GPS Accuracy

[Paul]

I was out on the lake this past Sunday and winds were 15mph gusting to 20mph. On some runs my handheld GPS (Garmen) indicated the speed in my Haven 12.5 at 6.5 mph. Would this be a true speed. I have checked the gps against my odometer and it was on the mark.

[merc412]

GPS doesn't care about wind or current, it only knows position vs time plus or minus 10-15 feet depending on how many sat's it had aquired. The longer your run the more accurate the speed.

[John Bell]

Is it your true speed? Generally speaking, yes. The speed you are reading is an average, albeit over a small enough time/distance interval that the displayed current speed is very close to your instantaneous speed. I find the speed output on mine to be in good agreement with just about every autombile speedometer I've checked it against.

What I don't trust is the maximum speed the unit records. Every now and then it seems to get quite confused and you get an impossible reading. For example, the other day we were doing some speed runs with a friend's jetksi. The first time we came back in, the maximum speed read 283 mph! Now that horrible beast is plenty fast, but not that fast! Resetting the unit again we got a more reasonable 54 mph. When I get 100 mph plus max speeds on a car or boat trip, or 1000 mph plus max speeds on a commercial airliner, I know the device has confused itself. What I don't know is how often it has a maximum speed error to a much lesser degree.

[Ian McColgin]

I recommend putting the averaging time as long as possible, like 120 seconds, to keep abberant readings to a minimum. It's not a speed log for timming, after all.

[Tom Lathrop]

It depends. The older GPS units give 10 meter accuracy since the GOV stopped the dither. This is pretty good if you are moving fairly fast. The slower you go, the greater the inaccuracy can be due to ransom error of samples at as much as 10 meters (33 feet) each sample. When you are moving at only 6 kts, this error can be a large percentage of the speed.

The newer GPS units with WAAS technology use an additional satellite to give correction signals that drop the error to 3 meters (10 feet) max. This makes the reading much more accurate.

If you are moving fast, the percent error is much smaller since the sample position error is the same at all speeds. As was mentioned, taking an average over a longer time will improve the accuracy, but only if the speed remains the same.

[Ian Wright]

Originally posted by John Bell:
the other day we were doing some speed runs with a friend's jetksi.You have a friend who has a jetskii?

For Shame!

IanW.

[John Bell]

Just trying to be open minded, Ian! smile.gif

[Bill Perkins]

I worked out the run times required to insure various accuracies of speed measurement back when the errors were big , but I guess it's still relevant for those of us with old machines .

A statistician could do better but I simply assume opposing errors at each end of the measured run .Garmin states the average error of their machines is currently 15 meters so my assumed error in distance measurement will be 30 meters or .0162NM .

Time = Distance/Speed . Similarly the minimum time of run required to achieve a given level of accuracy in a speed measurement is equal to the assumed error of the distance measurement divided by the allowable error in speed measurement .

For accuracy to the nearest half Knot : .0162NM /.5 KN = 1.9 minutes run time required . Call it 2 minutes .Quarter knot accuracy takes twice as long to acquire .It doesn't matter what speed you run .The percentage of error is larger at lower speeds whatever your measurement devise or level of accuracy.

For accuracy to the nearest tenth knot : .0162NM /.1KN = 9.7 minutes run time required .

[ 06-04-2003, 08:16 PM: Message edited by: Bill Perkins ]

[Smacksman]

These GPS thingys are magic.

Thank you USA for letting us use your satallites smile.gif

[Alan D. Hyde]

You're very welcome.

Our pleasure, Smacksman.

De nada. :D

Alan