Detection and Identification of EM Field Source by Using Support Vector Machines

An approach of tempest analysis is introduced and a method of detection and identification of electromagnetic emitting systems by intentionally or unintentionally is applied and discussed. Based on Support Vector Machine(SVM) algorithm, digital signal process method, this paper investigates an primary procedure involving red/black separation and classifying the emission sources radiated from electronics equipments such as the PCB layout or others. The transmission of digital clock and oscillation in ARM embedded system including tracks configuration and I/O Ports is taken into consideration. A set of computed spectra data of emission caused by PCB is adopted for the learning process. A simplify method to train the SVM is used by extracting the characters like the mean of magnitude, standard deviation and other parameters extracted from the data.