[GIS] How is Precise Point Positioning (PPP) exactly done

accuracygpsrtklib

I am looking for a positioning accuracy that is higher than normal consumer-grade GNSS receiver (mostly GPS/GLONASS hybrids with +- 5m precision). I came across RTK with dual receiver setup (base station + rover) which is supposed to do cm-level precision although it has the "disadvantage" while it needs a base station or another source of correction data that one may not be able to supply.

I am interested in getting the most of a single receiver setup. I read Precise Point Positioning (PPP) is able to achieve sub-meter (or maybe even 20cm-level) accuracy. I think PPP needs raw data from sensor which I can do. I will do readings from a moving vehicle in my application so I can not go with stationary "averaging" over time for getting the accuracy. However, I do not need a real-time solution as I will only collect mobile data and evaluate it later so I can do positioning in post-processing.

U-blox also has a page on PPP where I get the impression their solution is just an ordinary raw data chip with a "better firmware" that can do some additional calculations (quote: "… by stabilizing measurements of the distance between GNSS satellites and the receiver (pseudo-ranges) using carrier phase tracking") so it has better precision than ordinary chip.

I read the Wikipedia article but I do not really understand how PPP works. So how exactly is this PPP correction technically done? Where do the corrections come from?

I read RTKLIB can do PPP too. My location is Czech Republic.

Best Answer

PPP relies on post-processing using precise orbit and clock data from the satellites. This is data for the same timespan as your measurements, and is based on logged information from reference stations. If you require on-the-fly positioning I doubt you can use PPP.*

Depending on your location there may be correction services available where you instead of a base station retrieve corrections from reference stations by radio or telephone. For the Czech rep., take a look at CZEPOS

One-Centimeter Accuracy with PPP: The Future of Satellite Navigation from Inside GNSS seem to give a thorough run-through of the method.

*It seems you can get real-time data as well, based on broadcasted data, as announced in GPS World under the title IGS Launches Real-Time Service. This would presumably not be as accurate as post-processed data.

Related Question