Hoping some of the hardcore Forerunner users might have some data they can add to this discussion. I searched first, but didn’t find this specific issue discussed.
My Forerunner 301 uses GPS data, not barometric data, to determine the elevation gain and loss during a ride. I do not believe it is accurate, and seems to be significantly over-inflating the gain and loss info. Some of the readings seem to be double or even more what I would have estimated, eg rides where I estimate total gain to be 750 vertical feet my Garmin is saying 1,500 or more.
Has anyone else experienced this, and more importantly, has anybody been able to do some side-by-side comparisons between the Forerunner and a barometric altimeter? I’m hoping there is some sort of a predictable “inaccuracy factor” I can use, such as “always multiply your Garmin results by .65 to get a more accurate approximation of your true elevation gain”.
This is relevant to my current hills training in prep for the MS-150. Appreciate any relevant input. Likewise, if no known factor exists, I’d appreciate any recommends for a cheap barometric unit. I already have a GPS and a full-service cyclometer, so don’t want to spend a bunch more.
I’ve encountered the same problem using an Edge. If you look at the elevation profile you’ll find it’s very noisy with occasional large spikes in both directions thrown in; this all adds up to exaggerate the total elevation.
Some mapping programs will smooth the elevation, taking a rolling average over a certain period of time; I find this gives far better results that match what the map says it should be. SportTracks is one such program, and is what I use to record my rides because it’s a stunning piece of software and is free to boot.
I don’t think it’s as simple as an “inaccuracy factor” though; on some rides it’s been dead on, others twice the value it should be. I presume it all depends on whether your route has a few long climbs or many short ones and how good a position the GPS can read along the way.
These things work out your position by triangulation.
Put very simply, a satellite that is directly overhead can help to triangulate your latitude and longitude, but it cannot help with your altitude.
To work out altitude, the satellite has to be comparatively low on the horizon. All sorts of things (weather, trees, hills, houses) get in the way, and only a slight variation in angle is enough to make a good few metres of difference in recorded altitude.
When I switch mine on, it sometimes tells me I’m below sea level! Goodness knows what it’d tell me if ever I rode my unicycle into the canal.:o
(A friend of mine has a very accurate navigational system on his motorbike that he calls “sat-nag” - she’s sat on the back, nagging.)
My current solution is GPS/bluetooth module and PDA, this seems reasonable whilst riding but goes bananas when stopped at the pub (big height changes & very high speeds >50mph). However using SportsTracks, as mentioned by Phil above, it is pretty quick to remove the quirks.
I think the only solution is to ride an identical route a few times and get some idea of the total variation into a program like SportsTrack. On the site they have done half of the exercise showing multiple repeat trips of a single route, but haven’t mentioned height variation.
I am looking at getting the Forerunner 205 so any feedback on altitude variation with Garmin products will be interesting.
I’m using the 301 and have had good success with it… as long as there are not many tall trees or buildings. As a check of the accuracy in your specific location, try to compare the elevation using google earth. With google earth, try zooming in and point to a location. The elevation will be shown at the bottom. I’ve also heard of people using garmin data to retrace routes in google earth.
here’s some info I got from a friend on email last month
—If you’re into navigation by global positioning system (GPS), British roadie Martyn Davis is your kinda guy. He’s developed a free web service that allows the plotting of routes anywhere in the world using Google maps fed into a GPS device. Davis built his site after buying the Garmin Edge 205 GPS-enabled bike computer and finding its software “disappointing.” First you’ll want to read about his development of the service at http://www.marengo-ltd.com/gps. To see the service itself, click http://www.marengo-ltd.com/map. Even if this technology is beyond you, it’s fun to see your cycling roads (or those you might be visiting) represented by Google maps and satellite photos. When you get to Davis’s site, zoom out, center the map over your general area, then start zooming in. Once you’ve located your home turf, select the “Tools” button and click “Save” to make it your default start point.
Good, varied, interesting responses all…thanks much. I will take a look into SportTracks, but also–given Mikefule’s description of how it works–will likely try to find a good old barometric unit.
Hudson: you’ve heard correctly on using GPS data to re-trace routes on Google Earth. Nathan Hoover turned me on to some cool software from MotionBased which will download all your Garmin data, then allow you to both analyze it and map it against Google Earth. You can also save those files and send them, and when opened, you get the big blue marble, then a gradual zoom in down to the streets you were riding with the route highlighted in red. Very cool. They have both free accounts and paid versions with more functionality. So far, I’ve found the free version sufficient. But since the source data is still from the Garmin device, it does nothing to correct for inaccurate elevation info.
I thought I remembered once upon a time seeing a thread where someone had compared a long climb using Garmin with the same climb using barometric, but my searching couldn’t turn it up again. May have been dreaming it.
You are a human locomotive. If anything your GPS is underestimating your altitude differentials. I’ve ridden with you. You can probably ride up a vertical wall 1000 meters without breaking a sweat.
> Hudson: you’ve heard correctly on using GPS data to re-trace routes on
> Google Earth. Nathan Hoover turned me on to some cool software from
> ‘MotionBased’ (http://www.motionbased.com/) which will download all
> your Garmin data, then allow you to both analyze it and map it against
> Google Earth. You can also save those files and send them, and when
> opened, you get the big blue marble, then a gradual zoom in down to
> the streets you were riding with the route highlighted in red. Very
> cool. They have both free accounts and paid versions with more
> functionality. So far, I’ve found the free version sufficient. But
> since the source data is still from the Garmin device, it does nothing
> to correct for inaccurate elevation info.
I have also created some static maps based on speed, altitude etc, and I
would agree with you that the altitude seems to be one of the most
inaccurate aspects of a GPS system.
Altitude readings require a lock on 4 or more satellites while 2D (latitude and longitude) readings require a lock on only 3 or more satellites. There will be times where you can get a 2D location but the GPS receiver will not be able to calculate an altitude due to too few satellites in view or poor satellite geometry.
The GPS receivers also geared towards giving you a 2D location and the altitude is more secondary. The receiver will give you an accuracy number of say 42 feet. That accuracy number is for 2D location. You can assume that the 3D location accuracy is worse (may be much worse). That accuracy number is also just a guess by the receiver based on satellite geometry and other factors. It is only a guess. The 3D accuracy is going to be worse than your 2D accuracy. The elevation plots from the GPS are going to vary more than the 2D location plots.
The GPS calculations are done by triangulation with the satellites and the time it takes the signals from each satellite to reach your receiver. The accuracy of those calculations depends on the geometry (location) of the satellites in the sky. The wider the triangles are, and the more spread out the triangles are, the more accurate the calculations can be. The satellites in the sky are constantly moving so the geometry is also constantly changing over time.
Best bet for elevation plots from a GPS is to do some post processing on the data to smooth out the elevation readings and get rid of the jaggies. The jaggies in the elevation plot are probably erroneous readings. I believe that SportTracks can automatically smooth out the elevation data (I know you can manually edit the elevation data to smooth it out one data point at a time). There is probably other software that can edit or smooth out elevation data but I don’t know of any off-hand.
Another thing you can do is go to the high points and low points along the ride and sit there with the GPS till it gets a good satellite lock and good geometry and then take a waypoint. That waypoint should give you a more accurate elevation reading than the reading you get while on the move during a ride. Then use that more accurate waypoint elevation to help smooth out the elevation data from during a ride.
A GPS receiver with a barometric altimeter built in will give a smoother and more accurate elevation profile. But even then it can be off based on weather factors (say a front moving through) or through a poor calibration of the barometric sensor at the beginning of the ride.
I have the Garmin e-trex Vista C with the built in altimeter. The altimeter itself works good and appears fairly accurate, especially if it is calibrated to the actual elevation first.
But I sometimes get different cumulative elevation gain/loss information from the GPS display and the graph once I download it to my computer. For example, after a recent ride the GPS said I only had a cumulative elevation gain of 400’ but when I add the total elevations on the chart it is almost double that. The chart was more correct.
Not necessarily. I once spent a surprisingly easy day climbing Tryffan in Snowdonia, only to realise afterwards it was muddy thumb print on the map. I’d thought the view was a bit disappointing.
Actually, it’s exactly the opposite. Since GPS measures propagation delay, a satellite directly above you can’t measure your lat/long very well since as you move along the ground the change in propagation delay is very small because your motion is perpendicular to the satellite, rather than closer or further away from it. Satellites near the horizon are the ones that measure your lat/long because your motion causes you to move closer or further away from [some] of them, which is measurable.
The converse is true for altitude. Satellites near the horizon can’t measure your altitude very well because you’re not moving towards or away from them [very much]. Satellites above you, however can measure your altitude more accurately because you are moving closer or further away from them.
That said, “terrestrial” GPS units put the majority of their horsepower into lat/long (using satellites near the horizon) to make that as accurate as possible because the thinking is, to quote Buckaroo Banzai, “No matter where you go, there you are” – if you’re on the ground, you’re on the ground, and if you get back to the same point on the ground, you’re at the same altitude.
But as the processors get better and better, the manufacturers are devoting more horsepower to altitude calculations.
Sorry to be pedantic but you’re both wrong. You need satellites both high above you and near the horizon to get a good altitude fix. Mike would be right if his receiver could receive satellites both above and below him, but if one accepts the evidence of ones mind that there is ground beneath, and that satellite signals from the opposite side of the world won’t propagate through it, one is stuck with the next best of one above and some to the side.
Steve, thanks for the tip. I hadn’t checked that site since purchasing my unit, and I was a bit behind on updates. I like the new “auto-nag” feature. My unit keeps beeping, and flashing the message “You need to do a 50k ride today!”. I can’t get it to shut up.
When you draw routes on the map, it displays elevation data from the map. I dunno what it does with elevation when you use ‘import from gps’, it may ignore it and use their existing data.
It’s count of total elevation gain thing though seems pretty high for some of my local routes. It seems to count every tiny little bump or something, or maybe it’s broken.
but you’re both wrong. You need satellites both high above you and near the horizon to get a good altitude fix. Mike would be right if his receiver could receive satellites both above and below him, but if one accepts the evidence of ones mind that there is ground beneath, and that satellite signals from the opposite side of the world won’t propagate through it, one is stuck with the next best of one above and some to the side.