Least Squares Procedure

Hi Don,

Do you mean field procedures?

Over here we do double face reading for our traverses and always try to tie our traverse across to provide redundancy.

I find least squares works best with redundancies in the network and can be used as checks too.

We use Starnet for the adjustment and we will just upload the raw data for the least squares adjustment. Adjusting the angles before doing least squares would not give the true angular correction.

> ... It was recommended by some to perform a least squares adjustment given the right collection procedure.
Your collection procedure need not change, at least not at first. A closed loop traverse is still a very desirable thing with LS.

A Closed Loop Is A Required Thing In LS

A closed loop provides the required redundancies such that least squares has enough information to properly adjust the traverse.

Since your question came simplistically it is best to state a closed traverse requires setting an instrument at every traverse point and record every backsight to foresight angle. Angles should be record with each face and distances should be recorded for backsights in addition to foresights.

Paul in PA

A Closed Loop Is not A Required Thing In LS

> A closed loop provides the required redundancies such that least squares has enough information to properly adjust the traverse.
True, but a closed loop is not the only way to achieve redundancy.

Agreed, A Network And/Or GPS Provide Redundancy

With the expense of extra setups.

Paul in PA

1. Adjusting a single closed loop in LS will produce extremely similar results as a Compass Closure. Generally, we're talking thousandths.

2. Where is the advantage? When you grab an extra location. Run your closed loop, setup in the middle and resect to 2 or more points on the loop. Now you're grabbing data that Compass can't handle, but LS can.

3. Read a book. If you can handle the math, Wolf/Gilani is very useful.

4. What else can you add? 2d+1d solution for most surveying. 3d solution with the inclusion of GPS vectors.

I use survnet and I have a whole wonderful procedure.

I suggest working through the provided examples with Carlson SurvNET as they'll shed a lot of light on possible procedures. Cross ties won't do you much good if you don't know how to tell the software that you did one (same goes for loops, actually).

Also, get familiar with reading a Carlson raw file if you haven't yet. I would suggest going through the example lessons on Edit Raw file first.

You'll learn a lot just from seeing the BD,FD,FR,BR nomenclature on different lines in the raw file.

1. Adjusting a single closed loop in LS will produce extremely similar results as a Compass Closure. Generally, we're talking thousandths.

2. Where is the advantage? When you grab an extra location. Run your closed loop, setup in the middle and resect to 2 or more points on the loop. Now you're grabbing data that Compass can't handle, but LS can.

The big thing I liked about getting star-net back in the day was that I could enter the data and it would adjust the traverse so fast, really increased the work flow, but the results of our traverse adjustments were basically the same as a compass rule, minor, very minor differences.

It also became obvious that adjusting multiple traverses together was the real advantage and incorporating cross ties to strengthen the results, but a simple closed traverse,,,,,,,least squares, compass rule.......eh...big shrug.........

Thanks for the replies

We ran a 26-point traverse at Wentworth with about 10 crews. The first crew done setup inside a building and resected off a few points.

I was not on the crew that was first done and everybody behind my crew was slowed because we broke down every tripod setup, tightened the bolts and then proceeded with the traverse. The tripods were all from offices and should not have been so loose.

We were then given all the data sets. I ran iterations of all ten sets then created a 'prismatic' plan showing each point at 1" = 0.1' or so. I also ran one closure with Compass just to show how close it was. It was cool to see how the additional data changed the location by a little bit. I did not go so far as to determine where the tripods were repaired, but I did highlight my traverse as the likely best solution because of technique.

Good times, good times...

You can run an open traverse in LS, but without redundancy, it will not be able to adjust the traverse. One technique that can be used to provide redundancy on an open traverse is to shoot a set (forward and reverse, preferred) to intermediary points between the traverse points. For example, between the traverse points, set another nail midway. Shoot it forward from one traverse point, and then shoot back to the same nail from the next traverse point...

Don, if you have a RW5 file already, I'd be happy to try to help you process it in SurvNET. I'm rather new myself to the procedure, but I think I've learned enough to be helpful to someone new at it. Dean Goodman at Carlson has been very helpful to me, when I need technical expertise....

> One technique that can be used to provide redundancy on an open traverse is to shoot a set (forward and reverse, preferred) to intermediary points between the traverse points. For example, between the traverse points, set another nail midway. Shoot it forward from one traverse point, and then shoot back to the same nail from the next traverse point...
Another method would be to - after turning the angle BS to FS - make the FS the BS and the BS the FS and "close the horizon". This method has it's shortcomings, but it does provide redundancy. And set your dc to record distance to the BS as well as the FS. Doing so adds very little time in the field.

Personally I like to sprinkle in some GPS vectors for redundancy, which has the added advantage of putting the traverse on the grid.