Comparison of Artificial Intelligence Algorithms for ELINT Signals Classification

Abstract

Ability to automatically identify objects of interest and to automatically classify their signals belongs to essential functionalities of electronic intelligence systems. Objects identification results and signals classification results are conditioned by accurate measurement and technical analysis of objects signal parameters. Classification algorithm based on Feedforward Neural Network, and classification algorithms based on Support-Vector Machine, with three types of Kernel Function, are tested and compared in this paper as the first stage of objects identification in electronic intelligence systems.