| The space environment presents challenging operating conditions for a GPS attitude receiver. While spacecraft vehicle attitude dynamics may be at a lower rate than terrestrial vehicles or aircraft, they may undergo a much more general attitude motion. For example, for inertially fixed spacecraft or spinning spacecraft there is no preferred orientation of an array of antennas with aligned boresights that will give good GPS constellation visibility and ensure continuous attitude solution availability over an entire orbit. For these cases of general vehicle attitude motion, continuous attitude solution availability can be achieved using an array of antennas pointing in different directions. However, such a non-aligned antenna array requires modifications to the receiver signal processing algorithms to make attitude determination possible and to achieve accuracies comparable to an aligned array. Non-aligned antenna arrays introduce LOS dependent phase errors into the differential carrier phase measurements, the most significant of which are; the right hand circular polarization (RHCP) contribution to the carrier phase, antenna phase center variations, and multipath error.; In this dissertation, the results of two experimental laboratory demonstrations of attitude determination using measurements from a non-aligned antenna array are presented, along with a simulation study for the on-orbit case. In the first laboratory demonstration, attitude estimation was done (in post processing) using a passive, spinning platform with only single degree of freedom rotation, in a laboratory environment that had very significant reflected multipath signals, and near transmitters. The experiment used a 2-D air cushion vehicle on a granite table top with a single antenna baseline measurement. The results of this experiment showed that robust attitude solutions could be attained from differential carrier phase measurements taken from an array of non-aligned antennas, but also highlighted the importance of compensating for the RHCP in the measurement. Continuous attitude estimates with RMS errors on the order of 0.25°–0.50° were achieved using antenna baselines on the order of 1.0 in, after compensating for the limitations of an indoor GPS laboratory multipath environment.; A spinning 3-D test platform was also constructed to allow for full motion attitude sensing and closed loop control in a laboratory environment with much larger transmitter distances. The test platform has a non-aligned array of four antennas connected to a center node that is suspended from an overhead cable. The platform has three axis attitude control, and rate gyros to provide attitude truth measurements. This new 3-D experiment platform was used to demonstrate real-time attitude estimation, including real-time measurement bias estimation for periodically occluded measurements. |
