New Stable, Vectorizable Kalman Filter Algorithms Based on Orthogonal Transformations

摘要

A stable Kalman filter algorithm based on a Givens transformation is developed theoretically together with three separate update formulations. The backward filter of Rauch-Tung-Striebel is reformulated. As an alternative orthogonal transformation, the singular value decomposition is investigated. The advantage of the Kalman filter algorithm vs. conventional formulations is examined for the support of inertial navigation systems (simulation and flight data) on serial computers and a vector processor.