RTKPLOT does not plot the skyplot (Rinex data from Piksi)

[sahan dandeniyagamage]

Hi,

1. I collected Rinex observation  (L1 freq.) using a single frequency antenna and Piksi receiver.

2. Then I downloaded the Navigation message from IGS center 

3. I tried to get the skyplot using RTKPLOT tool in RTKLIB.

4. It does not show the skyplot for some reason which I cannot understand.

5. I have attached the used Rinex files

Could you please tell me what is wrong in these data.

Thank you.

Sahan

[Clive Turvey]

Have you run RTKPOST against the data before you plot it?
There is 38 minutes of data missing from the file, but the rest of the data is viable.

[sahan dandeniyagamage]

Hi Clive,

Thank you very much for taking your time. It helped me to solve the issue.

Regards,

Sahan.

[awat...@swift-nav.com]

Hi Sahan,

I'm curious:
- Where did you get a Rinex file created using output from your Piksi?
- What were you trying to accomplish with RTKPLOT? (Was it not a tool provided by the Piksi console?)

Thanks for any feedback.

Andrew
Swift Navigation

[Clive Turvey]

Well it looks reasonably apparent that the Rover data was recorded at the Base console via the observation pane, and the purpose would be to divine the quality of the data generated by the Piksi.

[sahan dandeniyagamage]

Hi Andrew,

1. I created Rinex by pressing the record button at Observation pane in Sift console.

2. I Just wanted an overview about data and satellite locations. (for GNSS-Reflectometry)

Hope this helped.

Thank you.

Sahan.

[Andrew Waterman]

Okay, thanks for the info.  It helps to understand what users like you are doing with our products.

[Ragas R.]

Hey, 

Can you make photo of the mane RTKPLOT setting window? like this 

[Clive Turvey]

This the plot I ran

[Ragas R.]

thanks a lot 

[Ragas R.]

Now i have collected data, for 1 hour, how  can i improve position with Differential station? Thank

I will add obs file from my rover and info from Differencial station near my house. For me i got accuracy about 20 m. I think it is not working. 
maybe my settings are bad, or something i dont know. 

https://drive.google.com/folderview?id=0B2UtCF0t3krgMFNRSU9TamhUOUE&usp=sharing

Thanks a lot. 

[Salil Goel]

I am trying to do a similar thing. I am using a CORS station data. The base station we have here is from Leica. So i tried processing the piksi rover data with Leica base station using Leica Geo office software. The software is unable to process this.

So i am using Rtkpost  (from Rtklib) to process the data with the same base station and piksi rover and the accuracy i am getting is in meters (something like of the order of 6-7 meters, i still need to quantify it).

Any ideas on how to go about it?

I thought one of the reasons could be that i was using broadcast ephemeris. So i tried using precise ephemeris (Rapid and Ultra Rapid)  from IGS but I don't see any significant change.

Would love to see if someone has been able to achieve accuracy of the order of cm using post processing with base station.

Thanks!
Salil

[Clive Turvey]

No, it is the nature of the RINEX data generated by the Piksi, you'd have to wait for them to provide more normative output.

We use other receivers to survey the base antenna locations, and plug that data into the Piksi's broadcast settings. In fixed antenna tests, where we know the location of both, and provide the Base Piksi with that information, the Rover Piksi can provide 2cm rms level results against the survey location. So they are demonstrably capable of the specifications set forth, within their own eco-system.

Good RINEX data against a CORS in the 5-7KM range should provide sub cm rms convergence numbers, with 30-60 minute observation data. I can do this with u-Blox NEO-M8N devices, which might run you $25 on a bad day. I've got a plot of JAVAD vs UBLOX offset for receivers sharing a common antenna, I should dig that up.

[Salil Goel]

Clive, what do you mean by good rinex data?

The solution you are proposing where you enter the location of Piksi base and get the location of rover within 2cm works well for RTK. Due to some constraints on my side, I can only use post processed solution and not RTK. Thus i am trying to use Piksi rover and CORS base station. And just to be clear, my rover is not static.

Fortunately for me, we have one CORS station at the university. The distance of the rover from CORS is within 2 km. So it should work in theory.

[dzo...@swift-nav.com]

Salil,

Right now our observation format is unique to Piksi and doesn't follow many of the typical conventions, which means that post-processing fails through many 3rd party tools.

I expect that our next firmware release will produce observations that will work better with 3rd party post-processing tools and play better with other receivers.  Stay tuned as we hope to have something this month!  

Best,

Dennis

[Clive Turvey]

Well I mean clean solid data, data can be noisy, and have a lot of cycle slips or loss of lock. You want to have a lot of continuous data from a lot of satellites, with good overlap, and they also need to have good geometry in the sky. The post processing software needs to be able to identify and fix cycle slips, and resolve the integer ambiguities, and frankly recognize it is dealing with poor data. Now on really clean data I can get results to converge to 3mm rms, on noisier receivers/antennas 5-7mm is typical, at 9mm I know I can do better. At 30-35KM I'd expect to see cm level numbers, because I'm basically fighting the horizon, the curvature of the Earth and a general lack of commonality in the observations.

2KM should be a good distance, I mentioned 5-7KM as the kind of maximum I'd want to deal with, I'm a tad over 7KM (7.1) from my nearest station, they tore out the old choke-ring antennas and replaced them a few years back and replace with some poorer performing GNSS antennas, and pushed the sample interval to 10 seconds. If you get too far away the atmospheric paths become too dissimilar. I have a much better NOAA CORS about 5KM from my office, and my neighbouring county has a network of stations where you are never more that 5-7KM away, as the circles of coverage overlap.

I like static tests as an initial validation, if those don't go well, nothing else will. Dynamic tests are very hard to reproduce accurately. Tethered circles would be the easiest to do, and validate. One could run a second receiver in parallel, but you are still basically dealing with similar levels of accuracy, and each is going to have it's own relative motion due to architecture/clocks/noise.

Not a big fan of RTKLIB, waiting for a new version, had about zero success with it in GPS+GLONASS mode.

I do hope Swift hurries up, this has been going on for far too long.

[seattledrainservice]

How does one decide on the rinex Nav file to download, in the beginning of this thread 16N was chosen, If I run the data in rtk post it works great! I know when I go to the servers for the latest Nav file it is arranged by date but how is the correct file chosen does the 16n in this case a particular satellite for australia?

[Clive Turvey]

The Nav files generated by a local receiver just include the ephemeris for the satellites they can see, and downloaded from the individual satellites navigation data stream (ie sub frames, 1,2 and 3), this data is satellite unique, unlike the almanac (subframe 4 and 5)

Now places like the NOAA FTP server have the BRDC versions which are a consolidation of data from all satellites visible across the entire network.


ftp://geodesy.noaa.gov/cors/rinex/2016/075/brdc0750.16n.gz

The stuff is categorized by GPS Week, GPS Day-Of-Week, Day-Of-Year and Year. The above being for the 75th (075) day of 2016 (16N), the below further decomposed into Week (1882) / Day (2)

ftp://geodesy.noaa.gov/cors/rinex/2016/075/igr18882.sp3.gz

Full day 75 navigation and observation would be the '0' file (0750), where as hourly would be 'A-X' (075B)

[seattledrainservice]

Thanks Clive!

My problem now as it turns out is I am not making a working pos file like you posted at the beginning of this thread, when I take the ops data (and or even mine) it does not produce a pos file with anything but the headers. I think my problem is the settings not shown in rtk plot.

[seattledrainservice]

Hi Clive

You said not a big fan of RTKLIB can you recommend another software to use with the Piksi's and also the ublox C94-M8P-2, I bought the ublox just to use with rtklib. Doesn't have to be a free suggestion either.

Thanks in advance. 

[seattledrainservice]

Thanks Clive for the PM