Detection Of Bearing-Only Trajectory Of Multi-Target Using Hidden Markov Model

Detection and track of the weak trajectory in the Bearing-Time Record (BTR) is an important method to discover and localize the target automatically in passive sonar. An algorithm to detect and track the trajectories of multi-targets in the BTR is proposed in this paper. It doesn't need the number of targets as prior information, and initial or terminate the trajectory automatically. The tracking of each trajectory is independent and can be parallelized. The detection performance is analyzed by the simulation. The effectiveness of this algorithm is also shown by a real acoustic data from experiments on the sea.