Least Squares Adjustments

> Hello,
>
> Please ask your advise if anyone knows if GNSS Solutions uses Least Squares Adjustments for GPS post processing? I was under the impression that most if not all post-processing software uses LSA in their algorithms. Thank you.

yes and no. LSA and GPS post processing are two different things. GNSS solutions does have LSA with the "F7" adjust network command, but that would be subsequent to processing vectors though

Ditto Moe's response.....

> Please ask your advise if anyone knows if GNSS Solutions uses Least Squares Adjustments for GPS post processing? I was under the impression that most if not all post-processing software uses LSA in their algorithms.

The subject is outside my area of expertise, but here's what I *think* I know: that least squares estimation is a special case of the more generalized Kalman filtering, and that Kalman filtering is well-suited to analyzing noisy data streams like those comprising GPS data. So I think as long as one is willing to accept some imprecision in the use of the terms, it's correct to say that GPS vector processing engines use a kind of least squares adjustment in calculating the most likely vector values.

I think I'd agree with that.

I would point out, though, that the Kalman filtering is done over one set of data (the time delays from each of the satellites) to produce an estimate of one position, and the typical least squares processing done in other software is on another set of data to relate multiple measurements of relative point positions.

I thank your responses.
Reason I asked your opinions is because a client wanted to have all GPS Control Points to be adjusted with LSA method. I informed him that the output from post-processing software for repeated observations (vector adjustments) was done in LSA. I could not get documentary proof though. LSA was mentioned in the Help section of GNSS Solutions but not specifically saying that it was used in the software.

Your client wants you to do redundant measurements then adjust the vectors to provide reliable control values. This is best done by performing multiple occupations of each point and adjusting between them and any fixed control you are using.

In addition to redundancy, consider a network of interdependent observations. That will do more to validate the field portion of the survey...

Control should be done using at least 4 receivers together, more is preferable, leap-frog is the simple way to run it, need to remember that each point be tied to the nearest point and all points have at least two occupations. When done properly there is no doubt that the control is good....