The marginalized Kalman filter algorithm for multi-plane passive tracking

The measurement equation of multi-station passive tracking system is strongly nonlinear and thus more demands are required for the tracking algorithm. To realize robust and fast tracking, a novel passive tracking algorithm is proposed based on the marginalized Kalman filter. The proposed algorithm expresses the nonlinear measurement equation as a weighted sum of Hermite polynomials up to p order and then the prior distribution of the weight matrix is modeled as a Gaussian process. The influence of the weight matrix is removed by marginalizing it when its posterior distribution is available and then the close-form solution of the target state and its covariance can be got. The bearings-only tracking problem is taken as an example to verify the performance of the proposed algorithm. Simulation results indicate that compared to the extended Kalman filter (EKF) algorithm, the unscented Kalman filter (UKF) algorithm and the cubature Kalman filter (CKF), the proposed algorithm has improved tracking performance.