In situations like the "urban canyon", where the problem is that only a small slice of sky is visible, I would think that the advantages of using both systems w

In situations like the "urban canyon", where the problem is that only a small slice of sky is visible, I would think that the advantages of using both systems would only be marginal. The most accurate GPS fix is had when the satellites you're seeing are widely separated in the sky. In the urban canyon, the problem isn't a lack of satellites, but the fact that you can only see those that happen to be in a narrow slice. And narrow slices are bad for constraining the solution to the equations. The way I think about it is each satellite being used in the solution is like a rope tied to the satellite (where stretchiness of the rope is analogous to measurement error). Your GPS receiver can calculate your location to within the accuracy of how far you can move while keeping all ropes taut. So if you had a fix from two satellites that are smack next to each other, you really don't have a fix any better than a single sat. If you've got two ropes at 180-degrees opposite each other (i.e., at either end of the street in the urban canyon), you've still got ambiguity due to your potential for lateral movement. For a good fix, you've got to bring in another rope at a wide angle from those -- but the tall buildings interfere with getting that rope, even if there are lots of satellites up there that could be offering it.