Controllab has automated its workflow for collision detection
In steel crane simulators there are a lot of objects that interact through collisions; crane hooks colliding with lifting eyes, ladles colliding with the ground, etc. To simulate the collision physics correctly, a number of steps are taken:
- Define and import the shapes of the objects that can collide.
- Detect the point on these shapes where objects make contact.
- Calculate the impact force of the contact.
- Calculate the resulting object motion as a function of the impact forces.
- Show the resulting motion in 3D.
Three years ago we had to write manual code for each of these steps. This is a lot of work and limited the amount of collision objects that we could handle. Step-by-step we have been working on the automation of this workflow.
This video shows where we are right now: We import CAD files in Blender, define collision shapes and from there everything is automated. Data is exported to 20-sim and Unity, and models are generated. We only have to run a simulation for visual verification.
We are now building crane simulators with hundreds of collision objects running in real time. See Steel Sim VR’s video for an impression.