RE: [Algorithms] Car engine behaviour

[Madoc E. <tm...@ti...>]

Hey guys,

Thanks for all the replies (except for Tom's 'cos he's just 
bragging about his cars - another Brit boy racer :)). Damn, I've always 
wanted a skyline.

So, most of the replies talk about advanced 
considerations that I would likely experiment with once I've got the 
basics in place. For the time being my problem is more basic. This 
isn't to say your replies were not useful, they have clarified a number 
of things and are great help. I'm at the prototyping stage for a game 
that still has a lot of development ahead of it. It is not a racing 
game, and as such I don't believe it will be subject to quite so much 
scrutiny but we want stats like HP and so on that reflect in the 
behaviour of the vehicle so that someone who knows his stuff won't say 
"this is c**p".

I don't know if my original post did a very good job 
of explaining what my current setup is. No one has commented but I'm 
still inclined to think that my basic wheel-engine feedback setup is 
good. From what I can tell, it's pretty much exactly what Jon described 
without the RPM-dependant torque (which Tom seems to suggest that it 
could be mostly negligeable). 

The unfortunate result of it however is 
that high gears are not really much worse at accellerating from a stand-
still than low gears, and in fact the feed-back from the wheels acts 
more as break limiting top-speed rather than accelleration. I can see 
that this might be a result of drag at higher velocities resulting in a 
greater shift in velocity at each integration step.

Out of my 
ignorance, I'm thinking that this could be the lack of consideration of 
the vehicle's mass and momentum. Something along the lines of: 
accellerating a mass from 4 to 5 kph takes far more torque than 
accellerating it from 40 to 41 (bare with me for the simplication). But 
this isn't coming into play anywhere. When calculating wheel-engine 
feedback, I only consider the difference in velocity between what the 
engine revs "want" and what the actual wheel velocity is, and this 
difference tends to be greater at higher velocities.

So, to narrow my 
most immediate problem down: How should I describe the fact that higher 
gears will have a harder time at accellerating at lower velocities 
while they fair pretty well at higher velocities? I get the idea that 
this is simple dynamics and I'm just being dumb by not just working it 
out but I don't feel confident with basing it on my own assumptions.


I'll have too look at some dyno charts and see if I can generate some 
suitable curves or get away with some constants. I won't be simulating 
any real-world models and the vehicles will be subject to a lot of 
(probably unpredicted) variation.

As for traction, I'm thinking the 
problem is similar to static vs dynamic friction which so far I've only 
confronted in terms of hard thresholding and contact forces which I 
guess isn't very suitable for this kind of simulation. From my 
experience with wheel spins and hard braking it seems almost right but 
hard thresholding is usually evil...


Thanks again,
Madoc