Timeline for Is it possible to estimate road grade while riding?
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 18, 2012 at 20:42 | comment | added | R. Chung | let us continue this discussion in chat | |
| Jul 18, 2012 at 20:34 | comment | added | R. Chung | A virtual elevation module has been built into Golden Cheetah; it's called Aerolab, and can use wind data if its available. As for a wind sensor people are definitely working on them. | |
| Jul 18, 2012 at 19:03 | comment | added | Rider_X | @R.Chung - Perusing your whitepaper/presentation I can see that you have a really bright idea here. The lap constraint added to the calculations is genius. The wind constraint isn't ideal, but I wonder if a more complex model using information from either wind meter (iBike) or an altimeter could overcome this. This would mean a slightly more complex statistical framework, but I think your lap constraint would nicely pull it together. Is there an R package for this yet? | |
| Jul 18, 2012 at 16:38 | comment | added | R. Chung | No smoothing kernel needed. We do need accurate speed and power data. There are other examples in that paper of comparing this method with altimetric barometers on big climbs and against public works road plans like here. | |
| Jul 18, 2012 at 16:29 | comment | added | R. Chung | Right, it was a slight headwind out / tailwind back (slight meaning a bit less than 2mph). If we hadn't known the wind speed we could actually have figured it out by guessing various wind speeds until the green and red profiles matched. But notice that the "inaccuracy" of the green profile is exaggerated because the scale is so small -- the vertical scale is in meters while the horizontal scale is in km, and the largest absolute error over the entire course was only about 40 meters and the absolute error in estimated slope was only about 0.25% | |
| Jul 18, 2012 at 15:28 | comment | added | Rider_X | @amcnabb - I am guessing unfavourable conditions = wind. | |
| Jul 18, 2012 at 15:25 | comment | added | Rider_X | @R.Chung – Crazy number crunching! When you are computing the elevation from the power data are you adding a smoothing kernel? Also would you be able to overlay the altimeter readings from the good and bad day as well as the true profile (e.g. from one the the USGS maps)? Could be quite interesting. My only caveat with this example is that we are talking about small elevation changes (-+ 35 m). Well small compared to many of our climbs which are upwards of 1200m. | |
| Jul 18, 2012 at 14:59 | comment | added | amcnabb | So what were the unfavorable conditions that caused the green trace to be so inaccurate? That seems interesting. | |
| Jul 18, 2012 at 9:22 | history | answered | R. Chung | CC BY-SA 3.0 |