Recognizing aircraft type using a support vector machine and a higher order cumulant

Two methods were examined that could provide aircraft type recognition from background sounds in shortwave communications with aircrew.Physical characteristics of vectors of the acoustic signals from aircraft were extracted by use of wavelet packet decomposition(WPD) and higher order cumulant(HOC).A support vector machine(SVM) and a back propagation neural network were adopted as classifiers in turn to identify five kinds of aircraft.Experimental results of aircraft type recognition showed the WPD-HOC-SVM system identified significant features of aircraft cabin background sounds,producing excellent results——a 93% recognition rate.