Kalman filter for extended object tracking

In this work, we present a novel method for tracking an elliptical shape approximation of an extended object based on a varying number of spatially distributed measurements. For this purpose, an explicit nonlinear measurement equation is formulated that relates the kinematic and shape parameters to a measurement by means of a multiplicative noise term. Based on the measurement equation, we derive an extended Kalman filter (EKF) for a closed-form recursive measurement update. The performance of the proposed method is demonstrated with simulations.