The reformulation and modification on iterated EKF for applications with large-dimension measurement

This paper presents a reformulation on the iterated extended Kalman filter (IEKF) using inverse covariance form, and two modifications on it. The reformulation makes a computational saving, particularly when the dimension of measurement is large. The modifications reduce the mean and standard deviation of estimation error. A performance comparison of the proposed algorithms against EKF is presented using a number of simulation examples in a video tracking for small target application.