I’ve been working on a physics simulation for a 2D game where I have a box subject to gravity, and I’m running into some issues regarding collision response, particularly when the relative velocities are low. The scenario I have is pretty straightforward: I’ve got a box rotated at 30 degrees, sitting on the ground, with gravity pulling it down. When a collision occurs, my impulse-based collision resolution model is supposed to handle it, but it’s behaving kind of oddly.

So here’s what happens: because the box is essentially at rest with very low relative velocity at the point of collision, the impulse generated is super tiny. This causes the box to precariously balance on its edge for what feels like forever, and it doesn’t settle back onto its side until it finally accumulates enough angular velocity. It definitely feels off, like something fundamental is missing from how I’m dealing with the physics.

I’m starting to think that I need to incorporate something that takes external forces like gravity more into account during collisions when we’re dealing with these small relative velocities. I’ve shared my collision impulse calculation code because I suspect there might be a bug or maybe just a missing component in the logic.

For instance, I’m wondering if my calculation of `angular_effect` is enough, or if I should add a function that injects additional impulse based on the gravitational force acting on the box or something similar. It’d be great to get some fresh eyes on this to see if anyone has suggestions for how I can improve the collision response here.

Have you guys tackled similar issues, or do you have any recommendations for modifying the impulse calculation to better align with real-world physics? I’d love to hear how you might approach fixing this so it feels more natural in the simulation. Thanks for any insights!