The invention discloses a coherent matching field tracking method and
system based on trajectory Poisson multi-Bernoulli
hybrid filtering, and belongs to the technical field of
underwater sound. Local peak values are extracted from a multi-frame
ambiguity surface obtained through coherent matching field
processing, clustering is carried out, and a three-dimensional measurement set containing target distance, depth and radial speed
estimation information is constructed; on this basis, a matching field trajectory
state space model is established, a trajectory Poisson-Bernoulli
hybrid filter is utilized to perform
state prediction in a trajectory space, and a prediction result is updated in combination with a three-dimensional measurement set; and realizing
data association based on the likelihood consistency of the distance, the depth and the radial speed so as to complete target trajectory
estimation. According to the method, clustering
processing is carried out on the peak values extracted from the three-dimensional
ambiguity surface, a large number of highly similar redundant trajectories near a real trajectory are effectively reduced, and the calculation burden of an
algorithm and the uncertainty of
data association are reduced. The method aims at a matching field tracking problem in a low
signal-to-
noise ratio environment.