Between the Lines: A Fortuitous Meeting

September 24, 2010 By: James L. Farrell

On the last day of ION-GNSS 2010, a serendipitous thing happened. While walking toward the upstairs exit I met Brad Parkinson, alone for just 5 minutes, waiting for his wife. Conversation immediately turned, not surprisingly, toward interoperability. He was very interested in how the Ohio University flight test achieved 1-cm RMS velocity accuracy, over a nominal 1-hour duration (until the flight recorder filled up). What really brightened him up was the method I used to get those results: for sequential (1-sec in my case) changes in carrier phase, it clearly makes no difference whether they came from GPS, Galileo, GLONASS, or any other constellation. Mixing all of them together is completely effortless.

I then noted some of the additional benefits that are more subtle: not only are we relieved of the sometimes-vulnerable cycle ambiguity resolution (integers cancel) but, when phase track is lost, we don't have to wait for the lengthy settling of reacquisition transients; the SV we lost is instantly available to us as soon as we have two consecutive phase measurements that differ by an amount that passes my single-measurement RAIM test.

[Pages 121-125 and 133 of my 2007 book GNSS Aided Navigation and Tracking show rigorous conformance to parity based on matrix factoring but, as explained in the footnote on page 122, without any need to perform those decompositions in operation. Those details wouldn't be discussed at length for a GNSS panel talk; a simple statement that single-measurement RAIM has been developed and validated by flight test would suffice.]

Another subtle — but very important — benefit comes from below-mask SVs. 1-sec IONO/TROPO changes are typically a few millimeters, so despite the unaccepability of near-horizon pseudoranges, we can use their sequential phase changes. That happened with one SV during the OU flight test. It was imperative to reject its pseudoranges (single-measurement RAIM easily did that) but the 1-sec carrier phase changes were as good as gold — and very beneficial for the geometry — all the way down to horizon until it was no longer visible.

That five-minute converstion allowed enough time for me to describe one more feature: the FFT-based processing approach pioneered at OU. Being a quick study, Brad saw the significance of how that dovetails in with my 1-sec-phase-changes-cum-single-measurement-RAIM method. His overall observation was that this offers interoperability plus a way to obtain dead reckoning without an IMU. Sure enough, I used 1-sec-phase-changes-cum-single-measurement-RAIM both with an IMU (producing 1-cm/sec RMS velocity) and without one (producing a decimeter/sec RMS velocity — and all of that is also documented).