[Bob Kuczewski]

... it needs to be understood that during certain quick (flying) transitions even a couple second sample rate could miss critical or revealing information.

That's a good observation Scott.

By the way, I was going to suggest tying your GPS unit to a long rope and spinning it over your head at varying distances and speeds. But then I thought about what might happen if it crashed, so I stayed silent. I think your car driving idea was a much better choice.

[wingspan33]

I think I could have done the over head spinning on a rope deal and not damaged my GPS, but laying it on my dash and driving around the block a couple times was much easier!

I'll include an image of the two different test tracks overlaid on each other, below. With the much rougher 10 second sample rate track, it's very clear that the GPS is only using the locations it registers. And in thinking about it, to create "compensating" curves, the GPS would have to make-up a large number of fictitious data points. I'm sure that a GPS could be designed to not give full merit to the fictitious data points - and only "show" the user the ones that were legitimately recorded, but that gets tricky. Does such a GPS figure the track distance by the true data points, or by the curve compensated and adjusted track? And if you happened to know you didn't hike, or fly, or drive where the adjusted track shows your path, . . . you would probably think the GPS unit was defective.

Considering the above, I'll bet that no GPS unit adds curves where it (may otherwise) think curves may have been happening. The goal is to record your (close to precise) location - when you've told the GPS to record it. If the user wants more data points - for finer definition - then most units allow such an adjustment.


Now, my GPSmap 60CSx takes a removable mini SD card (I've got a 4 Gb card in mine) - so its got plenty of space for recording tracks. But the previous 2 GPS units I've used did not have expandable memory. Yet, if I remember correctly, I had them set for a higher sample rate and they never (in any one flight) exceeded their memory capacity. But I'll have to check into that. Perhaps one or the other did not allow the data point sample rate to be adjusted. My point is, however, if your GPS can be set for a 5 or 2 second sample rate (but it defaults to 10 or 20 second sample rate), try it out and see if a higher sample rate creates TOO much data for your GPS's memory capacity. You might be pleasantly surprised.

PS - Clearly the map program I used in the image below didn't have the most accurate local street maps. I wouldn't want readers to think I was driving across peoples' front or back lawns!

[Bob Kuczewski]

Hi Scott,

It's funny to see how much the GPS track conflicts with the map.

With regard to the curve fitting, I believe that's probably happening when it displays the data. I think you're right that it would most likely store the data points themselves as raw data (without any mathematical interpretation). But the curve fitting happens when the data is displayed. You can see it in this portion of your test track:

Scott_Wise_track_logs_test_9-21-12_mod.png (63.86 KiB) Viewed 4322 times

The display is using some kind of curve fitting to draw that little bit of a curve rather than connecting the points in a straight line. I think the length of that interpolated curve depends on how far apart your sampling is. In the original postings of your flight, it only showed the curves, so it was hard to interpret where the actual sampled points were (although the straight line segments give a pretty convincing clue).

Of course, the estimation of velocity is another matter. It's clear from this little picture that the display does do some curve fitting and it represents your ground track as a result of that fit (in some cases). But it's hard to know if the velocity calculation is based on the fitted curve or if they just calculate it based on times between adjacent points. You'd have to do some experiments showing velocity, and you might even have to do the calculations on their actual numbers because their displays might not be clear enough to show what they're really doing.

Fun stuff!!!

[wingspan33]

I'll throw you a curve, Bob. The picture below is the most magnified view you can get in Map Source. While the less magnified version of the same track appears to have curves between data points, the fully magnified view has straight lines between all the data points. I only just discovered this difference.

So, there is something different going on with the display graphic, associated with resolution.

Also note that there are around 8 data points in the curve in fully zoomed in view, but there are only 3 data points in the curve in the more zoomed out view. ??? Still, I think that in either view, each point is associated with one straight line - but in the somewhat zoomed out view the straight line breaks (ends) half way between indicated data points. It could be that the breaks are where the extra data points should lie - but aren't shown due to the display program's resolution "choices".

So, . . . just when we think we've gotten it figured out, . . .

[Bill Cummings]

Here is another curve that a GPS can throw at a pilot.
As I stated earlier I fly with my GPS in the Track Up mode to more easily reconnect with a recently lost thermal. Once I started doing that my thermaling prowess took at least a two fold leap. I was even directing pilots into my thermals with me so that we all benefited.

Track Up did throw me a big curve one flight while soaring the Guadalupe Rim, NM.

I looked at my screen only to find it completely backwards. Way points that were to the back and right of me were showing on the screen ahead and to the left of me. WHAT THE HECK!!????
I even rebooted the GPS thinking it was itself lost or broke.
It turns out that the higher I got on the ridge the stronger laminar air flow was actually blowing me down wind (backwards.) The GPS was doing it’s job and showing my new track backwards at the top of the screen. To make the screen reverse itself I had to pull in a little to penetrate into the wind. My progress then was one mph over the ground into the wind with the GPS screen agreeing once again with the real world.
Talk about pilot confusion in flight.

[Bill Cummings]

741 6/13/2012 4:53:53 PM 1126 ft 12 ft 0:00:02 4 mph 239° true N42 19.523 W77 18.532
742 6/13/2012 4:53:55 PM 1128 ft 5 ft 0:00:02 2 mph 228° true N42 19.522 W77 18.534
743 6/13/2012 4:53:57 PM 1128 ft 4 ft 0:00:02 1.3 mph 226° true N42 19.522 W77 18.535
744 6/13/2012 4:53:59 PM 1128 ft 4 ft 0:00:02 1.3 mph 169° true N42 19.521 W77 18.536
745 6/13/2012 4:54:01 PM 1128 ft 7 ft 0:00:02 2 mph 138° true N42 19.521 W77 18.535

A different scenario:
What if ground elevation is a few feet lower than the at rest time stamps?
Still wondering where the GPS was mounted. (and or came to rest.)
Still assuming you had a ground loop to the left.
Could the 742 data point have been when the glider was still upright and ground looping (left tip on the ground) and the remaining data points with an altitude of 1128 feet being after the nose whack and when the glider came to rest upside down after your body pounded and came to rest on top of the upside down glider? (Vario a few feet above the ground level?)
Data point 742 to 745 has a constant altitude of 1128 yet the coordinates still showing a movement of 5,’ 4,’ 4,’ and 7.’ Is this the deviation that my GPS always shows me when I’m in a fixed position? Or was the wind strong enough to be skittering you across the ground in different directions.
If this is normal deviation? Might all other data points also have an average plus or minus error of five feet?
Set at two seconds might the error encompass a deviation anywhere from 0 to ten feet of your actual location?

[wingspan33]

Bill,

My GPS was mounted on my base tube. However, while I think I did catch my left wing at the last moment before "touch down" (i.e., pound in) the glider did not flip or invert. One witness, a near by highway construction worker who was the first to reach me, told me that as he approached me and my glider, I was shouting to get the glider off of me (I have no memory of this). What happened is that my right down tube broke on impact and the glider, at least partially, flattened down on top of me. That means that my GPS was always on the base tube, which was itself on the ground. The place I alighted was about as level as you could get. No one or two foot rise or falls. I later found the exact location after finding my missing left tip fairing that had come off at some point. There were also medical supply wrappers right where me and my control bar should have been, if the tip fairing had fallen reasonably close to the left wing tip of the glider.

As to varying lat. and long. readings, they are solely derived from received satellite info. But altitude can be gathered by satellite info and/or the barometric altimeter included in my GPS model. I'd bet that it generally defaults to the barometric altimeter - which I am guessing is fairly stable once it's sitting on the ground. (?)

And an interesting side effect of my landing is that more than a couple local pilots now think that the general place where I landed is (always?) DANGEROUS!!! (It's an alternate LZ because the usual one is planted in corn) But, in looking at the data, I think that one, two, or three unusual factors combined to create a rotor and/or turbulence condition that grabbed me and effectively tossed me down. I feel fairly sure that the mishap was a fluke.

But those local pilots who have no idea of, or interest in, this accident analysis are infected with the hebee jebees! They don't want to fly the site until the regular LZ is once again available. The site requires north winds that aren't all that common, but if a good day came along I do think I would fly it again - not being afraid to land near where I came in last time I flew there. But I will admit that I will do my best to manage a top landing. lol

741 6/13/2012 4:53:53 PM 1126 ft 12 ft 0:00:02 4 mph 239° true N42 19.523 W77 18.532
742 6/13/2012 4:53:55 PM 1128 ft 5 ft 0:00:02 2 mph 228° true N42 19.522 W77 18.534
743 6/13/2012 4:53:57 PM 1128 ft 4 ft 0:00:02 1.3 mph 226° true N42 19.522 W77 18.535
744 6/13/2012 4:53:59 PM 1128 ft 4 ft 0:00:02 1.3 mph 169° true N42 19.521 W77 18.536
745 6/13/2012 4:54:01 PM 1128 ft 7 ft 0:00:02 2 mph 138° true N42 19.521 W77 18.535

A different scenario:
What if ground elevation is a few feet lower than the at rest time stamps?
Still wondering where the GPS was mounted. (and or came to rest.)
Still assuming you had a ground loop to the left.
Could the 742 data point have been when the glider was still upright and ground looping (left tip on the ground) and the remaining data points with an altitude of 1128 feet being after the nose whack and when the glider came to rest upside down after your body pounded and came to rest on top of the upside down glider? (Vario a few feet above the ground level?)
Data point 742 to 745 has a constant altitude of 1128 yet the coordinates still showing a movement of 5,’ 4,’ 4,’ and 7.’ Is this the deviation that my GPS always shows me when I’m in a fixed position? Or was the wind strong enough to be skittering you across the ground in different directions.
If this is normal deviation? Might all other data points also have an average plus or minus error of five feet?
Set at two seconds might the error encompass a deviation anywhere from 0 to ten feet of your actual location?