Unconventional dynamic differential GPS orbit solutions

The first of these unconventional GPS applications to be seriously examined was precise orbit determination (POD) in support of high precision ocean altimetry. A global differential GPS technique for achieving sub-decimeter orbit accuracy on the joint U.S.- French TopexlPoseidon mission was first proposed at the Jet Propulsion Laboratory in 1981. The basic elements of the proposed differential GPS system-a small global ground network, a precision flight receiver, the GPS constellation, and an analysis center-are depicted in the picture below.

Over the years, a variety of refinements to the proposed orbit estimation technique, evaluated through simulation studies and covariance analysis, revealed the surprisingly rich potential of tracking device for few-centimeter tracking of orbiters at low altitudes. The Topex/Poseidon ocean altimetry satellite was launched into a 1300 kIn orbit on an Ariane rocket in August of 1992. It carried an experimental dual-frequency P-code receiver built by Motorola to test these new tracking techniques [Melbourne et ai, 1994]. The Topex GPS POD demonstration has now surpassed pre-launch expectations of 5- 10 cm radial orbit accuracy by about a factor of three. A number of aspects of this experiment are notable: 

(1) conventional dynamic differential GPS orbit solutions were essentially equivalent to dynamic solutions obtained with laser and DORIS (Doppler) tracking data, with radial accuracies of 3-4 cm RMS [Schutz et ai, 1994]; (2) reduced dynamic orbit solutions, in which the unique geometric strength of GPS data is used to minimize sensitivity to force model errors [Wu et ai, 1991] consistently improved upon dynamic solutions (judged primarily by altimeter crossover agreements) to yield radial orbit accuracies of 2-3 cm RMS [Yunck et ai, 1994; Bertiger et ai, 1994; Hesper et ai, 1994]; (3) University of Texas investigators used GPS data from TopexiPoseidon to improve the Earth gravity model over what had earlier been achieved by tuning with laser and DORIS data, leading to significantly reduced geographically correlated dynamic orbit error [Bertiger et ai, 1994]; (4) dynamic orbits with a GPS-tuned gravity model surpass those with a laser/Doppler-tuned model, but fall short (by -1 cm RMS) of GPS reduced dynamic orbits [Bertiger et ai, 1995]; (5) GPS tracker based orbits of the highest accuracy are now obtained with a fully automated, unattended processing system; (6) analysis based on Topex results suggests that reduced dynamic orbit accuracies of a few centimeters should be achievable for future missions at altitudes below 500 km [Melbourne et ai, 1994; Bertiger et ai, 1994]; (7) recent unpublished results by Ron Muellerschoen at JPL indicate that carefully tuned onboard dynamic filtering could yield real time non-differential orbit accuracies of a few meters under nominal levels of selective availability.

Since the TopexlPoseidon receiver cannot decrypt the Y-codes, the GPS demonstration has been partially in abeyance since anti-spoofing came on nearly full time in January of 1994. Routine processing continues, however, with Ll C/A-code data, yielding radial accuracies in the range of 4-5 cm RMS, itself a somewhat surprising result. In the wake of the Topex success, GPS-based POD has been adopted for several future altimetry missions, including the U.S. Navy's Geosat Follow-On, which will carry a Rockwell MAGR and is slated for a 1996 launch, and the TopexiPoseidon Follow-On, proposed for launch later in the decade.

More GPS tracking solutions at http://www.jimilab.com/ .