Automatic classification of multiple signals using 2D matching of magnitude–frequency density features

摘要: Signal classification is an important function of modern communication systems in software defined radio applications. The ability to quickly recognize the type received signals allows a system automatically adapt processor properly decode signals. Many techniques assume that signal space occupied by only one signal, and frequency operation known. However, some systems, receiver may be completely blind number characteristics within bandwidth interest. technique introduced this study proposes collapsing localized magnitude peaks from consecutive short time discrete fourier transform bins into histograms create two dimensional image frequency–magnitude density space. This can useful visualization tool characterization user assisted modes classification. Alternatively, process could automated utilizing pattern recognition processing algorithms. Automatic explored study.