DeepASTC:Antenna Scan Type Classification Using Deep Learning

Abstract

In this work, we propose a new method which we call DeepASTC, for antenna scanning type classification in Electronic Warfare Systems. DeepASTC is a deep neural network composed of LSTMs. Amplitude patterns of the deinterleaved radar pulses are fed into our network, and the corresponding scanning type is automatically obtained. DeepASTC and the Multiclass Support Vector Machine (SVM) based classifier method are compared. It is observed that the proposed DeepASTC is able to achieve 93.8% correct classification rate on average, whereas the corresponding rate for the Multiclass SVM method is 86.3%. Conducted experiments show that, the proposed DeepASTC performs successfully on the synthetic data sets.