Page 1 of 1

BEPU HoveCraft

Posted: Thu Jul 27, 2017 12:35 pm
by ApppStudio
Hi firstly Ih ave only just started using bepu and have managed to move it in to .netStandard and have found it a pleasure to use. I have one question I am want my object to hover over my terrain. Is the some sort of hover force I can add to the entity to make it hover just above the terrain?

Re: BEPU HoveCraft

Posted: Thu Jul 27, 2017 4:09 pm
by Norbo
The easiest thing to do would be to use something like Space.RayCast (or even the Terrain.RayCast, if you don't care about any hovering over any other objects) to detect where the terrain is relative to some positions on your hovercraft.

Then, you can modify the velocity of the hovercraft to keep it above the ground. The simplest version of this would be to just incrementally add to the entity.LinearVelocity/LinearMomentum- if your hovercraft is spherelike, that would work okay. If you want the hovercraft to rotate to adhere to the surface better, you'd want to instead use something like entity.ApplyImpulse, where you apply the impulse at locations on the hovercraft in response to the per-raycast-detected distance. Like a set of individual 'hover engines'.

Re: BEPU HoveCraft

Posted: Fri Jul 28, 2017 8:26 am
by ApppStudio
Thanks Norbo that a big help thank you and apologies for the awful typing I was on my mobile. IS there any news when version will be in a release candidate state?

Re: BEPU HoveCraft

Posted: Fri Jul 28, 2017 8:25 pm
by Norbo
I'm aiming for about 2 months for the core parts of v2 to be reasonably complete. That means things like a fully implemented collision pair matrix, a useful zoo of constraints, CCD, deactivation, and the required underlying infrastructure.

That first release probably won't include some useful features though- things like the full-featured character controller will require some additional time, and I might not yet have a 'easy mode' API example done.

However, I should note that my timelines are often pretty questionable. I have a habit of pulling features forward that delay the release and things like that.

You can keep an eye on continued development over at https://github.com/RossNordby/scratchpad. I will probably move to a dedicated bepuphysics2 repository in the next few weeks, since it's quite a bit more than a 'solver prototype' at this point.

Re: BEPU HoveCraft

Posted: Sat Jul 29, 2017 11:33 pm
by mcmonkey
Norbo wrote: Fri Jul 28, 2017 8:25 pm I'm aiming for about 2 months for the core parts of v2 to be reasonably complete. That means things like a fully implemented collision pair matrix, a useful zoo of constraints, CCD, deactivation, and the required underlying infrastructure.

That first release probably won't include some useful features though- things like the full-featured character controller will require some additional time, and I might not yet have a 'easy mode' API example done.

However, I should note that my timelines are often pretty questionable. I have a habit of pulling features forward that delay the release and things like that.

You can keep an eye on continued development over at https://github.com/RossNordby/scratchpad. I will probably move to a dedicated bepuphysics2 repository in the next few weeks, since it's quite a bit more than a 'solver prototype' at this point.
WOOO! :D <3

Great to hear it's progressed! I'll definitely be trying that out in 2 months (or whenever it's available), and following along until it's complete enough to move my main projects over :) (Which I imagine will be still a bit far out, as I'll need a reference point for EG voxel implementation and all)

Also @ OP: If you want to make it really cool, you could also apply a counter-force to the object detected in the ray cast, to simulate the engines pressing down on whatever's below (EG another hovercraft would be pushed downward a bit, or the ground might be caused to shake, etc).

Re: BEPU HoveCraft

Posted: Sun Jul 30, 2017 7:44 pm
by Norbo
oh dear now that I've attached a vague near-term date to it, I'm gonna have to suppress the urge to re-rewrite the broadphase and cut five bytes out of each contact manifold constraint and build a fallback solver for nodes of high degree and try alternate position correction modes and build maximum force variants of contact manifold constraints for higher quality fracture and experiment with CPU-side deformable bodies with volume preservation and build a direct solver for difficult constraint graph subsections and contact manifold simultaneous solving and create raytracing kernels for AVX512 and buoyancy simulation for large nonplanar surfaces and

Re: BEPU HoveCraft

Posted: Wed Aug 02, 2017 1:38 pm
by ApppStudio
Thanks Norbo if its anything version 1 it will be amazing