A Kalman Filter Design for the Space Shuttle Orbiter Inertial Measuring Unit During Deorbit/Reentry Using Global Positioning System Satellite Information.

摘要

This report studies three Kalman Filter implementations for the deorbit/reentry phase of a Space Shuttle Orbiter mission profile. Each design uses measurements from Global Positioning System satellites to update Kalman Filter estimates of selected error states and to bound the Shuttle's INS error growth. Covariance analysis techniques are used to compare the performances of 17-state, 11-state, and 8-state filter designs using a digital computer simulation of a NASA OFT-1 Space Shuttle mission trajectory. A 73-state truth model of the Space Shuttle IMU, the Orbiter GPS receiver Master Tuning Unit, and the GPS satellite constellation is used as a system simulation to examine the performance of the design filters. The results are shown primarily in the form of computer-generated plots of both filter indicated errors and true system errors versus time. It is concluded from the results that an 8-state filter design is able to perform well enough to be a primary candidate for design implementation, and also that further refinements are necessary in the user clock model of the design Kalman Filter. (Author)