Radar approaches for sequential human activity classification

Abstract of the Topic: Human activity recognition with radar sensors has attracted a lot of attention, starting initially from fall detection and moving to the classification of more complex pattern of activities as well as hand gestures and vital signs. While initial research in this domain considered activities and human body movements as artificially separated, individual ‘snapshot-like’ data, more recent work is exploring techniques that can deal with more realistic, unconstrained sequences of continuous activities. This talk will provide a brief overview of recently proposed research and techniques in the context of radar-based human activity recognition, focussing on the radar signal processing tools and machine learning techniques proposed in state of the art literature.

Short Bio of the Speaker: Francesco Fioranelli received his Laurea (BEng, cum laude) and Laurea Specialistica (MEng, cum laude) degrees in telecommunication engineering from the Università Politecnica delle Marche, Ancona, Italy, in 2007 and 2010, respectively, and the Ph.D. degree in electronic engineering from Durham University, U.K., in 2014. He is currently Associate Professor at TU Delft, Microwave Sensing Signals & Systems Group, in the Netherlands, and was an Assistant Professor at the University of Glasgow (2016-2019) and Research Associate at University College London (2014-2016).

His research interests include the development of radar systems and automatic classification for human signatures analysis in healthcare and security, drones and UAVs detection and classification, and automotive radar. He has authored over 145 publications between book chapters, journal and conference papers, edited the books on “Micro-Doppler Radar and Its Applications” and “Radar Countermeasures for Unmanned Aerial Vehicles” published by IET-Scitech in 2020, and received three best paper awards.