The application of neural networks to radar classification

The authors investigate the temporal and spectral features from a variety of different (but similar) radars using a multi-layer perceptron (MLP) classifier. This nonlinear neural network (NN), trained using error back-propagation, has been shown to be able to classify these different radars with a reasonably low probability of error. Additionally an attempt was made to interpret the NN as a (fast) digital filter structure so that it could be readily implemented in discrete technology