Analysis of measured radar data for specific emitter identification

Abstract

Measured radar data assisted in the successful development and implementation of Specific Emitter Identification (SEI) signal processing algorithms. The aim of the algorithm is the identification of a specific emitter within a single class of emitters. The processes developed are pulse extraction, feature calculation, dimensionality reduction and classification. A pulse is detected whenever the phase changes from being random to being linear. Time domain features are then calculated from these extracted pulses to populate a feature vector. Some redundancy exists in the feature vector and a process of dimensionality reduction is followed to obtain the optimum set of features. A fuzzy artmap classifier completes the process of specific emitter identification to demonstrate the improvement in Correct Classification Decisions (CCD) using the reduced feature vector.