An integrated GPS attitude determination system for small satellites

This dissertation develops attitude determination methods based on the Global Positioning System (GPS) for small satellites. A GPS attitude receiver is used in combination with other sensors planned for a small, three-axis stabilized satellite called JAWSAT. The other attitude sensors include fiber optic gyros and digital sun sensors.; The development of integrated attitude determination systems contributes to critical national technological objectives identified for small spacecraft. A recent study by the National Research Council addresses key technologies for small satellite programs. One of their principal recommendations was that, "GPS in various combinations with other guidance components can determine position and attitude very accurately, probably at significantly reduced weight and cost" (NRC, 1994, p. 4). The report also identifies specific potential benefits of integrating GPS with other sensors on small spacecraft. "Combining GPS and an inertial measurement unit (with gyroscopes, accelerometers, or trackers) offers major advantages by bounding errors of the inertial set, providing exceptionally good long-term references and thereby ensuring precise, on-board navigation and, with appropriate complimentary techniques, providing a higher level of redundancy and/or accuracy for position, velocity, and attitude" (NRC, 1994, p. 61). This dissertation develops algorithms that result in improved accuracy and redundancy through the development of complimentary techniques for combining GPS measurements with gyroscopes and sun sensors.; A measurement differencing Kalman filter algorithm for spacecraft attitude determination is developed that results in improved accuracy in the presence of GPS multipath errors, the primary error source in GPS based attitude determination systems. Multipath causes time-correlated errors in the GPS attitude measurements; these time-correlated errors are mitigated using the measurement differencing approach. Improved redundancy is achieved by a decentralized state estimation method based on the federated Kalman filter. Using a similar measurement differencing technique, the decentralized filtering approach achieves improved accuracy in addition to improved redundancy for an integrated system consisting of GPS, gyros, and sun sensors. Finally, this dissertation addresses additional spacecraft applications where integrating measurements for a GPS attitude receiver with other sensors may lead to significant improvements in cost and performance.