of the lack of available radio spectrum. One promising direction for dealing with spectrum scarcity is devising ways to opportunistically reuse large chunks of the radio spectrum that are allocated but are severely underutilized. However, we must develop efficient, adaptive, and secure techniques/systems/protocols for making opportunistic spectrum sharing a reality. My research aims at achieving these goals using novel solutions that are assisted by machine learning and building practical systems using my solutions. Incorporating machine learning into my solutions is crucial for attaining intelligence and adaptability.
In this talk, first, I will talk about DeepRadar, which is a robust spectrum sensing system that enables spectrum sharing in the 3.5 GHz CBRS band. I will also briefly discuss some other works on intelligent spectrum sensing. Then, I will present an overview of my works on crowdsourced localization of spectrum offenders, which is essential for making spectrum, especially shared spectrum, secure. Next, I will briefly talk about a couple of ongoing works on physical layer authentication, which is a promising direction in IoT security. Finally, I will conclude with my future research plans toward agile but secure spectrum sharing among heterogeneous wireless technologies and next-generation communication systems.
Speaker Bio: Shamik Sarkar is a postdoctoral researcher at the University of California, Los Angeles, where he is advised by Dr. Danijela Cabric. Prior to that, Shamik obtained his Ph.D. in Computer Science from the University of Utah, where he was advised by Dr. Sneha Kumar Kasera. Shamik’s research interests lie in the intersection of communication systems, spectrum sharing, and machine learning.