koobas.hobune.stream
How to convert Torque Stabilizer to RigidBody2D?
Can someone help me? I'm trying to translate this TorqueStabilizer concept to work with RigidBody2D. http://wiki.unity3d.com/index.php/TorqueStabilizer
I only vaguely understand Quaternions, so it's difficult to tell how to translate the quaternions used by RigidBody for AngularVelocity and Torque to the floats used by RigidBody2D. Could anyone here offer suggestions?
Thanks - Slapworth
It is a poor algorithm, but at least it's simple.
You could try these changes which preserve the existing structure - I just removed the "magnitude" conversion from the angular velocity, and fixed it to be about Vector3.forward (the Z axis), and made the AddTorque call only pass through the Z component of the calculated torque (the other components ought to be zero anyway):
public class TorqueStabilizer2D : MonoBehaviour {
public float stability = 0.3f;
public float speed = 2.0f;
void FixedUpdate () {
Vector3 predictedUp = Quaternion.AngleAxis(
rigidbody2D.angularVelocity * Mathf.Rad2Deg * stability / speed,
Vector3.forward
) * transform.up;
Vector3 torqueVector = Vector3.Cross(predictedUp, Vector3.up);
rigidbody2D.AddTorque(torqueVector.z * speed);
}
}
All the messing about with vectors is very indirect though, it's essentially the same as this simpler code:
public class TorqueStabilizer2D : MonoBehaviour {
public float stability = 0.3f;
public float speed = 2.0f;
void FixedUpdate () {
float currentAngle = Mathf.Atan2(transform.up.x, transform.up.y);
float predictedAngle = currentAngle - rigidbody2D.angularVelocity * stability / speed;
float torque = Mathf.Sin(predictedAngle);
rigidbody2D.AddTorque(torque * speed);
}
}
Thanks, gfoot!
The simplfied 2D system worked for me, though I did some tweaking.
The issue was that the behavior seemed to be unresponsive, and generated a slow unidirectional torque no matter how I configured the system. I noticed that some of the math was in radians, and some was in degrees, so I converted angularVelocity to radians for consistencty using $$anonymous$$athf.Deg2Rad.
I also jacked up the constants heavily to get the level of stability that I wanted... which may mean that my fix I introduced new errors. Curious what you think.
In any case, I've got the results I was looking for. Thanks so much for the help!
Slapworth
