Robust Decoding of Costas DFC Waveforms With Hidden Markov Model

Robust frequency coding sequence estimation of the Costas discrete frequency-coded (DFC) waveform signal is an important task for radar reconnaissance interpretation in radar electronic warfare. However, due to the unknown waveform characteristics of the signal, conventional frequency coding sequence estimation algorithms are not applied in noncooperative target scenarios, especially those with a low signal-to-noise ratio (SNR). Therefore, based on the hidden Markov model (HMM), a frequency coding sequence estimation method for Costas DFC waveforms is developed in this article. Through associating the time–frequency distribution of the Costas DFC waveform signal with the HMM, the frequency coding estimation of the waveform can be transformed into the sequence decoding problem of the HMM. In the proposed algorithm, an effective decoding model library can be obtained after HMMs are initialized and trained. Then, the time–frequency observation sequence of the Costas DFC waveform signal is used to select the optimal decoding model from the library. Finally, the Costas frequency coding sequence of the waveform is estimated by the Viterbi algorithm. Simulation experiments verify that the proposed algorithm has strong decoding capability and robustness for various Costas DFC waveforms under different SNR conditions.