Today I received this question:
I have a question regarding recommended occupation times for a MM120 system (using internal antenna) with post-processing. I watched your videos on ‘best GPS practices’, your field collection and the follow-on post-processed example. I wanted to
a) figure out if your results (i.e. a few cm variation with hand-held unit) were typical, and
b) to experiment with occupation times to figure out the time-vs-precision tradeoff
I ran this experiment yesterday with a MM 120 mounted on a tripod, and acquired a number of points for 60, 120, 300, 600 and even 1200 sec, all in the same location. I did not move the tripod in between collections. I post-processed things this morning and they are *all in the same location*. I am skeptical of anything with zero uncertainty, so can you explain what is going on? Is there a different method I should use to run this experiment? Do you have any general guidelines for optimal occupation times with this system?
Thanks for the videos and support,
This is a great question that is asked often. I wrote this (what I think is a pretty good, albeit snarky) reply and I think others will benefit from my answer:
Dear Kind and Most Valued Customer:
Yup, All The Same Position.
The mistake that you made was you should have closed the file and opened a new file… Or, typically what I will do is have an external antenna on a quick connect and shuttle it in a figure-8 pattern between the two points. (You don’t want to move along a linear line between the points as the receivers are smart enough to do some funny business and hot-wire the position knowing the velocity and direction.)
If you don’t move the receiver, the post-processing software treats the entire occupation as a SINGLE occupation (which it clearly is) and computes a solution for the entire occupation and then assigns that single computed value to all of the points that contributed to the seemingly extremely long occupation. Would you have it otherwise? J
So, that said, let’s make a list of things that affect the solution:
O Total length of carrier phase tracking in the file before (and after) the occupation (We can talk about this for hours)
O Distance to CORS, or in the case of a Virtual Base the effective distance to CORS?
O Geometry of surrounding CORS?
O Dual Frequency or Single Frequency?
O How many constellations are you tracking? (GPS, GLONASS, BeiDou, Galileo)
O Which SV’s are in common between the base (CORS) and rover?
O SBAS, and then WAAS (twice as good as) or ENGOS?
O Canopy: none, moderate, heavy, forest or urban?
O Is the receiver’s antenna hand held (and moving slightly) or is the receiver securely mounted on a tripod?
O Multipath: hard surfaces (like steel buildings and roofs) nearby?
O Internal Antenna at waist level vs. External Antenna on a pole above your head and shoulders
Since there are so many variables, I can’t answer your question with a simple answer. In fact, if I am in the field, I often will over and under estimate the results that I will obtain. I can make this prediction:
If you collect GPS only, L1/L2 data and process in OPUS a 4-hour occupation in open or light canopy, low multi-path conditions will converge to 4-cm almost anywhere in the world.
The solution is of course to purchase an RTK Base / Rover pair. Then you can evaluate the accuracy of the baseline from the base to the rover in real time. In real-time you would know the precision of the result. Typically you would usually have enough base data that an OPUS solution would nail down the base position well enough.
However, if you are working under heavy canopy I would also like to make these comments:
O when you move, carrier phase tracking is most certainly fully lost between points. Each point is a new, stand-alone occupation.
O the occupation data may be so trashed by SV’s moving behind tree limbs that post-processed data will be even worse than just WAAS corrected real time data.
O And RTK, even with 4-constellation tracking, won’t ever produce a fixed solution!
In this case, you won’t ever get the precision you seek using GPS measurements.