Home Learn thebasics Hardware reviews Software reviews Ship's Locker LEI Tips and tricks Technical issues Adventures with GPS Favorites Canadian Digest

DeLorme’s Earthmate™/GPS PostPro™ Combination

I.    Product Description & Function

 

  

Recently, DeLorme, who brought you Street Atlas USA® 7.0, DeLorme Topo USA™ 2.0, AAA Map'n'Go 5.0, and various other mapping programs, announced a new software selection entitled, GPS PostPro™. The GPS PostPro™ program is a software solution that will enable the user to record positions with DeLorme’s Earthmate GPS receiver, which can be processed later, via an internet connection, to yield much more accurate results than possible with an uncorrected standard GPS receiver. A laptop or Palmtop computer is also required to log the points in the field.

 

 

For a long time now, many posters on the sci.geo.satellite-nav newsgroup have expressed their frustration with Selective Availability (S/A), the intentional scrambling of the civilian signals that makes it accurate only to 100 meters, 95% of the time. While helpful for many applications, the possibility of a 100 meter error in the position of a critical waypoint or trail location may present a serious problem for some users, especially when attempting to accurately map them. To increase the accuracy down to under 10 meters, 95% of the time, two methods are currently available to civilian users: 1) DGPS, and 2) Post-processing.

 

DGPS does its corrections in real time, but to receive the free US Coast Guard correction signals requires a differential beacon receiver ($300+ USD), an antenna (4’ whip or expensive H-field), and a power supply for the beacon receiver. Although the coverage areas are expanding, if you don’t live near coastal areas presently, you may be forced to use one of the commercial sources of DGPS correction signals, which can be very expensive.

 

Post processing of GPS positions has traditionally been available only for very expensive receivers and has the disadvantage of not being available in real time, but for mapping applications, it is excellent. Popular GPS receivers such as the Garmin 12XL, III+, Lowrance GlobalMap 100, Magellan 315, etc. are not capable of post-processing, because they do not record the pseudorange measurements needed. Handheld receivers normally associated with post-processing will run anywhere from $2000+ to $10,000+ USD.

 

The attractive feature of the Earthmate™/GPS PostPro™ combo is that it can be purchased for about $220 USD, assuming you already have the laptop or palmtop computer needed for logging the points. To be fair, the more expensive units are capable of sub-meter accuracy, because in addition to the pseudorange measurements, they also record the carrier-phase data needed to obtain this level of accuracy and also exhibit much lower noise levels in the receiver. The Earthmate/Post Pro combo is advertised to obtain 1-5 meter accuracy.

 

In this review, we will try to accomplish two goals. The first is to describe the Earthmate™/GPS PostPro™ combo features and function and to decide if it is a truly portable solution. The second is to report on some field testing carried out with this combination, so that a potential user may know what to expect in terms of performance and accuracy.

 

The Earthmate™ is a compact rectangular receiver measuring about 4.78" (12.1 cm) X 2.85" (7.2 cm) X 0.77" (2.0 cm) and weighs in at about 189 grams (7 oz.) with batteries and serial cable. It outputs data in the Rockwell Binary format and will work with virtually all of DeLorme's mapping products, and it is the ONLY receiver that will work with GPS PostPro™. It is powered by 4 AAA's which will last about 10 hours (alkaline) according to DeLorme (we got about the same life in our tests), or by an optional 12 Volt cigarette lighter adapter cable available from DeLorme. Another option for laptop computer users is an adapter cable to allow the Earthmate to draw power from the laptop itself. The Earthmate receiver comes with a standard serial connector, which will interface with any serial port on a desktop or laptop computer. To use the Earthmate with a Palm Computing Organizer or Windows CE device, you will have to purchase an adapter cable for $19.95 USD.

 

The GPS PostPro™ software installed on the desktop and Windows CE device that I was using (HP 360LX Palmtop PC) without a hitch. The only application that is actually installed on the Palmtop is the GPS PostPro™ Logger program. You must use this logging program if you intend to post-process the points later. For those of you who have used Solus Pro 1.5, you are probably aware that it also has a logging function that you can use with the Earthmate, but its files can NOT be post-processed.

 

The next step is to hook up the Earthmate receiver to the Palmtop's serial port via the adapter cable, sold by DeLorme. (To connect with a laptop, no adapter cable is needed.)  Start the Logger program on the Palmtop and when it opens, tap the GPS button in the upper left hand corner. Select the Initialize option to start the process of getting the program to recognize the Earthmate receiver. This step must be done EVERY time you start the Logger program. If a laptop is used, then merely launch the GPS PostPro™ program and the Initialization step can be done from there.

 

On the Initialization page, make sure that all boxes are filled out correctly. The first time you initialize, select the state you are in from the State pull down menu. For subsequent initializations, choose the Last Location option for faster lock on times, unless you have moved a great distance. Make sure that you hit the OK button after checking the correctness of all options. The initialization process does not start until you do this! To monitor the progress of the locking on process, you may select the Satellites or Sky View options. The Satellite screen will show the ID #, Elevation, Azimuth, and SNR values for the satellites being tracked. The Sky View option screen will show the satellite positions on a compass grid with circles indicating elevation angles, much like you would find on just about any handheld GPS unit.

 

When enough satellites are found, a GPS Status window will indicate 3D Fix. You are now ready to start logging points. Hit OK to exit the monitoring screen you selected. This will get you back to the main screen. To start logging just tap the Start Logging button. A horizontal bar will indicate how much memory you have left as the points are logged. A 20 minute log file with points collected at 1 second intervals uses anywhere from 135 to 195 KB of memory. Setting the logging interval at a longer time would use up less memory, if this is a problem. When the logging is finished, just tap the Stop Logging button. The program will then prompt you to give a name to the .RAW file you just created and the folder in which you want to store it. Now you are ready to start the post-processing phase.

 

The post-processing portion of the GPS PostPro™ software resides on the desktop computer. Once you have collected the data with the Logger program on the Palmtop, you must download the file (saved in .RAW format) to the desktop computer. If you use a laptop computer, the Logger program and downloading step are not necessary and the data file is saved in RINEX format.

Now it's time to launch the GPS PostPro™ software from the desktop (or laptop). A small vertical tool bar with 4 options will pop up on the screen. The choices are Help, Log GPS, Post Process, and RINEX Tools>>.

 

 

 

bullet

The Help option takes you to a brief but informative help section that should be read before attempting to use the GPS PostPro™ software.

bullet

The Log GPS option is used when logging with a laptop. If a palmtop is being used, a separate Logging program is used.

bullet

The Post Process option launches the Post-Processing Wizard.

bullet

The RINEX Tools>> option allows the user to post-process the files manually, but the Post-Processing Wizard will accomplish the same tasks automatically.

 

 

To Post-process the logged file, you may simply select the Post Process option to launch the Post-Processing Wizard, which does all of the work for you. There are 4 prompts that you must answer. The first is to select the .RAW file that you want to post-process. Next you will be asked to choose which correction source file you want to use. The selections are:

 

bulletCORS (USA) See  http://www.ngs.noaa.gov/CORS/
bulletEUREF (Europe) See http://homepage.oma.be/euref/eurefhome.html 
bulletCDDIS (International) See http://cddisa.gsfc.nasa.gov/cddis_welcome.html 
bulletEnter your own base file

 

Next you will be prompted to select a base station from which to get the correction data. The default is to let the program automatically determine the nearest base station for you.

 

The final step is to select the type of file you want the post-processed data to be stored in. There are 3 DeLorme formats, that will allow you to display your points/track on any of the DeLorme mapping programs. The 4th choice is to save the data in a comma-delimited Standard Text File (.txt). If you plan to do any calculations with the data, such as averaging, or wish to display it on another mapping program such as Fugawi or OziExplorer, then the Standard Text File is the option you want to use. At this point, all that is left to do is to select a name and folder for the post-processed file and hit the FINISH button. The rest is done automatically by the Post-Processing Wizard. For a 20 minute log file (collected at 1 second intervals), it takes just a couple of minutes to complete the post-processing.

 

Depending on the format the post-processed file was saved in, you can now display it on one of DeLorme's mapping programs, or convert it to a file that can be imported into Fugawi or OziExplorer as a track. You may also open the file in any word processor or spreadsheet program to see the data. Averaging can also be done at this time.

Below is a portion of a post-processed file from a stationary point log, that was opened in Microsoft Excel.

Date

Time

Latitude

Longitude

Altitude

Heading

Speed

Status

RMS

1/9/2000

17:37:22

38.189731

-75.407447

-24.449

0

0.27

3

1.7

1/9/2000

17:37:24

38.189733

-75.407448

-23.922

0

0.22

3

1.66

1/9/2000

17:37:25

38.189733

-75.407448

-23.841

0

0.43

3

1.67

1/9/2000

17:37:26

38.189733

-75.407449

-23.737

0

0.22

3

1.65

1/9/2000

17:37:27

38.189733

-75.407449

-23.658

0

0.32

3

1.61

1/9/2000

17:37:29

38.189734

-75.407449

-23.653

0

0.22

3

1.58

1/9/2000

17:37:30

38.189734

-75.40745

-23.727

356.99

0.7

3

1.52

1/9/2000

17:37:31

38.189734

-75.40745

-23.466

0

0.31

3

1.52

1/9/2000

17:37:32

38.189734

-75.407451

-23.546

0

0.09

3

1.47

1/9/2000

17:37:34

38.189734

-75.407452

-23.551

0

0.33

3

1.39

1/9/2000

17:37:35

38.189734

-75.407452

-23.72

0

0.32

3

1.33

1/9/2000

17:37:36

38.189734

-75.407453

-23.66

0

0.21

3

1.29

Notice the STATUS column. The possible values for this field are "3", which indicates good data, "2" indicates questionable data and "1" indicates bad data. The RMS (Root Mean Square) column is an estimate of distance variance in meters for each logged point. We have no idea how the software calculates this for a single point.

II. Accuracy Testing Results and Performance

We performed two types of testing with the Earthmate™/GPS PostPro™ combo; moving track logs and stationary waypoint logs. As mentioned before, data points were recorded using a HP 360LX Palmtop PC. This is a Windows CE device, but Palm Computing organizers will also work with this combination. A palmtop was used, because we wanted to test the portability of the EM/PostPro combination.

Moving Track Logs:

Moving track logs were recorded and post-processed for 2 loop trails. The first trail (see image below) was recorded with the Earthmate receiver on top of the car. The Eastern half of this route was through relatively open roads, with little or no tree cover present. Note the outstanding matching of the track line (blue-green line) with the road below it in this section. For comparison purposes, a Lowrance GlobalMap 100 recorded a track simultaneously (without post-processing). This track is shown with the yellow line. The Western half of the track traversed a heavily wooded stretch of mostly tall pine trees. On this section of the EM/PostPro track (and the GM100 track), you can see a few places where the receiver apparently lost a few satellites and the corrected positions weren't as accurate.

 

 

To explore this potential weakness further, we did a second track on a woods road, that was completely surrounded by a tall pine woods (the only trees we could find this time of year with leaves on them). Once again, we did a simultaneous comparison log with a GM100 and then even a 3rd track was recorded while walking. The three tracks are displayed below.

Notice the straight line sections on the driving EM track (blue-green line), indicating that it did not have sufficient satellites to get good post-processed points. The GM wandered around somewhat (red line), indicating a diminished level of accuracy, but never seemed to loose lock. The loop trail was repeated, this time while walking with the EM velcroed to my hat. (This act earned Ron the GEEK of the WEEK award from Andrew, but it did prove that it was a portable combination.) The results from this track (yellow line) were somewhat better, but still indicated some lapses in accurate positioning.

This might seem to imply that the EM receiver is not as sensitive as the GM100’s, but it may not be as simple as this. While recording the moving track from the car, we noticed that the EM maintained a 3D lock most of the time, with an occasional 2D appearing. Actually, considering the extremely tough cover conditions of this trail, this was a very reasonable performance. The problem appears to be that when the data is post-processed, the marginal reception conditions are exaggerated, especially when moving. Apparently, the more satellites used in the post-processing solution, the better. Several stationary waypoint logs were taken along this trail and the number of bad data points (STATUS = 1) was extremely low (< 10%). If you look at the number of satellites locked on as a criteria, the EM seemed to be quite sensitive in heavy cover conditions and basically equal to the GM 100. The problem seems to appear when post-processing is introduced to a plot recorded while moving under heavy cover.

 

Stationary Waypoint Logs

To determine the absolute accuracy of the Earthmate™/GPS PostPro™ combo under ideal conditions, a series of waypoint logs were done over National Geodetic Survey (NGS) benchmarks. Only benchmarks that had been recently verified by GPS measurements were selected. A series of 15 logs were recorded, varying from 5 to 20 minutes each, over 9 different NGS benchmarks, covering an area of over 1000 square miles.  The testing covered nearly a one month period.  For comparison purposes, a GM 100 was also set to average points simultaneously with the EM. The two units were placed about 1 meter apart, to minimize any possible interference, and were both equidistant from the benchmark. It should be mentioned that all of the benchmarks were positioned in open situations, with almost no obstructions in any direction.

 

A complete table of results from the testing is given in the table below, but let us highlight the important points at this time. One of the 15 waypoint logs had very large errors and was not included in the table. The reasons for this will be discussed later.

 

  1. The errors for the averaged post-processed positions of the EM ranged from 0.36 meters to 2.44 meters, with the average error being 1.22 meters.
  2. The errors for the GM 100 (no post-processing) position averages ranged from 5.18 meters to 22.67 meters, with the average error being 11.7 meters.
  3. Out of 14,644 individual points recorded in the EM logging, 98% of them had errors of 5 meters or less. The greatest single point error was 8.38 meters.
  4. The 2DRMS values (95% confidence level) for the EM ranged from 3.4 meters to 7.5 meters, with an average value of 4.83 meters.
  5. The average lock on time for the EM was 53.6 seconds, while for the GM 100 averaged 34.2. Many of these were cold starts (over 30 minutes since being turned on).
  6. For the 14 logging sessions that are reported in the table below, which consisted of 14.644 individual points for the EM, EVERY point recorded had a STATUS reading of 3.

 

Benchmark Accuracy Testing Results

Test Number

Test 8

Test 9

Test 10

Test 13

Test 14

Test 15

Test 16

Test 17

Test 19

Test 21

Test 22

Test 23

Test 24

Test 25

Average

Type of Test

5 min.

10 min.

10 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

 

Earthmate Error of averaged position (in meters)

2.31

0.43

1.9

0.78

2.25

0.87

0.45

1.1

0.36

2.44

0.74

1.06

0.42

1.97

1.22

Earthmate RMS error (in meters)

2.44

2.02

2.45

2.23

3.33

1.72

1.78

2.23

1.85

3.06

3.79

2.22

2.42

2.63

2.44

Earthmate 95% Error (in meters)

4.84

4.01

4.86

4.41

6.59

3.4

3.53

4.42

3.66

6.05

7.5

4.4

4.8

5.21

4.83

Smallest EM error (in meters)

0.73

0.24

0.035

0.026

0.67

0.021

0.18

0.1

0.026

0.033

0.35

0.076

0.19

0.57

0.23

Largest EM error (in meters)

3.46

4.95

8.38

4.34

5.48

3.2

4.6

6.56

5.02

5.14

7.74

4.2

4.81

4.78

5.19

EM lock on time (in seconds)

60

75

51

37

41

48

54

80

61

39

42

50

55

57

53.57

GM100 Error of averaged positon (in meters)

8.68

15.32

6.21

20.79

5.18

5.9

5.89

10.03

16.83

6.58

19.02

14.37

22.67

6.28

11.7

GM100 lock on time (in seconds)

44

44

26

22

48

48

20

25

41

42

23

22

49

25

34.21

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Number of Points with Error > 5 meters (EM)

0

0

5

0

72

0

0

20

1

13

185

0

0

0

Total  

Number of Points taken (EM)

278

597

592

1184

1091

1179

1247

1212

1212

1260

1184

1173

1244

1191

Total   14644

 

Errors > 5 meters 2.02%

 

The 15th logging file was not reported in the above table. That test was also conducted over a NGS benchmark in open canopy conditions. Right from the start, the EM had difficulties locking on. This was the last site to be tested, so we had 14 other previous experiences with the EM. We had not seen anything like this before or since. The satellite screen would indicate 3 satellites being tracked, then quickly jump up to 8 satellites and then just as quickly jump back to 3 satellites. Eventually, after nearly 3 minutes of this, the EM did indicate a 3D lock, but it seemed tenuous at best. The GM 100 was in its usual position, 1 meter away, and had no problems at all at this site. We decided to log points for the normal 20 minutes, as we had done for most of the other benchmarks. The results showed a great deal of STATUS = 1 points and the accuracy reflected this.

We returned to the site two times and got similar results. The cause of the poor lock on appeared to be a Cell Phone tower about 200 meters away. We moved around in the general area of the benchmark and still got similar results. To test the cell phone tower interference theory, we then drove about 3 blocks away and turned on the EM again. This time it acquired and locked on to the satellites in its normal manner, taking less than a minute to do so.

Our theory seems to have been validated, but it didn't explain why the GM100 receiver had no problems with interference in the same area. A couple of theories came to mind to explain the difference. One involves the difference in antenna angles and mask angles selected by the two receivers. Both units were laid flat on the ground, but the GM100 antenna is mounted on an angle, while the EM patch antenna is flat. The second theory might involve the difference in sensitivities of the two receivers. At any rate, the EM did appear to be adversely affected by the cell phone tower when in close proximity.

We did not extend the testing to determine if all cell phone towers have this affect on the EM receiver and we did not try to quantify the distances involved other than to say that 200 meters will certainly cause problems and at a half a mile away, it certainly won't cause interference. Somewhere in between those two extremes, there is a minimum safe distance. Our recommendation is to be careful if using the Earthmate in close proximity to a cell phone tower. For use in areas off the beaten path, as this review was intended to address, we don't think this problem would ever present itself.

 

III. EM/PostPro Altitude Accuracy Summary

As reported above, a series of 14 waypoint logs were done over National Geodetic Survey (NGS) benchmarks. In this section of the review, we will analyze the altitude accuracy of the EM/PostPro combination. In the chart below, there is a complete set of altitude data as well as horizontal data for comparison:

Benchmark Altitude Accuracy Testing Results

Test Number

Test 8

Test 9

Test 10

Test 13

Test 14

Test 15

Test 16

Test 17

Test 19

Test 21

Test 22

Test 23

Test 24

Test 25

Average

Type of Test

5 min.

10 min.

10 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

20 min.

 

Horizontal Errors

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Earthmate average horizontal error (in meters)

2.31

0.43

1.9

0.78

2.25

0.87

0.45

1.1

0.36

2.44

0.74

1.06

0.42

1.97

1.22

Earthmate horizontal RMS error (in meters)

2.44

2.02

2.45

2.23

3.33

1.72

1.78

2.23

1.85

3.06

3.79

2.22

2.42

2.63

2.44

Earthmate horizontal 95% Error (in meters)

4.84

4.01

4.86

4.41

6.59

3.4

3.53

4.42

3.66

6.05

7.5

4.4

4.8

5.21

4.83

Vertical Errors

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EM average altitude error (in meters)

1.78

2.84

6.42

0.11

3.13

1.15

0.41

0.53

3.09

0.28

1.07

1.22

2.5

3.67

2.01

EM altitude RMS error (in meters)

2.52

3.74

7.82

4.52

6.95

2.43

4.81

5.22

7.24

4.54

5.41

6.38

4.87

5.72

5.16

EM 95% altitude Error (in meters)

5

7.41

15.48

8.95

13.76

4.81

9.52

10.34

14.34

8.99

10.71

12.63

9.65

11.32

10.21

 

Here are some of the important points to consider in the altitude error data. The average of the altitude errors for the 14 stationary points was only 2.01 meters, ranging from a low of 0.11 meters to a high of 6.42 meters. The RMS errors for individual points, however, weren't nearly as good for the altitude measurements. The average 95% error was 10.21 meters for altitude, but one benchmark point ranged all the way out to 15.48 meters.

The rule of thumb for altitude error is that it will tend to be about 1.5 times the horizontal error for a given measurement. Our measurements resulted in a factor that tended to be closer to 2 times the horizontal error. It should be pointed out, however, that even if you take the data for our most inaccurate benchmark (15.48 meters for a 95% error), that's still very good, when compared to standard non-corrected GPS measurements, which could easily be off by 150 meters in altitude. With just a few minutes of averaging, you can expect this error to drop dramatically. Remember, of the AVERAGED altitudes for the 14 benchmarks, the greatest error encountered was still only 6.42 meters! That's about 21 feet.

 

IV. Here is a summary of our recommendations based on our conclusions dealing with the testing of the Earthmate™/GPS PostPro™ combination.

 

  1. For moving tracks recorded in open canopy situations, the results are excellent. Errors seemed to be in the 10 meter or less range for this application.
  2. For moving tracks recorded with tree cover overhead, the results are mixed. If the cover is extremely heavy, post processing may not yield great accuracy. Your mileage may vary. If the overhead cover is not extremely thick, then decent results should be obtained, though not always in the 10 meter or less accuracy range.
  3. For data recorded while stationary with open canopy, the results were outstanding for such an inexpensive solution. The 2DRMS accuracy (95% error) was under 5 meters for a single reading and if averaging is done, for even periods of 5 minutes, an accuracy of 2 meters can be expected and even better for 20 minute averages.
  4. For data recorded while stationary and under tree cover, the accuracy wasn't quite as good, but if care is taken to discard the STATUS = 1 points, then decent results should be obtained. (No NGS GPS-surveyed benchmarks were located under tree cover for absolute accuracy testing under these conditions.)
  5. When using the EM/PostPro combo for altitude measurements, averaging of the data is highly recommended for maximum accuracy.  If this is not possible, then maximum vertical errors for single measurements in the 15-16 meter range are possible.  With even 5 minutes worth of averaging, this figure should be cut by more than half.  Twenty minute averages, if possible, should cut the error in half again.

 

For further details, you can check out the Earthmate™/GPS PostPro™ combination at http://www.delorme.com/gpspostpro/ .

 

Ron and Andrew

the

GPS Nuts

 

Home Learn thebasics Hardware reviews Software reviews Ship's Locker LEI Tips and tricks Technical issues Adventures with GPS Favorites Canadian Digest