Particle Filter for Underwater Passive Bearings-Only Target Tracking with Random Missing Measurements

In underwater passive bearings-only tracking (BOT), it is generally assumed that bearing measurements are valid and available at each time index. However, in practice, some of the measurements could be randomly unavailable or invalid on the account of low signal-to-noise ratio (SNR). This work proposes a particle filter (PF) with modified importance weights to tackle random missing instances in the observed bearings. A measurement model which replaces the missing measurements according to the last received measurement and the knowledge of process dynamics, is formulated. Finally, the proposed PF is implemented on a real-life underwater BOT scenario with random missing measurements. Root mean square error and track-loss count obtained from the simulation show that the proposed PF which accounts for the randomness in missing measurements, performs with improved accuracy compared to the existing PF and other Gaussian filters.