I’m looking for advice on procedural animation techniques.
I’m visualizing the results of a sim with lat/lon outputs converting to coordinate data. I want to control these two degrees of freedom of a land vehicle and have the rest be simulated/calculated in real time. The obvious solution if I didn’t have constraints would be to run a basic physics sim but I’m not sure if that’s applicable. The collider of the ground would also be quite detailed and might be a concern as well.
The goal is to:
-Have random terrain [done]
-Keep surface contact at the wheels by calculating height[1dof] and orientation[3dof]
-Have the x and y coordinate of the vehicle origin be governed by an external data source
The way I was considering doing it was to raycast down from the wheels and do the math every frame, either calculating orientation or setting up a linkage model in isaac sim and just transforming the wheels. This strategy seems really expensive/inefficient so I figured I would ask around for advice.
Can I have physX govern the orientation and z translation but not affect the x and y translation of the vehicle?
Hi,
I am not quite sure I am following the problem. Is the problem that you want to constraint the movement along X,Y axis? You could add a joint to the vehicle rigid body and constraint the movement. Eventually there are lock X,Y axis flags on the rigid body.
But I am not quite sure I do understand what you are trying to solve.
My objective is to control movement along the x and y axes without having z axis data. I want to have the z axis to adapt to a mesh surface at the given x and y coordinate.
I could get the height of a mesh at that given point but that wouldn’t rotate my vehicle for me.
I could teleport the vehicle from place to place and use a physics scene to orient the vehicle but I’m looking for smooth (or at least 5-10 updates per second) animation.
we don’t have any built in capability that just does this for you. In general any problem dealing with forcing a physical system with momentum to follow an arbitrary animation is going to be difficult. I think Ales’ suggestion of using a joint to force the car movement in the plane might however be a good starting point for experimentation.