GPS Drift Correction & Overlay Adjustment

[JayDee]

Hi,
We intend to extend our use GPS track mapping, specifically overlaying driven lines of novices and pro's.
I intend to do exploratory laps at the start of each weekend, this should provide reasonably accurate internal and external lap lines.

In the past however we have noticed significant GPS drift. We performed a test by repeatedly tracing out a 10m square over a sustained period. When displaying all of the laps in i2, an individual ‘lap’ was fairly accurate but each subsequent lap would have drifted slightly. The error is soon in the order of several meters.

If performing overlays of a novice with a Pro run (done earlier in the morning) and against track outlines performed the previous day, the resulting data is of debatable use.

Has anyone a quick way to adjust/align the lap overlays?
@ The motec guys , if there is no such current feature, is this likely in the foreseeable? Or even the option to define a datum point for each lap or similar…..

Cheers,
John

[Mazi]

You have to find your local GPS Correction Signal provider first. Usually this signal is received via GSM GPRS Modem with RS 232 output. You just have to connect it to the correction input of your GPS module and you will be inside 1m circle
However... I can't find correction input on Motec GPS modules But it must be somewhere.

[JayDee]

I thought the correction signal ( assuming you mean the WAAS system ) was only avilable in north america?
I heard the European guys were developing a correction system call EGNOS but dont know if this actually happened. I suppose the other question is can the MoTeC system use the correction even if it is available?

[Mazi]

I thought the correction signal ( assuming you mean the WAAS system ) was only avilable in north america?

True

I heard the European guys were developing a correction system call EGNOS but dont know if this actually happened

Is still under development... And will stay as it is for the next 99 years...

But... There is another way!
For another application I'm using differential correction RTCM signal provided by The Surveying and Mapping Authority of the Republic of Slovenia (SIGNAL Network service) as a part of EPN . More details here:
http://www.gu.gov.si/en/
http://www.epncb.oma.be/
If you are in Europe your problem is solved nearly

I suppose the other question is can the MoTeC system use the correction even if it is available?

Hm... Hello MoTeC guys

[JamieA]

At present, we just log and use the lat and long channels as they are, there is no correction available, unless it is done in the unit before calculating lat and long.

This is an areas that we will look into more, but that is it for now.

-Jamie

[DarrenR]

There are GPS systems available that have their own local base station and radio modem for correction of the Lat and Long as Jamie says.
We have not tested these systems, but if what they say is true then this would be a very good (although expencive) solution.
If anyone is willing to test one of these types of systems we would love to see the data!

[toby]

The only way to get accurate drive line analysis in my view is to have a differential correction beacon broadcasting a correction signal in the local area to your track. This is very expensive and equipment can be bulky. However position information can be as accurate as 20mm repeatable over long periods of time. My other operation deals with the Racelogic GPS based equipment which has these features if you are prepared to pay for them. A useful feature in their software is once the track map has been drawn - both inside and outside - it can be "picked up" with the mouse and "dropped" on the drive line. This effectively gives the same result - almost - as having DGPS. Over to you software guys in Melbourne!

[JamieA]

The problem I see with the manual shifting of the race line, is that it really isnt increasing accuracy. You are just guessing where the race line might have been. How much trust can you really place in the race line when you can just put it wherever it suits you?

The only real fix to the problem is to make the GPS more accurate, and diff GPS is what I see as the real genuine fix to making the data more accuate.

As it currently stands, we can use any brand of GPS, and we just log/report the gps position exactly as the GPS reports it.

-Jamie

[toby]

Jamie, I agree with you regarding the positioning of a track over a drive line. It is not the best method for sure but it does give a reasonable view when using data from the same session with overlays. There would be a problem when you overlay data from a different session/day as there would be drift between the 2 sessions. A differential correction beacon is the ultimate way. There would have to be radio contact between the beacon and the vehicle (not allowed in some formula and bulky for bikes) with all the terratorial problems associated with using it. Personally, I feel that if you look at the data carefully with a GPS drive line, a track overlay is of little benefit. But I can understand the requirements of the original post on the forum from JayDee. I think he need the benefits of DGPS.

[Chris]

I would recommend reading up on GPS via Wikipedia, http://en.wikipedia.org/wiki/Global_Positioning_System

Most GPS systems are now being designed to use not only the orbiting satellites but the geostationary satellites which can provide corrections near equal to that of fixed base stations. These do not require a set of data radios, rather the GPS unit receives corrections directly from the WAAS or SBAS satellites. These corrections also exist in Europe, but the GPS unit must be programmed differently for the different PRN#'s.

MoTeC USA sells a 10 Hz or 20 Hz unit that includes DGPS via WAAS. There is currently not a manual or spec sheet online yet, but I will try to get one up next week. I believe MoTeC Europe also carries them. Accuracy is quoted as 0.7 meters per 24hr but I typically see much greater accuracy as long as the satellite count is good. My recommendation is to maintain 8+ satellites. If the satellite count drops below this, you will see drift in the positional data. 6-7 is OK but not perfect like 8-10. 4-5 is simply not enough and should be ignored.

Chris Brown