Next-Generation Polymorphic Wireless Systems Through Real-Time Deep Spectrum Learning
The impressive scale and requirements of next-generation wireless systems will impose a never-before-seen burden on today's wireless infrastructure. Yet, existing wireless protocols are deeply rooted in inflexible designs, and thus unable to address the demands of the IoT and 5/6G networks. In this talk, I am going to explain our recent advances toward changing the state of the art and realizing polymorphic wireless systems. In our vision, wireless polymorphism will concretely realize a minimalistic, protocol-free, inference-based, on-the-fly approach to wireless communications where transmitters and receivers operate using a limited set of physical-layer parameters — seamlessly changed by the transmitter and inferred by the receiver by using real-time deep learning. I am also going to talk about our recent research on spectrum security, and in particular on the DARPA-funded radio-frequency machine learning systems (RFMLS) project.
Francesco Restuccia is an Associate Research Scientist with the Institute for the Wireless Internet of Things, Department of Electrical and Computer Engineering, Northeastern University, Boston. His research interests lie at the crossroads of embedded mobile systems, wireless networks and machine learning. He has published more than 25 papers in top-tier venues such as IEEE INFOCOM, ACM MobiHoc, and ACM SenSys. His research on AI-based wireless devices has been recognized with the 2019 ISSNAF Mario Gerla Award for Young Investigators in Computer Science. He has co-authored 7 pending US patents and 2 book chapters. Francesco serves on the TPC of several ACM and IEEE conferences, including IEEE INFOCOM, ACM MobiHoc, IEEE MASS, IEEE WoWMoM, and ACM MSWiM. Francesco is part of the only all-university team selected by DARPA to work on their new radio-frequency machine learning systems (RFMLS) program. He is a Member of the IEEE and ACM.
Host: Professor Ulkuhan Guler