Satellite Attitude From a Single GPS Antenna

The motion of an unpropelled satellite in orbit is predictable to a high degree of accuracy, and so enables effective filtering of GPS navigation data received, both in terms of position and attitude. The experimental GPS Navigation Unit on the PoSAT microsatellite offers a novel opportunity for the implementation of such real-time data filtering. The dual transputer system offers a powerful space borne parallel processing host for the GPS receiver. Following the launch in September 1993, it will be possible to upload the data processing algorithms to the PoSAT Navigation Unit to enable comparisons of simulations and real results. An eight-state Extended Kalman Filter has been developed which will accept pseudoranges and Doppler shifts from the GPS satellites as observables and maintain an accurate orbit model in real time. This will directly benefit the GPS experiment on PoSAT by providing the microsatellite’s position when GPS satellites are lost, or even when the GPS receiver is switched off. This paper describes how the Kalman filter was extended to 15 states to include attitude information from the GPS receiver. Although the receiver has only one antenna on PoSAT, it is still possible to derive limited attitude information from the antenna gain pattern and the signal cut-off effects. The simulation results are presented together with predictions for the practical application to PoSAT data. Future UoSAT microsatellite missions will contain a full GPS attitude determination facility within the Navigation Unit, based on the filtered measurements received from multiple GPS antennas.