Underwater acoustic channel tracking by multi-bernoulli filter

Most underwater acoustic channels exhibit multipath arrivals. Such channels are often time-varying with extensive delay spread which challenge the channel tracking. In this work, we propose a model-based channel tracking method which captures the signal propagation physics. A newly-proposed tracker named Multi-Bernoulli filter (MBF) is used in this model under the framework of Kalman filter (KF) to improve the tracking performance. This Bayesian tracker has a good balance between measurements and prediction. Simulation results presents a significant improvement on parameter estimation comparing to channel estimation by Compressed Sensing (CS) method. Furthermore, more relative channel physics such as the establishment, hiddenness and vanishment can be analyzed from tracking results which is also examined by MACE10 experiment data.