Angular rate estimation by multiplicative Kalman filtering techniques

Abstract

Spacecraft attitude estimation and pointing accuracy have always been limited by imperfect sensors. The rate gyroscope is one of the most critical instruments used in spacecraft attitude estimation and unfortunately historical trends show this instrument degrades significantly with time. Degraded rate gyroscopes have impacted the missions for several NASA and ESA spacecraft, including the Hubble Telescope. A possible solution to this problem is using a mathematically modeled dynamic gyroscope in lieu of a real one. In this thesis, data from such a gyro is presented and integrated into a spacecraft attitude estimation algorithm. The impediment to spacecraft attitude estimation presented by imperfect sensors has been overcome by developing more accurate sensors and using Kalman filters to reduce the effect of noisy measurements. Kalman filters for spacecraft attitude estimation have historically been based on an Euler angle or quaternion formulation. Though Euler angles and quaternions are arguably the easiest methods with which to describe the attitude of a spacecraft, other methods of describing attitudes do exist - including the Gibbs and Rodriguez parameters. A Kalman filter based upon the Gibbs parameter is presented and analyzed in this thesis.