Bayesian Estimation-Based Algorithm for Multistat-ic Sonobuoy Fusion Location

This paper proposes a new Bayesian estimation-based multistatic sonobuoy fusion localization algorithm for the precise localization of underwater target sources in com-plex marine environments. The algorithm combines multiple observations, including the duration of a single signal transmission, time of arrival (TOA), and azimuth of each receiving buoy, while considering the positional uncertainty of transmitting and receiving buoys and the correlation between these parameters. The conditional probability density function is constructed to estimate the target position by maximizing its probability distribution, and it is combined with the covariance matrix to improve the positioning accuracy. Compared with the traditional localization methods, which only rely on single time or bearing information, this paper proposes a Bayesian framework that jointly estimates both the target position and sound velocity while optimizing the search peak using the lim-ited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) algorithm. Simulation results show that this algorithm outper-forms traditional localization algorithms in scenarios involving multistatic detection with significant errors in station location and time measurements.