A unified approach to post-mission processing of intertial data

The outpout of inertial survey systems is available to the user in two basic forms: as Kalman filtered information at updates or as integrated velocity and position information at regular time intervals. In case of the second data type, the post-mission processing starts with the approximation of the velocity error curve. This approximation is either based on a system error model as in Kalman filtering or uses curve fitting techniques. From there on, smoothing or adjustment procedures are used as further steps in the post-mission treatment of both data types. A unified treatment of the various post-mission approaches starts with the formulation of appropriate error models of the system outputs. It is then possible to present all existing methods as intermediate steps of a rigorous adjustment procedure. This unified approach gives insight into the limitations of individual methods and provides a means to detect inconsistencies in post-mission processing strategies. An analysis of existing approaches is made and a new method, spectral decomposition, is treated in detail. Compared to the existing procedures, it has advantages with respect to a rigorous covariance propagation and blunder detection.