Tracking exit data

[dthames 0]

Even before I learned to skydive, I knew tracking would be something of interest to me. If I do a hop and pop and the exit is 4000 or better, you can be sure it will be an exit right into a tracking posture. Something about a tracking exit is just fun for me.

Saturday we had low clouds and hop and pops were about all we could do. I did a couple from a C-182 wearing a long sleeve fleece type shirt and jeans. I would not claim to be a high performance tracker dressed like that.

GPS data shows my vertical speed lower than I had expected. I knew my vertical was slower tracking than normal belly jumps but under 95 MPH without a good suit, I was a bit surprised. I go into a normal boxman right before I wave and pitch, which results in a slight increase in vertical speed right before deployment.

I enjoy looking at the data and trying to understand all of the reasons. I am sharing this for others that might be interested as well.

CTRL + and CTRL - will zoom in and out with Internet Explorer if you can't read the graphs.

Vertical speed in pink, MPH
Horizontal speed in blue, MPH

5,000 exit, pull at 3,000
[inline jump1.jpg]

4,500 exit, pull at 3,000
[inline jump2.jpg]

Instructor quote, “What's weird is that you're older than my dad!”

[platypii 20]

Very cool. I've been working on tracking for my last 2 days at the DZ. Not exactly "exit" data, but since we're nerding out with data, I thought I would share.

These are 3 of those jumps:
- Red: My first jump of the day, doing what I thought was "tracking" but apparently not so much.
- Orange: 5th jump of the day, and hired a coach to help (thanks Patrick!)
- Yellow: First jump of 2nd day, right after reviewing video from previous coach jump.

[inline trackingjumps.png]

I was wearing a normal belly suit for all these jumps. I was using the Recon Flight HUD goggles to get feedback during solo tracks. The last jump got an average glide ratio of 0.7, and sustained fallrate under 90mph (confirmed with N3).

Next weekend trying out my new PF track suit!

BASEline - Wingsuit Flight Computer

[mjosparky 3]

Just for clarification….it seems people have lost the meaning of “hop and pop”. It does not mean a jump from 5,000 feet or even 4,000 feet. It is a jump where as soon as you leave the plane you open your canopy, hence “hop and pop”.

Sparky

My idea of a fair fight is clubbing baby seals

[dthames 0]

mjosparky

Just for clarification….it seems people have lost the meaning of “hop and pop”. It does not mean a jump from 5,000 feet or even 4,000 feet. It is a jump where as soon as you leave the plane you open your canopy, hence “hop and pop”.

Sparky

I stand corrected..... replace "Hop and Pop" with "Low altitude ticket".

Instructor quote, “What's weird is that you're older than my dad!”

[mjosparky 3]

dthames

***Just for clarification….it seems people have lost the meaning of “hop and pop”. It does not mean a jump from 5,000 feet or even 4,000 feet. It is a jump where as soon as you leave the plane you open your canopy, hence “hop and pop”.

Sparky

I stand corrected..... replace "Hop and Pop" with "Low altitude ticket".

My idea of a fair fight is clubbing baby seals

[potatoman 0]

Please post the graphs of the PF track suit. I would like to see how that compares.

You have the right to your opinion, and I have the right to tell you how Fu***** stupid it is.
Davelepka - "This isn't an x-box, or a Chevy truck forum"
Whatever you do, don't listen to ChrisD.

[footballhokie 0]

what did you uses to get this data? i see you said GPS, but what make, model, price... ect?

[FlyingRhenquest 1]

I was quite surprised early on when my N3 told me I'd had an average of 98 mph on a tracking dive shortly after I got the altimeter. I haven't seen one that low since then. My average back then was about 145 mph on a normal belly jump.

What did you use to collect your GPS data? I've tried my cell phone but it doesn't work very well. It very frequently loses points (Often losing half the jump or more) or is off by several thousand feet for one altitude point. See jump 1 for an example, it lost a pretty good chunk of my canopy time there, right after I gained several hundred feet of altitude. Bad GPS is bad.

Someone kindly provided me a flysight GPS trace, and it looks like that works much better. I'm going to buy one as soon as I'm done with the payments on the rest of my gear. You can view both of these files with the free version of google earth. My program doesn't have exit detection yet, though the canopy deployment works surprisingly well. I convert all the coordinates in the jump to ECEF, which are just X/Y/Z coordinates from the center of the earth in meters. My code just checks your location every second and flags the first point that's less than 10 meters from the last point as canopy deployment.

The source to my program to convert android mytracks and flysight CSV files to KML documents like this is available under a BSD-style license (free as long as I'm attributed as the original developer of the components in the application that you use) on Github.

I'm trying to teach myself how to set things on fire with my mind. Hey... is it hot in here?

[Boomerdog 0]

That's why I call them a "low pass" as I'm going to get about 10 -15 seconds of free fall and pull! I have more confidence in a good deploymenet when I'm terminal V.

[dthames 0]

footballhokie

what did you uses to get this data? i see you said GPS, but what make, model, price... ect?

I use a Flysight. They cost about $250. It functions as a recorder and also provides real time feedback for one of three preset flight indicators, horz speed, vert speed, or glide ratio.

The Flysight produces a text file that you can use in Excel. Flysight has a free viewer. Also, Hellis (DZ.com user) has developed an Excel addin that will make graphs in Excel and also put out a Google Earth type KML file. The graphs that I included are from the Google Earth display of that data.

There are some other related threads on this forum and also the wingsuit forum. Search Flysight and you should be able to find them.

Instructor quote, “What's weird is that you're older than my dad!”

[footballhokie 0]

dthames

I use a Flysight. They cost about $250. It functions as a recorder and also provides real time feedback for one of three preset flight indicators, horz speed, vert speed, or glide ratio.

Thanks!

[Hellis 0]

FlyingRhenquest

I convert all the coordinates in the jump to ECEF, which are just X/Y/Z coordinates from the center of the earth in meters.

It's not center of earth.
Your first datapoint show 4576 m, earth radius is aprox. 6371000 m.

The datapoints you have a MSL, Mean Sea Level.
Altitude above average sea level.

Sorry for the correction.

Nice work on the convertiontool.
I just want to warn you about using "absolute" altitudes.
Google earth is not accurate at groudlevel, just looking at the file you uploaded here you can see the problem..

[inline high_flare.jpg]

So on this DZ Google Earth shows groundlevel as too low.
On other DZ´s the groundlevel is too high, and it looks like you landed below ground.

The solution is to use "RelativeToGround" and calculate the groundlevel using the data from the FlySight/or GPS.

I attached one of my files created in my FlySight macro.
Here I calculate the groundlevel and use "RelativeToGround".
You have to save it as *.kml to be able to open it in Google Earth.

[dthames 0]

Quote

So on this DZ Google Earth shows groundlevel as too low.
On other DZ´s the groundlevel is too high, and it looks like you landed below ground.

That has to hurt.

Instructor quote, “What's weird is that you're older than my dad!”

[flyhi 24]

Quote

sustained fallrate under 90mph (confirmed with N3).

Where do you wear your N3 and how much do you trust it? I wear one (wrist) and really like it, but it can differ quite a bit from other devices. Last weekend, we had about an 8 mph delta. Of course in FS, that's only a big deal if you're the low guy.

Aloha

Shit happens. And it usually happens because of physics.

[Hellis 0]

dthames

Quote

So on this DZ Google Earth shows groundlevel as too low. On other DZ´s the groundlevel is too high, and it looks like you landed below ground.

That has to hurt.

look at the landing in the file jump1 posted earlier.

THAT HURTS! :-)

[FlyingRhenquest 1]

Google earth's points are in lat/long, so I have to convert back to that prior to generating the KML file. It's a lot easier to work in ECEF if you need to get distances or velocities since everything's in meters.

Sadly the GPS altitudes are all in MSL. I still have a few features I want to add -- landing/ground detection, exit detection and a jump factory object so you can just leave your GPS running all day and then shovel all the points into the program and have it separate out your jumps. Once I add ground detection, it'll be easy to add an AGL notation to the points. I'll probably leave the MSL one in too, since 12 grand here is a mile higher than 12 grand at sea level.

It's just depressing working with GPS data off my phone (GAAH! It didn't record anything all day! AGAIN!) so I'd kind of put the project on the shelf until I can afford a flysight. I'd had some troubles with "relativetoground" but I forget exactly what they were. I'll play with those tags again and see if I can get a good one. It's easy enough to alter the tags and I could even add a command line flag to allow the user to specify it.

I'm trying to teach myself how to set things on fire with my mind. Hey... is it hot in here?

[FlyingRhenquest 1]

flyhi

Quote

sustained fallrate under 90mph (confirmed with N3).

Where do you wear your N3 and how much do you trust it? I wear one (wrist) and really like it, but it can differ quite a bit from other devices. Last weekend, we had about an 8 mph delta. Of course in FS, that's only a big deal if you're the low guy.

Aloha

I trust my eyes more than my altimeter. Last year I was getting ground rush and my N3 said I was still at 6 grand. I pulled anyway. I believe I noticed an anomaly once I was under canopy, though it wasn't as big of one as I expected. It's the only time I've ever had that happen, but if my eyes and instruments disagree, I'm siding with my eyes.

I've also had it claim a number of times that my 3K speed was 170-200. All I can think for those is it's taking the reading mid-deployment and something about that process is throwing off the reading.

I wear mine between my wrist and index finger. This position usually works better for me than my wrist, since my jumpsuit is a little baggy and can obscure the altimeter if I wear it there. I've had people dock and obscure it where I wear it, but then I just check THEIR altimeter!

I'm trying to teach myself how to set things on fire with my mind. Hey... is it hot in here?

[Hellis 0]

FlyingRhenquest

Google earth's points are in lat/long, so I have to convert back to that prior to generating the KML file. It's a lot easier to work in ECEF if you need to get distances or velocities since everything's in meters.

Sadly the GPS altitudes are all in MSL. I still have a few features I want to add -- landing/ground detection, exit detection and a jump factory object so you can just leave your GPS running all day and then shovel all the points into the program and have it separate out your jumps. Once I add ground detection, it'll be easy to add an AGL notation to the points. I'll probably leave the MSL one in too, since 12 grand here is a mile higher than 12 grand at sea level.

It's just depressing working with GPS data off my phone (GAAH! It didn't record anything all day! AGAIN!) so I'd kind of put the project on the shelf until I can afford a flysight. I'd had some troubles with "relativetoground" but I forget exactly what they were. I'll play with those tags again and see if I can get a good one. It's easy enough to alter the tags and I could even add a command line flag to allow the user to specify it.

So you have the data in lat/lon and then you convert them to this ECEF and then back to lat/lon?
I really don't see the point.

Every time you convert numbers they get less accurate.
Does that not happen with this data to?

[FlyingRhenquest 1]

Hellis

So you have the data in lat/lon and then you convert them to this ECEF and then back to lat/lon?
I really don't see the point.

Every time you convert numbers they get less accurate.
Does that not happen with this data to?

Lat/long is measured in degrees, and the distance one degree represents changes along the spheroid (One degree is less distance at the poles than the equator.) So it's very hard to get a linear distance between two points using the lat/long coordinate system. Since points in ecef are measured in meters, converting to ECEF makes is very easy to measure the distance between two points. Once I do that, I can compute your velocity between any two points in your jump (The points also have a time stamp.) It's also a lot easier to interpolate points in a linear coordinate system like that. That lets me put five samples per second down in the data file, despite only getting one sample per second from my phone's GPS.

The conversion doesn't seem to introduce any inaccuracy. One of my unit tests does the conversion from latlong to ECEF and back to latlong to confirm that the numbers I started with is the same as the numbers I got back from the two conversions.

Funnily enough, I seem to recall that GPS hardware actually calculates coordinates in ECEF and converts them to latlongs for display. I imagine its a lot easier to triangulate your location in a linear coordinate system. Sadly there doesn't seem to be a way to tell the GPS unit to just give you the ECEF coordinates, since this would save me the initial conversion.

I'm trying to teach myself how to set things on fire with my mind. Hey... is it hot in here?

[dthames 0]

FlyingRhenquest

***


So you have the data in lat/lon and then you convert them to this ECEF and then back to lat/lon?
I really don't see the point.

Every time you convert numbers they get less accurate.
Does that not happen with this data to?

Lat/long is measured in degrees, and the distance one degree represents changes along the spheroid (One degree is less distance at the poles than the equator.) So it's very hard to get a linear distance between two points using the lat/long coordinate system. Since points in ecef are measured in meters, converting to ECEF makes is very easy to measure the distance between two points. Once I do that, I can compute your velocity between any two points in your jump (The points also have a time stamp.) It's also a lot easier to interpolate points in a linear coordinate system like that. That lets me put five samples per second down in the data file, despite only getting one sample per second from my phone's GPS.

The conversion doesn't seem to introduce any inaccuracy. One of my unit tests does the conversion from latlong to ECEF and back to latlong to confirm that the numbers I started with is the same as the numbers I got back from the two conversions.

Funnily enough, I seem to recall that GPS hardware actually calculates coordinates in ECEF and converts them to latlongs for display. I imagine its a lot easier to triangulate your location in a linear coordinate system. Sadly there doesn't seem to be a way to tell the GPS unit to just give you the ECEF coordinates, since this would save me the initial conversion.

See attached. You can apply this formula to sequential data points and sum all of the differences to come up with distance over the ground.

Instructor quote, “What's weird is that you're older than my dad!”