Find Collision Point of two RigidBody2Ds

This is a READ-ONLY snapshot of the Godot Community Forums when TwistedTwigleg was the owner, which is hosted for archival reasons only. The current Godot Community Forums can be found here.

Some links may point to external websites, including the current Godot Community Forums. These websites are not associated with TwistedTwigleg.

Find Collision Point of two RigidBody2Ds

jgrauman

Hi,

I have two RigidBody2Ds that collide. I am trying to figure out the collision point. I see that Shape2D has a function collide_and_get_contacts, but unfortunately, my RigidBody2Ds both have CollisionPolygon2Ds to define their shape which isn't derived from Shape2D. Does anyone know of a function that would find the point of collision? And better yet, also the collision normal?

Josh

jgrauman

Can I bump this? Surely there has to be a way to figure out where RigidBody2Ds collide? Is there a prefered way of contacting the devs or raising a feature request??

TwistedTwigleg

I meant to reply to this earlier but forgot. Sorry.

You can detect collisions between RigidBody2D nodes using the body_entered signal. That will return the body the RigidBody2D collided with, but not the position or normal.

The easiest way to get collision position is to use the test_move function combined with the body_entered signal. The test_move function, it will return a Physics2DTestMotionResult containing the collision position, normal, and other collision data. If you pass a velocity of zero to the test_move when you detect a collision, it should return the collision info for that collision.

Is there a prefered way of contacting the devs or raising a feature request??

The new way to make a feature request is to make an issue on the Godot proposals repository.

jgrauman

Super helpful! Thanks, it seems to work! Is there any documentation on what the motion, motion_remainer, and collider_velocity Vectors in Physics2DTestMotionResult mean? I'm wondering if they would be useful in calculating the collision energy? Ideally, if I have two KineticBody2Ds that collide, I would love to have both of their velocity's before and after the collision, and then I can calculate how much energy the collision was...

TwistedTwigleg

Great! I'm glad it is working. I only recently learned about the test_move function and its uses, and for certain things (like detection collision positions) it has been super useful.

I do not know 100% for sure what some of the properties are, here is what I think they are for:

motion : The velocity that lead to the impact. I think this should be the same value as the motion/velocity passed in to the test_move function.
motion_remainder : I believe this is the resulting motion/velocity from the collision. That said, I'm not totally sure how it calculates the motion/velocity remainder.
collider_velocity : I think this is the velocity of the other collider that caused the collision.

For calculating the energy of a collision, I think you'd probably need to look at the difference of motion and motion_remainder, where the difference is the energy lost in the collision. Hopefully this helps!

jgrauman

I have come across an unfortunate 'bug' of using this method. test_move does not always find a collision inside of body_entered. I'd say 90% of the time it returns the collision information, but some of the time it just returns false to there being a collision. I've tried messing with the parameters to test_move, but haven't gotten a 100% success rate. I would love to see body_entered itself be able to return a Physics2DTestMotionResult for the collision that caused body_entered to be called...

TwistedTwigleg

Maybe open a proposal to the Godot proposal repository for including a Physics2DTestMotionResult with the body_entered signal? I did a quick search and it looks like it has not been proposed. I could certainly see how it would make the body_entered signal way more useful than it is currently.

Unfortunately, I don't really know of any workarounds for when the test_move function fails. If I stumble across a potential solution though, I'll let you know.