Tracking Before Detecting Weak Targets Based on Dynamic Programing in Passive Sonar Scenarios

In passive scenarios, effective detection and precise tracking of weak targets represent key research directions in underwater acoustics. The traditional detect-before-track method is easy to miss detection in low signal-to-noise ratio (SNR) scenarios. Track-before-detect based on dynamic programing (DP-TBD) has more advantages in handling weak target scenarios. However, in scenarios with target crossovers, there is a risk of trajectory merging, where two target trajectories combine into a single trajectory. This study proposes a speed weight-based method to solve trajectory merging problem. When constructing the value function, adding a speed-weighted function enables the trajectory to be correctly separated in the case of intersection.