by Alant » Sun Sep 10, 2017 1:13 am
P.S
This is a brilliant start to automated GCA.
GCA looks simple, but it relies on two human operators - the controller and the pilot,
The GCA system being developed simulates the controller.
The pilot does what he decides with the instructions. Some pilot are better than others.
The aircraft instruments (in particular the compass) may not indicate the same as the controller´s radar display.
Cross winds add a further complication, but the ground controller should be aware of this.and be able to compensate.
The GCA controller must be able to recognise that when he requests a given heading, and that the localiser track is not corrected as , he expects, then he should assume that his commands gave been followed and therefore he should change his heading (track) datum accordingly.
The same logic applies to the descent/glideslope commands.
The logic for commands is therefore:
Start with best estimate, based upon wind, magnetic variation and expected airspeed.
If the aircraft response is not as expected, modify the commands , using new estimates, or just differences from the last command. i.e If the last track command was turn port/left, heading 167 did not correct a drift to the right, issue the next command turn port/left heading 166.
If this was not sufficient to show that the error is showing signs of being corrected, the next instruction may be turn port/left, heading 165/164/163 etc. Cross winds and compass errors will both result in this confusion.
All of this is relatively simple for a human controller, but our controller must rely on the artificial intelligence that he is programmed with.
Alan