Page 2 of 2

Re: Rollercoaster AI

Posted: 15 Oct 2007, 20:15
by petermcleland
dswanson wrote:Peter,

Thanks for that explanation. I've set up a number of AI flights for formation flying and noticed the same effect on approach where the AI plane takes a dive on finals only to rise up miraculously at the same (slow) speed just before the threshold to then descend to land.

Do you have a "rule of thumb" relationship between IAS and TAS that we could use generically to amend the CFG data you quoted ?

Also, is there any simple way of amending the final approach speed of AI planes ? When I use the same plane (= same FD files) for AI as the one I fly in pursuit, the AI plane always seems to hold final approach speeds that I am incapable of holding with any combination of flaps, power, spoiler and pitch. I end up having to do a "hold" loop to keep a good distance between the two planes. :think:
It would appear that the speed over the hedge is pretty fixed and I don't think you can change that. It is possible to greatly change the approach behaviour...i.e. how much flap is selected and when, what speeds are flown on the approach prior to the final reduction:-

{Reference Speeds}
flaps_up_stall_speed=120.000
full_flaps_stall_speed=90.000
cruise_speed=600
max_indicated_speed=620
max_mach=1.20

Note in the above there are 30 knots between the two stall speeds and the amount of this difference will cause change of behaviour. Also the difference between the full_flaps_stall_speed and the observed "Over the hedge speed" will have considerable effect on the touchdown quality. Pushing up that top stall speed will increase the speed at which take-off lift off occurs and therefore the length of the ground run.

The "Rules" can only be found by trial and error and it takes a long time :roll:

To find the cruise_speed that you should put there...Find what Mach number and altitude the actual aircraft cruises and and then go and fly a fast aircraft in FS9 at that altitude and mach number...Note the IAS and, if there is any instrumentation to show the TAS then note that. If you could not read off the TAS then convert the IAS on an E6-B or similar aviation computer to get the TAS...That TAS is what should be set in "cruise_speed=". At 40,000 feet 200 knots IAS is 400 knots TAS...At 35,000 feet 200 knots IAS is 360 knots TAS...At 30,000 feet 200 knots IAS is 325 knots TAS...at 25,000 feet 200 knots IAS is 300 knots TAS...Those figures should be enough to make a reasonable conversion :)

Re: Rollercoaster AI

Posted: 29 Oct 2007, 17:25
by dswanson
Peter,

Thanks for the help - your tips have been put to good use. By changing the two stall speeds, notably the "flaps-up" one, I've been able to make the AI planes' approach a little faster and I can now follow them in much easier. :)

A bit of research revealed that when at lower altitudes, say <10,000 ft, the difference between IAS and TAS is neglible so I was able to take the approach speeds in IAS from my own landings and program them straight into the AI cfg - very successful. :welldone:

The changed data seems to make no difference when flying the same plane manually so a win all round. I hadn't realised that these data were read for AI purposes - something to bear in mind for other AI conversions. :think:

Re: Rollercoaster AI

Posted: 29 Oct 2007, 18:44
by petermcleland
Degsy,

Glad to have been some help...AI is a very mysterious business, and we never seem to stop finding new solutions to its problems :)