Unscented particle filter for tracking a magnetic dipole target

In this paper we present a recursive Bayesian solution to the problem of joint tracking and classification of a target modeled at a distance by an equivalent magnetic dipole. Tracking/classification of a magnetic dipole from noisy magnetic field measurements involves the modeling of a non-linear non-Gaussian system. This system allows for complications due to multiple directions of arrival and target maneuver. The determination of target position, velocity and magnetic moment is formulated as an optimal stochastic estimation problem, which could be solved using a sequential Monte Carlo based approach known as the particle filter. In addition to the conventional particle filter, the proposed tracking and classification algorithm uses the unscented Kalman filter (UKF) to generate the transition prior as the proposal distribution.