Robust radar waveform recognition algorithm based on random projections and sparse classification

To solve the limitations existing in the feature-based recognition method in information completeness and redundant of the representation and the noise robustness, a radar waveform recognition algorithm based on random projections and sparse classification (SC) is presented. Construction of the framework of the proposed algorithm consists of two phases. In the first phase, the compressed signals by random projections are used instead of traditional signal features to represent the original signal. In the second, the robust SC approach is investigated to the radar waveform recognition. Compared with the existing feature-based recognition method, the random projections and sparse classification-based algorithm can improve the information completeness, efficiency and the noise robust. The validity of the recognition algorithm is demonstrated with the analysis and simulations.