Radar target recognition using multilayer back propagation neural network (MLBPN)

The authors have attempted to characterise the Doppler modulation aspects of radar echoes from rotating blades (experimental radar targets) using fast Fourier transforms, cepstrum and linear predictive coding. This paper attempts to recognise the experimental radar targets by extracting target features through conventional signal processing techniques and then feeding the feature vectors to a multilayer back propagation neural network. The results point to a major increase in capabilities of future radar systems