GPS Telemetry

[MattS]

In the spirit of assigning mathematical odds to everything up to and including death (hint: the odds are 1 in 1, but they increase as you get older), I recently embarked on a small project (it was raining and everybody was in meetings) to see if it is possible to mathematically predict a good road.

 

For the purposes of this exercise, good means multiple corners and the resulting speed changes on entry and exit; while bad means straight lines and constant speeds -- or straight lines and high speeds.

 

What I'm really looking for are those corners that have a higher-than-average average speed through the corner. Since the data I'm using to predict is the telemetry of my own ride, I can assume that any sections of road that are run at lower-than-average rates are done so because (a) of the threat of cross-traffic or (b) in-town speed limits ... in both cases these would be places to be avoided.

 

So I took an aggregate of three days' worth of data, about 200 miles of local explorations. About 3000 rows of this:

 

LAT		LON		ALT	LEG TIME	SPD	HDG	ALTDIF	SPDDIF	TURN
36.00497	-87.02734	589	00:00:07	60	165	7	3	11
36.00353	-87.02635	592	00:00:07	59	151	3	1	14
36.00223	-87.02478	597	00:00:08	57	136	5	2	15
36.00099	-87.02227	598	00:00:10	59	121	1	2	15

 

What I added to the "stock telemetry" is three difference fields toward the end, the magnitude change in altitude, speed and heading. Ultimately, I made the executive decision to show only above-average speeds (43 being the average) of 45mph and higher.

 

Next I discounted any erratic changes of speed, allowing a value up to ±15mph ... the "smooth" factor. Finally I looked at only turns of 15° or more ... the "fun" factor.

 

Wound up with a data set of about 500 points. Not sure about the presentation, but the net-net is shown here in four stages, 15°+, 22°+, 30°+ and 40°+ turns, all with an average speed of 45mph or greater. Not surprisingly, the head of the Natchez Trace is loaded with good spots, as are parts of (but not nearly all off) western Williamson County.

 

 

Riding through these places, it's not easy to remember where all of the good places are, and it's a hassle to leave the bars for a MARK, so it remains to be seen if this is a useful tool or an incredible waste of time that should instead be devoted to boosting the GDP.

[johnlt]

Matt, you obviously don't have enought to do!

[Kinsley]

Interesting RouteMan

 

I just got a 2720, so this is new to me.

I suppose you are looking at the track logs and downloading them (excel?). It seems my track log capacity gets filled up very fast, are the track points made by time or mileage? If they're adjustable, what do you set yours at?

 

So, is the goal to link the highest "fun" factors together for the ulimate ride

[MattS]

I suppose you are looking at the track logs and downloading them (excel?). It seems my track log capacity gets filled up very fast, are the track points made by time or mileage? If they're adjustable, what do you set yours at?

The 2720 has an itty bitty little track log that can't be customized ... 2000 points for about 300 miles highway, 180 country. I'm using the 376C, which gives you 10000 points plus space for 15 saved logs plus full customization (on time, distance or resolution). I am in telemetry heaven at the moment.