Extended Kalman-Bucy filters for radar tracking and identification

A class of extended Kalman-Bucy filters used for tracking ballistic and lifting objects as well as vehicle identification is described. The primary filter, in this class, is based on estimating position, velocity, and aerodynamic drag of a given object from radar measurements. One key design feature of these filters is to model the vehicle dynamics in a Cartesian coordinate system while modeling measurements in a range-direction cosine (i.e., spherical) frame. A consequence of this selection of multiple coordinate systems is the elimination of bias induced from the linearization of the measurement model. The drag state and aerodynamic vector estimates are used as deceleration gages, whereas position and velocity estimates are utilized to maintain track. The rationale for the use of distinct coordinate systems, design features, and state selection is discussed