Joint Estimation of Source Number and DOA for Underwater Passive Target Based on RJMCMC Method

Conventional wideband signal processing method of subspace class wideband requires more snapshots to get a better estimation result. We apply a Bayesian high resolution direction of arrival(DOA) estimation method based on the time-domain wideband signal model to underwater passive target DOA estimation. Compared with conventional subspace class methods, the proposed method can process signal directly in time-domain and needs remarkably less snapshots. We use a time-domain wideband signal model to construct a posterior probability density function(PDF) of the parameters according to Bayes law, and then use the reversible jump Markov chain Monte Carlo(RJMCMC) method for Bayesian estimation to search peaks of the posterior PDF. This method can jump among parameter spaces with different orders, thus can implement joint model order (i.e. source number) and DOA estimation. Simulation results show that the RJMCMC method exhibits good performance in joint estimation of model order and DOA.