Robust intra-pulse modulation recognition via sparse representation

Aimed at the problem that overlapped signals and noise interference in complex electromagnetic environment cause recognition rate declined, this paper proposed a robust radar signal intra-pulse modulation recognition method, instantaneous frequency and frequency spectrum feature are extracted for online dictionary learning which represent each type of modulated pulse, and recognition results are acquired by selecting the minimum residual of sparse reconstruction, the experiment results demonstrate that the proposed algorithm can obtain higher recognition accuracy with low SNR and overlapped pulses condition, which is more applicable for the practical applications.