A Time-frequency Atom Approach to Radar Emitter Signal Feature Extraction

In this paper, a novel approach to extract the features of radar emitter signals in the high density, complex and variable signal modulation environment is presented. Based on the over-complete time-frequency atom dictionary, the signals are decomposed into a linear expansion of atoms by the method of matching pursuit (MP). Then, improved quantum genetic algorithm is applied to effectively reduce the time-complexity at each search step of MP, and thus some optimal time-frequency atoms describing features of signals are obtained, which can provide some new feature parameters for the deinterleaving and recognition of the radar emitter signals subsequently. Experiment result proved the validity and feasibility of the approach and that the extracted atoms had the features of certain extent noise-suppression ability