Wavelet Based Approach to Transmitter Identification

摘要

Research is conducted to find a robust wavelet based algorithm for automaticidentification of push to talk radio transmitters. Digital data at an intermediate frequency (IF) is preprocessed to translate it into a form applicable to wavelet analysis. The magnitudes and the relative positions of the extrema from the best suited wavelet scales are used in conjunction with a distance measure algorithm to determine from which transmitter any particular signal may have originated. The distance measure algorithm can correctly identify the four signal sets provided for the research. Even at reduced signal to noise (SNR) ratios good identification is obtained. The current version of the algorithm can classify all sets, with at least an 80 percent reliability at an SNR on the average 10 dB worse than the original recordings. The procedure lends itself to an automated, minimizing human interaction.