2013 International Symposium on Wireless Communication Systems (ISWCS) Best Paper Award in Physical Layer Communications and Signal Processing for ‘‘Some Options for L1-Subspace Signal Processing.’’
Dimitris Pados, with his current Ph.D. student Panos Markopoulos and former Ph.D. student George Karystinos (Assoc. Prof. at Technical Univ. of Crete, Greece and 2012-13 UB visiting Prof.),received the 2013 Intern. Symposium on Wireless Communication Systems (ISWCS) Best Paper Award iin Physical Layer Communications and Signal Processing for their article ‘‘Some Options for L1-SubspaceSignal Processing.’’
In their groundbreaking work, Professor Pados and his students describe for the first time in the literature ways to define and calculate optimally L1-norm signal subspaces. In folk language, L1-norm distance is sometimes referred to as "Manhattan distance" or "Taxicab." L1-calculated signal subspaces are less sensitive to outlying (erroneous) data than L2-calculated subspaces. The work suggests the possibility now of establishing a new practical line of L1 data Principal-Component-Analysis (PCA) that parallels and robustifies the familiar L2 PCA approaches. Applications are projected to be as broad and as diverse as PCA itself, for example tweet analysis, astronomical data analysis, genomic signal processing, robust multidimensional statistical characterization (materials, processes), machine learning, feature extraction, image processing, spectrum sensing.