Welcome to Netrider ... Connecting Riders!

Interested in talking motorbikes with a terrific community of riders?
Signup (it's quick and free) to join the discussions and access the full suite of tools and information that Netrider has to offer.

A GPS question for the tech heads.

Discussion in 'The Pub' started by incitatus, Mar 26, 2007.

  1. Ok, let me explain. I was flying on Saturday surveying a very large recently extinguished area of burnt bush for re-ignition points. I have two GPS instruments in the aircraft, one Garmin and one Lowrance. Both use an identical map, the Jeppesen Aviation Pacific Region Chart. Both GPS instruments were set to record 'trails' at the same intervals, and these were downloaded to a PC for analysis. Using a tool called 'GPSBabel' both files were converted to the same format and overlaid on a chart. Now here is the funny bit, they didn't match exactly, they varied by as much as 1/2 nautical mile. Both recorded at the same time, in the same aircraft, on the same chart data......why the difference? :?

  2. I'd be a little suss about your conversion software. Maybe it has rounding issues? Can you plot both sets of raw data on the same graph and compare?

    Are the inconsistencies laterally from the centreline of your course or could it be a timimg offset?
  3. Yes, that is pretty much where I am too, but the error is not consistent though. The actual waypoints match perfectly after conversion by the way, its just the trails that vary.

    The inconsistencies seem to be mainly the radius of turns. Straight sections are pretty consistent between the two plots, but the greater the radius of the turn the greater the inconsistency.
  4. Sounds like rounding to me too. The software for each device may store the data in different numbers of digital bytes and therefore loose some accuracy. 8 bits = 1/255, 9 bits = 1/1023 etc.
    D'oh edit 9 bit = 1/511 , 10 bit = 1/1023
  5. OK, that sounds like a response time difference between the two GPS units. Sounds like one simply responds to changes quicker than the other. This could be caused by the antenna response or the processor speed of the GPS.

    But that's only a guess from an ex-meteorologist come road nazi!

    Yes Kishy, I know you hate me! :cool:
  6. Good point, and one I should be able to check....

    EDIT -

    Could you expand on that a bit? I can see how response time might affect measurement of velocity, but at a given position surely the position data should be the same?
  7. As I said, I'm no expert. My understanding is as follows:

    You fly your aircraft, your GPS plots your position at a preset frequency, say every second. But the position it plots is not the position at the time of plotting, but the position a certain time lag ago. I suspect your GPS may give a current time stamp to this position that is already a little old, especially if it's a GPS designed for walkers or maritime use and not for the speeds involved in aviation.

    Compare this with the output from a different GPS, one that may be designed for aviation, or simply with later software and then compare positions with the same time stamps you are comparing positions taken at slightly different times. This will obviously become more apparent at faster rates of change of your locations, ie during rapid acceleration or during turns.

    I suspect some GPS software may also use software to fill in datatpoints (interpolate) in between actual positions to make your track look smoother on the screen. If it interpolates during fast turns when it has a slow response you may well get interpolated points some way off your track before the next calculated position draws the plot back on track.

    The next issue would then be one of aliasing, where the sampling frequency of your datalogger may radically increase the apparent errors in your data.

    As I said, I'm quite rusty on this stuff, haven't looked at data analysis for many years...
  8. These are both panel mount, $10,000+ aviation instruments rated to in excess of 500 Knots. You could be correct, but I think rounding and the conversion software is a more likely culprit.
  9. Yep, you're probably right!

  10. Also, most portable GPS units "simplify" their recorded tracks in order to save space onboard the unit.

    The easiest example that springs to mind would be... simplifying a circle (an infinite number of straight lines to describe the curve) into a simple polygon, like an octagon... only 8 lines, but still the general "gist".