Classification of radar emitter signals based on the feature of time-frequency atoms
Ming Qiao Zhu, Weidong Jin, Yunwei Pu, Laizhao Hu · 2007
An effective approach based on the feature of time-frequency atoms for classification of the radar emitter signals is presented. Firstly, we introduce a fast matching pursuit (MP) algorithm, which using improved quantum genetic algorithm (IQGA) to reduce the time-complexity at each step of standard MP, to decompose the signal into a linear expansion of Gaussian chirplet time-frequency atoms. Then, the atoms characteristics are re-extracted to constitute the strong- discrimination atoms feature vector. Experimental results of atoms feature extraction of 5 typical radar emitter signals shows that the atom features have good performances of clustering the same radar signals and separating the different radar signals, which confirms the validity and feasibility of the proposed scheme of signals classification.