Can't seem to get any collisions

[JC Denton]

I have made a project based closely on the HMMWV using largely the same parameters. But when simulating, the vehicle's chassis falls through the floor, and the wheels remain in place. I tried PAC02 tire and Rigid Mesh tire, as well as Rigid tire without mesh, along with a Rigid Terrain but nothing seems to collide. Setting the steering input also doesn't seem to change the orientation of the wheels. Making gravity 0 or positive also has no effect, chassis still falls through floor.

Inside ChPac02Tire, it indicates each tire is in contact with the surface, m_data.in_contact = true. The chassis uses a box Primitive for collision, and it appears to be setup and created the physics colliders for it. Using NSC System with default settings. The positions all appear correct, with the vehicle and its wheels originating slightly above the terrain. I initialize my systems in the following order:
Vehicle -> Chassis->Steering->Axles->Driveline->Powertrain->Wheels->Tires->RigidTerrain->Driver

I update using
VehicleDriver->Synchronize(Time);
Landscape->Synchronize(Time);
PhysVehicle->Synchronize(Time,  DriverInputs, *Landscape->GetTerrain());                

VehicleDriver->Advance(PhysicsTimeStep);
Landscape->Advance(PhysicsTimeStep);
PhysVehicle->Advance(PhysicsTimeStep);

Clearly something is very amiss, and feels like my systems are not communicating with each other although inspecting their components doesn't reveal anything out of the ordinary. I'm out of ideas at this point. Is there something I might be missing? Or maybe steps I can take to dig deeper?

Thanks for your time

[Radu Serban]

Maybe you could provide a snapshot from the run-time visualization to clearly show what the issue is?  If I understand it correctly, your axles appear not to be properly attached to the chassis? Do you model this vehicle using a set of concrete C++ classes (as done in the Chrono::Vehicle models library) or through a set of JSON files?

 

I’m afraid this would be difficult for us to figure out unless you can share your model files.

 

--Radu

--
You received this message because you are subscribed to the Google Groups "ProjectChrono" group.
To unsubscribe from this group and stop receiving emails from it, send an email to projectchron...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/projectchrono/c5e3a0b3-2cdb-4bbc-9ec5-4b180a90b62cn%40googlegroups.com.

[JC Denton]

Hi Radu,

thanks for the reply!

I modeled it as a concrete class based on ChWheeledVehicle, with the same specs as the HMMWV, but with my own meshes. My meshes are scaled and oriented the same way as the HMMWV wheels, tires and collision mesh, although even when
trying to use non mesh collision types like primitives I still get the same results. Perhaps I missed a parameter, I will re-examine my values.

"If I understand it correctly, your axles appear not to be properly attached to the chassis?"

Thats what it feels like, but in debugging and stepping through the code I can confirm that the vehicle's wheels and tires are initialized and the vehicle knows about them since it calls its wheel->m_tire->Advance(step); logic on each of them inside ChWheeledVehicle

I tried to simplify my simulation yesterday, so I actually got rid of the vehicle entirely and only added a basic simulation, copied from demo_IRR_collisionNSC.cpp

So just a few primitive objects colliding added using the AddFallingItems and AddContainer, the NSC system,and same solver settings, and none of my positions seem to update. The  primitives remain stuck at their initial position, but I am getting nearly 500 physics iterations per frame. The positions the objects are created at are queried every frame on the graphics side and updates them, they are indeed at the correct positions, but those positions never change.

For the record, the demo_IRR_collisionNSC.cpp  app runs fine on my machine.


Images attached for the vehicle simulation showing the vehicle's wheels staying stationary, but the chassis falling through the floor

[Radu Serban]

While I have some suspicions of what may be wrong, this is all throwing darts in the dark.  Unless you can provide a complete piece of code that reproduces the issues you see, I don’t know how I can help.  Since both the original HMMWV demos and demo_IRR_collisionNSC work fine for you, this must be something amiss in your modified codes that could be easily tracked down with a debugger.

To view this discussion on the web visit https://groups.google.com/d/msgid/projectchrono/e7011cb8-1dff-4fd3-b679-47725cc097a4n%40googlegroups.com.

[JC Denton]

Okay I figured it out. Even when I simplified my sim to just simple colliders, no vehicle, the vehicle did in fact still exist in the world even though I wasn't using it. When I removed the vehicle entirely, the simple colliders starting moving as they should and all is well. One major symptom was that if I ran more then 2 time steps, the solver would quickly accumulate memory
until it ran out of memory and crashed.

I should be able to figure it out from here

Thanks!