Sea-surface targets viz ships, boats etc. exhibit benign motion and rarely make maneuvers. The problem of tracking slow moving surface targets has to be treated differently from that of tracking high speed airborne targets. Absence of maneuvers may simplify the kinematic model used for estimating the target position, but getting good accuracies of the target course estimate may become difficult with reduced angle-measurement accuracies (especially, in the absence of monopulse, and with targets having measurement extensions in multiple angular bins). With airborne radars, this may further degrade with biases in the platform motion sensor measurements. Tracking in such a scenario is challenging especially with respect to providing stable and accurate estimation of target speed and course. This paper brings out a comparative analysis of target-velocity and target-course estimation accuracies with three different filtering algorithms - namely, Kalman filter with a constant-velocity model and fixed coefficient smoothing, Extended Kalman filter with polar velocity state and Particle filter.
展开▼