KA72 required for GPSTC

sailquik said..






Thanks Mikey, you bring up some good points that have been covered before in these forums, but are well worth reiterating and expanding on. This is probably not the ideal thread for it, but this is where you brough it up.

Wearing 2 or more GPS devices is sensible and encouraged. It is specifically mentioned in the rules: gpsteamchallenge.com.au/pages/rules

"Using multiple GPS'sGPS devices sometimes can fail, or users can sometimes make errors in use that can result in lost data, or no recorded data on a device, so users are encouraged to use multiple GPS's if they wish to have backup redundancy. In such cases, users should select the complete set of data from only one of the devices to post from the device with the lowest +/- numbers, unless the technical panel gives an individual case approval."

There are a number of reasons including: One or more of them might loose satellite, lock or signal, and not have a complete record. (this happens - ask the guys from Budgie/Lake Mac/Wello etc)

Sometime GPS's dont work as we expect them to. Sometimes a GPS will just die or run flat. 90% of the time is is a user setting or wearing error, but it happens. It good to have a back up

In a very few cases, some sailors are trying for a GPS record where an average of 2 GPS's are required.

All approved gps's have been tested endlessly in side by side situations. We have proved that they will all give the same result within the reported error margin 99% + of the time. Therefore, for results within the error margin, every GPS can be said to be no more or less accurate than any other it is worn with. That said, it is also stated in the rules that the GPS with the lowest error should be posted from and all results should be posted from one GPS only. If in doubt, consult your team captain and/or the GPSTC.

'Fast' and 'Slow' Devices: I have heard some sailors say that one of their devices always gives a higher result. This is not as simple as it may look at a casual glance. One must also look at the way the gps was worn, the error margins reported, the number of satellites used, and have a look at quite a large sample. In all the hundreds of side by side testing sessions I have done with multiple GPS's, I have yet to see any GPS that consistently give a higher result when these factors are taken into account. If you think you have one, perhaps it is faulty and need to be properly tested and eliminated from use if it is faulty.

One other factor in this is the Hz rate (1Hz or 5Hz). A higher Hz rate (readings per second) will almost always find a slightly better result over the short time spans of 2 seconds and especially 10 seconds if the run is a typical accelleration curve from low to high speed and back down again. This is because it can find the optimum speed 2 or 10 second segment with finer resolution. It is still completely 'real'. But the difference is usually measured in 0.1 of a knot or less. It is, in this case, sometimes an advantage to use a 5Hz or 10Hz device, and the difference is both real, and quite small.

For the Alpha category, we have a situation where the technology is still not good enough to be up to 'world record' type standards as we must rely on the positional data (trackpoint) for calculating the 50m proximity circle (Doppler is used for the speed). Since we can't rely on much better than 3-6m accuracy for that parameter, results between GPS's may vary more if the return path is right at the 50m limits (as it is on many of the fastest Alphas). All we can do is accept it for what it is and give the sailors the benefit of the doubt. If we ever get cm accurate locational data (e.g. RTK), this may improve considerably.

Software analysis programs: All the approved programs are well tested and their authors have cooperated on the specifications of the various categories so they all give pretty much identical results. Any differences you find are usually due to different filters or the way the program handles errors or missed data points. RealSpeed may sometimes produce weird results with the new 5 and 10Hz devices, as it has not been updated for quite some time to account for some new developments. If you like using this program, check with me that you have the best version for the devices you use it with, and check your results against other programs if in doubt.

In the 1 hr category, some programs handle missing points and stops, or low speeds differently, but for most one HR results, especially higher speed results, there is no significant difference.

In the Distance calculation some of the same factors apply. Even without making the effort to set up the program to the same base parameters, the results for distance are usually within a small, and usually insignificant margin. If they are not, then something may be wrong with you program settings or your session file, and you should seek a second opinion from your team captain or the GPSTC.

Different programs are often used to find or resolve apparent errors, or for other validation situations. They are extremely useful for testing and research and they can sometimes highlight things in different ways or provide data in a more useful way. Some programs are a lot easier for doing certain things, like cutting off your drive home, a common issue (RealSpeed/GPS-Speedreader) or combining split files. or averaging results from two GPS (GPS-Results). GPSAR-Pro and RealSpeed have the facility to produce a complete tabulation of the entire session including columns for all the different types of data that a GPS device outputs. GPSAR can also run multiple tracks in 'replay' mode for you can visually compare yourself with others on the water near you. The newest sftware program, GPS-SpeedReader, can produce an excellent Google Earth trackmap Graphic. There are many more examples.

But the bottom line is that they will all produce identical, or almost identical results if set up correctly and running a fault free file. If you find a difference that makes a material difference in your results, you have probably found a data problem and need to seek expert assistance.

The KA-72 website program is an excellent service and has clearly made GPS session analysis and posting very much more accesable and easier for a large number of sailors. It is designed for this purpose and also to work with almost any GPS device file. It does an extremely good job if this, and this is why so many sailors use it and love it. I think Dylan would be the first to agree that this also nessasarily enforces limits in some other ways that dedicated analysis progams do not have.

To summerise, there is no one single analysis program that is perfect, or can be all things to all people. This is why the rules of the GPSTC are written as they are. We trust that the upcoming changes to the GPSTC website will help us to alleviate almost all the posting and session validation issues in a much simpler way, without tying the GPSTC to one 'unicorn' session analysis program. In the updated system, all session files will either have to be public on KA-72 or uploaded to a GPSTC database where they can easily be checked if nessasary. No more need for the GPSTC to ask for files and the often misinterpreted implications associated with that.

Oh, and Mikey, to brige that 0.1Kt gap, you just need to sail a tiny bit faster mate!

mikey100 said..
Having just spent time sailing with a large group of speed-sailors from around the country, I was bemused at the end of each day to see some running their tracks from their 3 or 4 worn gps devices through multiple programs in order to get the best result in the speed category that they were targeting. I was also amazed to see the wide variety of results that could be obtained from the same session. It appears that I may be letting myself down by using just 'one device and KA72'!
I also learned that there are 'fast' and 'slow' GPS devices, so I am still hoping to be able to buy one of the 'fast' ones if I see one come up for sale. Now, although not having any idea about coding or programming etc, I offer a possible opinion/solution to help alleviate some of the issues that have been voiced in this, and many other forums...
If GPSTC had its own specific, analysis program (similar to KA72) which ALL tracks had to be submitted through, then everyone's data would be run through the same program/filters. It could then directly fill in the data on the sailor's page. Also, all tracks could then be easily available for GPSTC committee review. This would solve many issues, wouldn't it? But again I admit... I have no idea about the logistics of this idea... so I patiently await the bullets!
On another note. My friend beat me by .010kts over '2sec' the other day- this equates to 1cm (yes 10mm) over the "max 2 sec" so, I was wondering if anyone could help me work out how I could bridge that gap. I don't like him beating me by so much.

Boston! said..




Nope. I don't save other peoples files. (it was last winter@Budgie)
Significant? More than half a knot.

mikey100 said..
Or put gps on top of mast... and get catapulted at max speed.

sailquik said..
Stretchy. On the shoulder facing forward is probably not too bad, but facing back towards the water will probably introduce a lot more mutipath refected signal and will decrease accuracy as a result.

It probably worth mentioning for others who are interested that wearing a GPS in the back pocket of your neoprene vest is definitely a problem. The vest material gets wet and water is a major blocker of radio signals. Not only that, the body blocks most of the direct sat signal.

Neoprene front pocket, or inside the wetsuit is also a big problem, but probably not as bad as in the back pocket.

An open mesh pocket on the front chest, or upper front shoulder of your vest is probably not too bad.

The first signs you are havng problems in high SDOP or Doppler Error values.
The last sign is that your gps misses large sections of you session through loss of satellite signal, especially when sat visibility/availability is low.