This filter uses bearings only measurements to estimate the target state in passive target tracking scenario. This work combines the MGEKF and the iteration method. The filter utilizes the updated state to re-linearize the measurement equation. Then the proposed work is tested in a two dimensional scenario. The simulation study compares the IMGEKF and some other filters to show the improvement. | Journal of Automation and Control Engineering Vol. 3, No. 6, December 2015 Iterated Modified Gain Extended Kalman Filter with Applications to Bearings Only Tracking Yuan Huang and Taek Lyul Song Department of Electronic Systems Engineering, Hanyang University, Republic of Korea Email: hy4335657@, tsong@ Abstract—A nonlinear filter called the iterated modified gain extended Kalman filter (IMGEKF) is presented in this paper. This filter uses bearings only measurements to estimate the target state in passive target tracking scenario. This work combines the MGEKF and the iteration method. The filter utilizes the updated state to re-linearize the measurement equation. Then the proposed work is tested in a two dimensional scenario. The simulation study compares the IMGEKF and some other filters to show the improvement. Index Terms—surveillance, target tracking, nonlinear estimation, bearings only, iteration method I. INTRODUCTION Target tracking problem arises in a variety of practical applications, such as antimissile, aircraft surveillance and GPS. The single and multiple target tracking algorithms are proposed to solve the tracking problem. The target tracking problem considers both linear and non-linear measurements. The bearings only target tracking is broadly used in many passive tracking applications. Typical examples are submarine tracking using a passive sonar or satellite to satellite passive tracking using a radar in passive mode [1], [2]. The bearings only target tracking is an inherent non-linear state estimation problem. The basic problem in bearings only target tracking is to estimate the target state (usually position and velocity) from noise corrupted angle data. For a single sensor tracking scenario, the angle data are obtained from a single moving observer. The problem of observability of the target parameter in passive localization is demonstrated in [3]. To make the target observable, careful designed maneuver must be .
