GPS

quote:

---

Originally posted by Rick Melanson:
I am a real novice when it comes to GPS gear. The only experience I have had is when a runner tells me that my Certified course is off because of their GPS reading 3.2 instead of 3.1 mi. I have been told that I should buy a GPS so that it would be easy to find "points" on a course that I have measured instead of having to look for nails, paint marks or simply re-measuring locations. I know this duty is usually up to the R.D. but since I measure 95% of the courses here in town I don't mind trying to help the R.D. out. They are usually lost when it comes to locating the mile markers. Is it possible to locate PK nails at the mile markers and then find them easy using the GPS? This might be a worthwhile investement if this is the case. Especially when I have courses at the local state park where there is absoulutely nothing to reference to in terms of mile locations. Would appreciate any suggestions.

Rick Melanson
Birmingham, AL

---

Rick--I have been using a Garmin forerunner GPS for my training. No matter what anyone tells you GPS isn't as accurate. Do not believe people that tell you your course is off--their GPS is more likely less accurate.
On my unit occasionally you might lose signal or go on the same run over the same course and have it come out many different lengths.
If you want to use it to relocate splits I would use the mark location feature which gives lattitude and longitude--that might in fact be quite helpful as a general reference to find the referenced nail/washer. I run over courses I have measured almost every day. Sometimes the splits are within a few feet and other times tenths of a mile..so giving the coordinates might be more specific. Others who use the GPS for referencing should pipe in here--just wanted to give you my support in not doubting the careful work you likely did certifying your courses and defending their accuracy over GPS!--Matt

I began putting GPS locations on my maps several months ago, just for reference. Today I returned to a course I measured in September to see if I could use the GPS to locate my marks. It's a 5K course with nails at the start, finish, and at each mile. I moved the GPS until the longitude and latitude matched what I had recorded when the course was measured. Here's the result, distances are how far away I was from the measured point:
Start: 2' north, 1' east
1 Mile: 12' due west
2 Mile: 15' south, 3' east
3 Mile: 20' north, 3' west
Finish: 2' north, 20' east
What's interesting is that the start and finish are less than 100 feet apart so you would expect the inaccuracy to be about the same at those two points. I did notice that the start is in the clear while there is a big tree about 30 feet east of the finish line. At this time of year there are no leaves on the tree, but it's a large tree with a lot of limbs. At mile 1 there are trees to the north and west, while miles 2 and 3 are essentially in the clear. My GPS unit is an old Garmin GPS 3. The newer units may be more accurate.
My guess is that the GPS could be a help in locating a point especially if you have a pretty good idea where it is. I don't think you could rely on it in a situation where the original marker is missing, especially for the endpoints of a course.

Danny -

Very nice confirmation. It would be interesting to see what other people using other GPS devices would find.

quote:

---

Originally posted by Pete Riegel:
Danny -

Very nice confirmation. It would be interesting to see what other people using other GPS devices would find.

---

We have been experimenting with GPS for course measurement for a little while. As most people in the course measurement field have already discovered, GPS is not suitable for course distance measurement. Mainly because the individual point positioning is too inaccurate but there are also problems with trees and buildings causing loss of signal and degradation of accuracy. There have been comments in this forum about people (runners) contesting the length of a race route that was measured using the calibrated bicycle method because the runner got a different answer using a GPS. Maybe the following is useful to your discussion.

We have tried different GPS methods for measuring course distances. We have conducted tests using differential GPS (or DGPS). DGPS involves using a second stationary GPS receiver (set over a known pre-surveyed point) to help improve the accuracy of results whilst the primary GPS unit is taken over the course route. This method also relies on specialist surveying software and data post-processing. With this, we can achieve precision at the metre level. Without this, individual GPS point positioning is precise at the 10m level. After processing results from a couple of different test sites, DGPS was found too inaccurate when compared to the JO counter. DGPS obviously also suffers from sky visibility problems.

As a step further, we also tried kinematic carrier phase GPS (at the centimetre level of precision). Besides using some very expensive and specialised kit (we're surveyors with equipment and software readily available), we failed to obtain reasonable results even in not particularly harsh GPS environments (e.g. we had clear sky visibility). The errors on each point measurement add up over the length of the route.

The overall result is, GPS is simply not able to compete with the JO counter. There is also the notion that different brands might be more precise, which is not true. There is an upper limitation as to how precise handheld GPS receivers are.

I have a few screen shots to illustrate what we experienced with the DGPS as well the final distances computed using DGPS and JO counters. In the figures, it is possible to see individual point measurements, which is a graphical view of Danny Michael's recent post. I have placed them on a website with a short description of our tests:

http://www.cage.curtin.edu.au/~gordonsj/course

We have also been experimenting with aerial photogrammetry to see if it can match it with the calibrated bicycle method and I have some results of these tests on the same website. This method has been successful at getting the level of accuracies that the JO counter provides and we are continuing to explore it. We hope to have some more results soon.

Regards,
Stuart Gordon
Australia

Dear Stuart,

Very nice piece of work. We can use more material like this.

Thank you, Stuart and Pete! I've been chafing about the GPS issue ever since a couple of local runners (who also happen to run on the same team as I) started using both the Garmin 201 and the Timex Bodylink on runs. They refuse to believe the degree of inaccuracy, nor that the JO is the best of all possible measurement...for now.

Addendum: In spite of posting Stuart's research on the local club's BB and explaining the ramifications of their 'irrational dependence and unwavering confidence' in personally-owned GPS units, some people fail to understand. Of course, these are the same folks who when asked to help ensure courses are measured correctly have other 'more important things' to do...until the day after the race when they decide to complain about measurement. The education continues.

I've been using a Garmin GPS for two years while measuring. I use a combination of notes and GPS readings to assist me in finding nails that indicate split locations on country roads where permanent landmarks may be hundreds of metres apart. I take note of where the nail is in relation to the center or the edge of the road, and to the closest landmark (even if it's hundreds of metres), and record a waypoint. I use the GPS to record the position of an obvious landmark so that I can use the device to measure its approximate distance to a split without using a measuring wheel or a steel tape. I can consistently park the car within 10 metres of a recorded waypoint without much trouble. I then walk with the GPS in hand, which usually brings me to within a metre or two of the nail if there is no tree canopy to obstruct the satellite signals. Knowing where to find the nail in relation to the centre or edge of the road makes it much simpler to find since I can look for it while walking in a straight line.

I also use the track log in conjunction with waypoints and map software to aid me in drawing maps and placing the splits at the right spot on the map.

The most telling evidence that the technology does not meet our needs is that the GPS's position readings change constantly when the device is stationary. GPS certainly lacks the precision to be useful as the only recorded data in locating endpoints of a calibration course or the location of start and finish lines but it's a nice toy to have if you can afford it.

Laurent Lacroix
Winnipeg, Canada
llacroix@mb.sympatico.ca

In my previous post, I gave a link to my website which presents results from our recent tests with GPS for road course measurement. As part of that research, we also performed some tests on the calibrated bicycle method. These tests were important because we were using the JO counter to benchmark the GPS measurements. However, we did not actually know how good the JO counter was. Therefore, along with releasing the results of the GPS tests, we are also making available the results from our tests of the JO counter.

In total, we conducted three experiments involving up to 4 cyclists. The results from these are posted to my website:

http://www.cage.curtin.edu.au/~gordonsj/course

We received a great response regarding our GPS results and we look forward to hearing more comments about the JO counter tests. My email address is at the footer of the webpage. As additional background information to these investigations, there are actually three of us involved: Mike Stewart (Curtin University, Australia), Nigel Penna (University of Newcastle-upon-Tyne, UK) and myself (also at Curtin University). Please bear in mind that we are not experienced road course measurers but we are experienced surveyors. Course measurers know that the calibrated bicycle method is a superior technique for measuring long and complex routes but we think that our results prove this and, with our GPS results, we hope to contribute to the on-going research into course measurement.

We have results from a third investigation where we test aerial photogrammetry for road course measurement. However, these aren't ready at the moment but I hope to upload them as soon as they are available.

Best regards,
Stuart Gordon
Perth, Australia

Message to Pete Riegel from Lou Joline:

Synopsis: Arm swing inflates the distance shown on a Garmin Forerunner 201 GPS.

I measure courses for certification. In the past years I have measured ten marathons and over a hundred shorter courses. After the running of every marathon I receive email from owners of Garmin GPS units complaining that the course was too long according to their Garmin GPS.

This afternoon I went to the Truman High School quarter mile track and ran two miles with my Garmin GPS. I ran as close as I could to the center of lane one, although a competitive runner would hug the line. For the first mile I ran with my right arm held closely to my chest. In other words, I did not let my arm swing. At the end of the mile the Garmin read 1.010 miles. For the second mile I let my arm swing normally. At the end of the second mile the Garmin read 1.037 miles. Over the course of a marathon this difference would accumulate an extra 7/10 mile which is consistent with what is being reported to me in these email complaints. Add to this the extra distance from not running on the "blue line" and the 138 feet that we must add as an error factor and you will see readings of extra distance well in excess of one mile.

So for your next marathon, if you want an accurate measurement, you had better staple the GPS to your head. If you insist on wearing it on your am, use duct tape to strap your arm to your body. Don't be calling the course long if you haven't done one of these two things.

Louis Joline, course measurer.

Reply to Lou from Ben Holmes

Interesting experiment, Lou.

In addition, the sample rate of your unit has a lot to do with accuracy.
Runner's GPS units vary from a 6-second to a 30-second sample rate, depending
upon the model you have. The sample rate is the number of times data is
collected in a given period of time (i.e. a 30-second sample rate means that
data is collected every 30 seconds). Therefore, buildings, trees, turns, and
rise and fall can affect the accuracy of GPS units.

Current GPS technology measurement can actually SHORTEN the distance that a GPS
will measure a run, especially a trail run with a lot of turns and hills.
Because of the sample rate issue, and the possibility of missing a few signal
points, GPS units will have the effect of straightening-out any extremely twisty
and hilly run, therefore shortening the instrument's distance measurement.

An alternative that seems to be more accurate for measuring the distance on
twisty trail runs is accelerometer-based technology.

Funny Story:
I was working the finish line of the St Pat's 4-miler, this year. The course
had a lot of tall buildings and narrow streets, but was a certified distance. A
guy finishing with his GPS complained to me that our course was 0.02 miles
short. The guy looked real fit, but he had just "run" a 4-miler in 58 minutes
and was complaining that the course was SHORT!

I guess the central theme to every GPS owner's complaint may subliminally be:
"I paid over $100 for this thingy on my arm; it has to be more accurate than
someone spending hours measuring and re-measuring a course. After all, it's
space-age technology, man!"

Yeah, and I believe everything that I read on the internet, too.

Ben

Reply from Sue Robinette

How about doing another experiment? Put the Garmin on a pendulum
(preferably a uniformly driven one) and let it keep swinging in place. What
happens? There is no distance covered, in that the pendulum stays in one
place but the Garmin covers distance during each arc.
Sue

Pete, funny you should post that - on LetsRun.com there was an account of someone running a race w/ a GPS taped to his hat - now THAT would eliminate armswing error!

I always use a GPS when measuring to help with generation of an elevation profile for the course. It is mounted to my handlebars so there's not a lot of arm swing. The GPS reading always comes out long. On a marathon measurement, I typically get between 26.4 - 26.6 (and mine cost $300 so it must be more accurate than those $100 arm thingies :-)). Stuart Gordon's research cited above explains why very nicely why it happens due the effects of loss of signal and the GPS then having to recover signal after going under bridges, tree cover, signal bounce off buildings, etc.

This is posted just to get the string back at the top. Look for Stuart Gordon's posting in this string. Lots of good stuff there.

quote:

Originally posted by Rick Melanson:
I am a real novice when it comes to GPS gear. The only experience I have had is when a runner tells me that my Certified course is off because of their GPS reading 3.2 instead of 3.1 mi. I have been told that I should buy a GPS so that it would be easy to find "points" on a course that I have measured instead of having to look for nails, paint marks or simply re-measuring locations. I know this duty is usually up to the R.D. but since I measure 95% of the courses here in town I don't mind trying to help the R.D. out. They are usually lost when it comes to locating the mile markers. Is it possible to locate PK nails at the mile markers and then find them easy using the GPS? This might be a worthwhile investement if this is the case. Especially when I have courses at the local state park where there is absoulutely nothing to reference to in terms of mile locations. Would appreciate any suggestions.

Rick Melanson
Birmingham, AL

A good GPS is nice to have but no hand held GPS will exceed the currently available standard of 2 - 3 meter accuracy. If that is close enough to help you find mile markers, that's great. Just make sure that what you buy is WAAS enabled. You can read a lot about WAAS on the internet.

Dick Blum