Design and Analysis of IF Signal Preprocessing Methods for the Classification of ELINT Signals

Abstract

Automatic object identification has been a key technology in electronic intelligence systems. Identification process is based on signal analysis. Especially, the signal modulation as level of encryption is important. A new method based on two-stage automatic electronic intelligence objects identification is described in this paper. In the first stage, an automatic classification is designed to automatically sort signals of these objects into one of the four kinds of modulation. The second stage is designed to realize electronic intelligence object recognition. The signal preprocessing and classification using neural network as the main part of the first stage is analyzed and described. Two algorithms of the preprocessing, A and B, are designed and compared in this paper. Simulation results of these algorithms are analyzed. While the algorithm A does not meet all the conditions, the algorithm B is suitable for classification of electronic intelligence object signals with the probability of correct classification higher than 0.95 when the signal–to–noise ratio is above −2dB.