A space-frequency joint detection and tracking method for line-spectrum components of underwater acoustic signals

Abstract

For the passive detection of azimuth and frequency of azimuths and frequencies of line-spectrum signals radiated from underwater weak targets, a space-frequency joint detection and tracking method based on track-before-detect (TBD) technology is proposed. This method works directly with the time sequence of frequency-azimuth (FRAZ) spectra to estimate the frequency and azimuth of line-spectrum signals. And the change of the frequency and azimuth states of a line-spectrum signal is modeled as a first-order hidden Markov process. The distribution of the FRAZ spectra of linear array received signal are derived in this paper to set the parameters of Hidden Markov Models (HMM). The FRAZ spectra at different times is joint process using Viterbi algorithm to obtain the maximum a posteriori estimation of the azimuth and frequency of line-spectrum signal. A dynamic searching method based on region segmentation and merging similar line-spectrum is designed to solve the problem of large computation of Viterbi algorithm in the detection of line-spectrum. The proposed method can make full use of the correlation between the target's information in the multi-frame FRAZ spectra to distinguish the target from the background noise. The processing results of simulation data and actual data show that the method has excellent line-spectrum detection performance at low SNR and can adapt to the complex situation such as there are multiple targets and multiple line-spectrum signals.