Wild hair idea: "Double VAMOS"

[reece146]

Correct anything here that is incorrect/imprecise/misunderstood.

VAMOS does suspension geometry via a simplified model. We have the pivot points for a "reduced" suspension and once they are defined the wheel follows the arc around that pivot point. As a result there is no wheel camber gain/loss curve that can be mapped to our suspension outside of this fixed geometry.

What if the suspension was modeled "twice" using VAMOS?

As an example for a high performance SLA suspension we model the top A arm and the bottom arm per corner with VAMOS. It is now known where the ends of the hub/upright assembly/balljoints will be relative to the ground.

Since the rest of the goemetry at the upright is fixed (excluding steering angle for the moment) we now know where the wheel is relative to the ground by extrapolating from the ball joints to the traction circle.

Would this be too computationally expensive? Are there better (faster & easier to code) libraries to get the same results? Maybe a simpler apprach would be to map a caster gain/loss curve to each wheel assembly instead and not both with "double VAMOS".

I can post a diagram of what I'm saying if needed - was planning on doing diagramming for the Wiki .car work-up/definition section anyway.

[xTs]

Somebody mentioned that it wouldn't be a bad idea to remove vamos, because vdrift is just using a small part of it...

[thelusiv]

We've almost gotten to the point that the remaining Vamos functionality in VDrift could be replaced and rewritten as VDrift's own simulation engine. This is currently taking a lower priority than fixing bugs, and I'd argue is a lower priority than some new features the game needs.

[xTs]

But VAMOS seems to have itself some bugs. So if you wanna replace vamos, you've got to do the bug hunting twice.

Is there a roadmap anywhere?

[cotharyus]

I agree. If we're bug hunting, lets figure out which ones are caused by Vamos, and either forget about them if we're going to replace Vamos, or just replace Vamos and get on with squashing all the bugs. I'd had to have to describe twice as many bugs as I need to