Tracking an Underwater Target in a Large Surveillance Region With Sensor Location Uncertainty

This paper provides an integrated solution of sensor selection and multi-sensor tracking to localize an underwater target using passive angle-only measurements received from sonobuoys, floating in a large surveillance region. It is evident that sonobuoys, airdropped in the surveillance region, undergo drift due to the influence of sea currents. To accurately capture this behavior, the positions of the sonobuoys are modeled using a stochastic difference equation. Due to some physical limitations, only a few selected sensors are allowed to send the measurement data for an interval of time, and they are chosen by solving an optimization problem whose cost function is formulated based on the Fisher information matrix. The effect of uncertainty in sensor location on measurement noise covariance matrix is calculated and the shifted Rayleigh filter (SRF) is modified so that it can be applied to a system with a nonlinear process model. The combined method of sensor selection and target localization is used to track a target moving in (i) a nearly straight path, and (ii) taking a turn with a constant but unknown turn rate. The tracking performance of the developed method is compared with the conventional Gaussian filters in terms of root mean square error (RMSE), averaged normalized estimation error squared (ANEES), percentage of track divergence, and execution time. The proposed methodology with the SRF provides an improved result when other existing Gaussian filters are compared.