Can anyone direct me to a text book or white paper describing, comparing and/or critiquing these methods? I have three Ashtech Locus receivers that have performed well with the rover method, i.e. 2 base stations on known locations (on NGS control checks). The results have been impressive. The Ashtech manual is generally helpful, but doesn't give clear examples of these methods. I have read several opionions comparing these methods which have been affected by experience. But, the user's haven't always been clear on their application or the geometry of their project (hard to do on-line). Our reason for purchasing the Locus receivers was to establish control on small photogrammetric mapping projects, say a few hundred acres or less. The rover method works fine with a one man crew. But for highway mapping over long stretches, it seems the leap frog method would be better with a two man crew. I'm still learning and would appreciate any info on these methods. Thank you.

Eric
I asked the same question on the old Ashtech site at:
http://www.rpls.com/locus/messages/488.html

and got these answers:
1/
IN THEORY YES, BUT YOU'LL HAVE to re-SOLVE CRD's by PAIRs, MEANING SOLVE B BASED on A, then C BASED on B and SO FOURTH. GUARANTEED CONFUSION HOWEVER.
2/
Your leapfrog traverse is a valid approach to performing a GPS survey. The LOCUS System will have difficulties in handling this.

2/ was later corrected:
Your leapfrog traverse is a valid approach to performing a GPS survey. The LOCUS System will NOT have difficulties in handling this.

Question3/
wouldn't you have a better solution leaving the base on A, and moving the rover to B, C, D and E? If you have an error in the first baseline, won't that translate all the way down the line?
Answer3/
You are correct in your statement. Just like a conventional traverse, error will accumulate and propagate through the observations. Each leg will have some setup error and some observation error that will add up as you move down the traverse. This is a disadvantage to the traverse approach.

But, there is an advantage to the traverse approach. If, when using the traverse approach, the user loops back to the original point to close the loop or begins on a known and ends on another known, you have a check of the quality of your data (beyond the quality indicators given to you by the processing software). Using the side-shot approach (leaving base on A and moving rover to B,C,D,etc.), there is no redundancy, i.e. no check. Although LOCUS Processor is good at giving dependable uncertainties to the observations, it is not perfect. It is possible that an observation will show good uncertainties but be poor (that's the true nature of statistics). The only check for this is redundancy.

Bill Martin
Ashtech Precision Products