Best way to get base station GPS Coordinates

[Jason Franciosa]

What is the best way to get the most accurate GPS Coordinates without an expensive hand held GPS or a known accurate point when using the piksi's with a pixhawk?

I am a bit confused as to how the Piksi's work when set up normally (Both connected to the ground station). If I hook both piksi's to the ground station computer and get RTK Lock will it give me a very precise GPS coordinate, or, will it simply be a very precise relative GPS coordinate?

What I am asking is can I use the piksi's to get a very accurate GPS coordinate for the ground station location. Then, use these Lat Long Alt parameters for when I set up the Piksi's on the Pixhawk?

If not, would the GPS Coordinates from the Ublox on the pixhawk be the next best solution as SPP accuracy is better than most other options?

Thank you,

Jason

[Clive Turvey]

>>If not, would the GPS Coordinates from the Ublox on the pixhawk be the next best solution as SPP accuracy is better than most other options?

I've covered this before. Taking average values reported by an unaided GPS device real time are next to useless.

When you're trying to work in the realm of centimetres and millimetres, you don't tie it to a location you estimate to 50-100cm, or to a device with 3m or 11m CEP95 specs.

If you're operating just within Google Earth imagery, find a feature you can see well, like the cross-hairs created from a parking lot stripping, curbs, fire hydrants, and use the lat/lon of that for your local relative world. Yes there are distortions there, but they are baked in, your results will look good, and be at least as good and arguably better that taking the average NMEA output position over 24 hours. Blair has demonstrated this technique.

>>If I hook both piksi's to the ground station computer and get RTK Lock will it give me a very precise GPS coordinate, or, will it simply be a very precise relative GPS coordinate?

You'll know to within a couple of centimetres where they are with respect to each other. You have to nail one of the end points to something you know, to know anything useful about where the other end is.

>>What I am asking is can I use the piksi's to get a very accurate GPS coordinate for the ground station location.

You might be able too, but it's not my experience. I perhaps need to try with the current firmware. The method you'd want to use is called Post Processing, where you take 20-30 minutes of raw measurements/observations in RINEX format, and then run those in your favorite software against contemporaneous RINEX data from a local GPS CORS. Ideally something within the 5-7 Km range. With a good set of measurements, you can get a convergence below 1cm rms. What does that mean? Basically that 68% of all the solutions found fall within a sphere of 1cm radius. With really good data you can get it into the millimeter domain, but at that point you're into things like the phase centre stability, and physical deflection of the antenna mount. And as others have pointed out the planet and the plates.

There's imagery in this thread that covers survey vs average, and distortion in Google Earth. My location has very good registration, other places it's significantly worse. If you use real lat/lon/alt you need some trim values to address local registration issue, otherwise you'll break people's reality distortion field that if it's on Google it must be right.
https://groups.google.com/forum/#!searchin/swiftnav-discuss/garden/swiftnav-discuss/Lei61Tq15z8/uNSfu3bvqksJ

[Jason Franciosa]

Excellent,

Thank you for the info. That makes perfect sense to align it up with google maps data. Sorry for repeating questions that have been answered. I'm very excited to get these up and running in a UAV, however, I want to ensure I fully understand how it works and it's limitations before going crazy with it or spreading misinformation.

Thanks again,

Jason