Hi Fellow Forumites
I have decided to resurrect this thread to post these results as it gives the reader some background into the thinking behind what we have been trying to achieve.
What follows is largely the text from an email message to Lee regarding the testing of Mode S alerts based on the signal strength as discussed earlier in this thread which Exfirepro and myself have been conducting in Scotland using an engineering version of the software provided by Lee some time ago. This software has gone through several changes as we have progressed these trials. Some of these changes are reflected in the current software version as the pre-selectable Mode S alert levels and these are what I am referring to in this email. The settings in the current public release of the software had already undergone quite a lot of testing but as has been commented about in various threads on the forum they were still a work in progress. It has always been recognised that because of the variations in equipment between CAT transponders and GA transponders, the variations of installations and the vagaries of RF propagation that Mode S acquisition could never be an exact science, but in order to make PilotAware as multi-functional as possible it was worth the effort to give this our best shot
I have always stated from the start of my involvement with this project that there is no such thing as a false alert. Sometimes the target may be further away and not an immediate threat but increased vigilance in lookout for a few seconds never killed a pilot yet, but not knowing something is out there could. Read on:-
Hi Lee
The new trigger levels for the Mode S alerts are looking good from my perspective on a flight yesterday. Whilst circumstances plotted against me to achieve the original objective of a flight with Peter (Exfirepro) our separation proved useful in that I picked up a mode S alert for a well know PA28 aircraft that is based the opposite side of the Forth from us.
When I got home I managed to download the Trk file and using “Notepad”, trawled through until I found the start of the alert for G-EVIE and using the time stamp from the $GPRMC message and tracking his recorded flight path in FR24 and my own one, and cross referencing my position from the $GPGGA message to plot the position in Google Earth. I was then able to approximate his position on Google Earth from his track on FR24 using the same time frame and came up with an estimate of our separation distances.
I first saw him on a Green flag in EVFR and then in a few seconds this was Amber. There were some fluctuations between green & Amber before it then became Red.
I hasten to add that I was giving the sky the “big eyeball” trying to find him but failed to get a visual on him. The alert, again after a few fluctuations reverted to Amber and then Green and then disappeared.
My finding from all this were that:-- We were flying a more or less parallel course and he was 1100’ below me hence the reason I probably couldn’t find him visually against the background.
At the initial Amber alert he was appox. 4.5 statute miles behind in my 7 o’clock.
At the initial Red alert he was approx. 2.7 sm in my 9 o’clock which was the nearest we got.
He then proceeded to pull ahead of me on a very slight diverging course.
His speed according to FR24 was about double mine. The whole alert period from first contact to last contact (Green, Amber, Red, Amber, Green) was about 3 minutes.
This was all carried out with the “Short” setting of the Mode S detection range.
As far as I’m concerned if we can replicate this a few more times I would find this level of warning completely adequate. It’s never going to be foolproof as this shows with the fluctuation of signal strength from what was essentially a fairly straight forward overtake manoeuvre which all things being equal should result in a gradual increase in signal received through a peak and then a gradual decrease but as can be determined from the fluctuations we are not operating in an ideal environment
This is the best result for me so far in this testing as with a bit of, (make that a lot of,) faffing and switching between programmes I have managed to take some of the guesswork out of the equation.
Hope this helps.
Alan
