Digital Modulation Identi Cation And Modulation Orders Estimation Using Wavelet Transform

Automatic modulation identication is rapidly evolving in many areas mainly inrnmilitary applications and research institutions. The identication methods arernbasically categorized as likelihood based (LB) and feature based (FB) approaches.rnIn this thesis FB is proposed to study modulation identication of received signalsrnin the presence of additive white Gaussian noise (AWGN) using wavelets. ThernHaar wavelet was used as the mother wavelet. The algorithm identies 13 modu-rnlation schemes 4 for FSK, 3 for QAM, 3 for ASK and 3 for PSK modulation typesrnwithout prior knowledge. The correct identi cation ratio has been analyzed basedrnon the confusion matrix for di erent modulation type at di erent signal to noisernratio (SNR) and the intra-class and inter-class identi cation of those modulationrnschemes are evaluted. The correct intra-class identi cation ratio was greater thanrn99%, 97%, 96% and 83% at thier lowerest SNR bounds 5dB, 8dB, 8dB and 25dBrnfor FSK, QAM, ASK and PSK modulations respectively. The proposed methodrnis relatively robust for noisy signal and identies more modulation schemes com-rnpared to related exsiting works.rnKey words: Feature based, Modulation identi cation,Waveletrnand Histogram

Get Full Work

Report copyright infringement or plagiarism