Title :
Feature Space Analysis of Modulation Classification Using Very High-Order Statistics
Author_Institution :
U.S. Army-CERDEC, Aberdeen Proving Ground, MD, USA
Abstract :
This paper explains how higher-order statistics work in classifying a unknown modulation scheme, why the very high-order statistics are needed in recognizing a large variety of signals, and what features should be selected in developing an effective classifier. The categories and properties of moments are discussed.
Keywords :
cognitive radio; higher order statistics; modulation; signal classification; cognitive radio; effective classifier; feature space analysis; high-order statistics; moment property; unknown modulation classification scheme; Color; Constellation diagram; Fading; Quadrature amplitude modulation; Shape; Transforms; Modulation classification; cognitive radio; cumulants; high-order statistics; moments;
Journal_Title :
Communications Letters, IEEE
DOI :
10.1109/LCOMM.2013.080613.130070
