Accuracy test

of

Consumer Grade GPS Receivers

by David Doyle 11/08/2003

updated 04/06/2005

Summary:   

One question that comes ups regarding GPS Receivers (GPSR) is how accurate is the measurement.  As with any measurement instrument the only way to measure the accuracy is to compare the measurement against a known standard.  DDS Engineering was kind enough to allow me to use the reference point by their office to check the accuracy of 2 consumer grade GPSRs I have access to.  The result were:

GPSR WAAS

Distance to point

95% of time.

(Meters)

Average Distance

from Waypoint

(Meters)

Magellan

SporTrak Pro

Yes 1.1 - 1.9 1.0
Garmin 12 No 5.8 - 9.9 4.8

It should be noted that the purpose here is not to compare Garmin to Magellan.  These 2 units were chosen because the Magellan has WAAS correction and the Garmin did not.  Also these were the 2 units I had access to.

Reference Point Information
Latitude N error Longitude E error Height H error
36 56' 37.09685" N 0.0131 m 086 26' 00.47600" W 0.0144 m 164.7630 m 0.0323 m

Datum = NAD 1983 (Conus)    Errors are reported using 1.96δ

 

Now a quick note about precision.  At 37 latitude each degree of latitude is 364,100 feet and each degree of longitude is 292,033 feet.  So the following table gives you the precision for the units.

GPS Unit Feet latitude Feet longitude
00.0001 36.4 29.2
00 00.01' 60.7 48.7
00 00' 01" 101.1 81.1
00 00' 00.01 1.01 0.81

So if your GPS unit records coordinates to 0.01 second the distance is 1.01 feet for latitude and 0.81 feet longitude.  Keep in mind this is only for 37 latitude.  To determine the distances for your area click here

TEST RESULTS

I preformed 2 type of test.  The "Total Point Test" was done by connecting the GPSR to a laptop and using the software SA Watch to capture the data from the GPSR.  The SA Watch software collected data from the unit over a period of time and then generated plots showing the error in reference to the known point.

The "Waypoint Test" was done by setting a waypoint after having the GPSR average the points for a given period of time.

I used GeoSky to determine the HDOP and PDOP for the days I did the test and I tried to conduct the test when both were low.  (Note:  GeoSky is no longer supported by thales but you can download Thales Mission Planning Software from Thales Navigation Software.  Once installed, you will need to download the most current Almanac.  The almanac is a file that contains the orbit, clock corrections and atmospheric delay parameters for all of the satellites.  The software will then calculate the PDOP for the area you will be working in and for the date you need.)

The results are as follows:

Total Point Test Results
Unit Date

Distance to point 95% of time.

(meters)

Distance to point 95% of time.

(feet)

Number of measurements

Magellan SporTrak Pro 12/23/03 1.1 3.6 769
01/01/03 1.9 6.2 244

01/10/03

1.6

5.2

342

Garmin 12 01/01/03 5.8 19.0 71
01/10/03 9.9 32.5 311

 Garmin Plot on Jan 10     /     Magellan Plot on Jan 10

 

Waypoint Test
GPSR Date

Time

Averaged

Latt. Long.

Error

(Meters)

Error

(Feet)

Magellan SporTrak Pro

w/ WAAS

03/03/03 1 Min 36 56' 37.0788" 086 26' 00.4812" 0.6 1.9
03/03/03 1 Min 36 56' 37.0788" 086 26' 00.4812" 0.6 1.9
03/04/03 1 Min 36 56' 37.0788" 086 26' 00.4812" 0.6 1.9
03/03/03 2 Min 36 56' 37.0788" 086 26' 00.4200" 1.5 4.9
03/04/03 2 Min 36 56' 37.1400 086 26' 00.4200" 1.9 6.3

Average

1.0 3.4

Garmin 12

wo/ WAAS

03/03/03 1 Min 36 56' 37.1760" 086 26' 00.4560" 2.5 8.2
03/03/03 1 Min 36 56' 37.2480" 086 26' 00.3120" 6.2 20.3
03/04/03 1 Min 36 56' 37.2480 086 26' 00.4200 4.8 15.9
03/03/03 2 Min 36 56' 37.1400" 086 26' 00.3840" 2.6 8.6
03/04/03 2 Min 36 56' 37.3200 086 26'  00.3120 8.0 26.2

Average

4.8

15.8

Other GPSR Accuracy web pages:

GPS WAAS ACCURACY-Garmin GPSMAP 76

GPS ACCURACY MONITOR (Garmin 12XL)

GPS ACCURACY MONITOR (Garmin GPSMAP 76) (no WAAS)

GPS ACCURACY MONITOR (Garmin GPSMAP 76) (with WAAS)

Garmin 12xl Post SA GPS Accuracy Measurements

Sam Wormley's GPS Errors & Estimating Your Receiver's Accuracy

Back to Doyle's GIS web page