An algorithm for space-based tracking with bearings-only measurements

We describe and discuss different algorithms for estimating the state of a target based on angles-only measurements from a space- borne platform. The target state is propagated using a quasi-linear dynamical model while the nonlinearity is explicit in the state-to-measurement map. A global linearization is formulated where the associated Jacobian depends explicitly on the measurements. This leads to a special form of the extended Kalman filter (EKF). We discuss several implementation issues including the effect of measurement bias and noise, target dynamical models, the on-line estimation of measurement noise statistics, and the assessment of the filter performance. We also contrast the performance of this filter with alternative implementations.