calculate the path of a body

[dieterweb]

Hi Scott,

is there a way to calculate the movement for a body in the coming n seconds? Collisions must not been taken into account.

Thomas

[slembcke]

I'm finishing up a Cocos2D tutorial that does just that:

Code: Select all

   cpBody body = *(cage.body.body);   
   cpPolyShape shape = *((cpPolyShape *)cage.shape.shape);
   shape.shape.body = &body;
   
   cpVect gravity = space.gravity;
   
   int count = 0;
   for(int i=0; i<300; i++){
      cpBodyUpdatePosition(&body, FIXED_TIMESTEP);
      cpBodyUpdateVelocity(&body, gravity, 1.0f, FIXED_TIMESTEP);
      
      if(cpSpaceShapeQuery(space.space, (cpShape *)&shape, NULL, NULL)){
         // draw shape
         break;
      }
      
      if(i%10==0){
         // draw dot
      }
   }


This is more or less the same code that we use in Twilight Golf to plot the path of the ball and it's first collision point.

It copies the body and the shape onto the stack and steps it forward. If you don't want the collision point, you can just leave out the shape and the shape query.

If you are using custom position or velocity functions you can do this:

Code: Select all

body.position_func(&body, dt);
body.velocity_func(&body, gravity, damping, dt);


[dieterweb]

Cool, I will give it a try tomorrow.

Thx!

[bollu]

lol, I wanted to ask the same question too!

Thanks slembcke

[dieterweb]

The code is working as long as I do not use any damping in the space. I added space->damping to the call of cpBodyUpdateVelocity. In this case the calculated path is not the same as the one simulated.

[slembcke]

Oh, I forgot about that. The damping value is pre-calculated in cpSpaceStep() like this:

Code: Select all

   cpFloat damping = cpfpow(space->damping, dt);


That way it is not dependent on the size of the timestep.

[dieterweb]

ok, that fixed the damping problem. Now I have another one

Calculating the path works for simple things now. But I it does not work for some bodies with custom velocity functions. I am calling the function when calculating the future path.

Here is my function I use to simulate a boomerang movement:

Code: Select all

void cpBodyUpdateVelocity_boomerang(cpBody *body, cpVect gravity, cpFloat damping, cpFloat dt)
{
   cpBodyUpdateVelocity(body, gravity, damping, dt);

   ZXEntityBoomerang *boomerang = (ZXEntityBoomerang*)((ShapeData*)body->data)->entity;
   if([boomerang flying])
   {
      cpBodySetAngVel(body, 60 * [boomerang sign]);
      cpBodyActivate(body);
      cpBodyApplyImpulse(body, [boomerang returnVect], cpvzero);
      boomerang.returnVect = cpvrotate([boomerang returnVect], cpvforangle(CC_DEGREES_TO_RADIANS(1.6 * [boomerang sign] * 60.0 * dt)));
   }   
}


The calculated path is not even near the real path.

I think I know the reason: The impulse I add every step does not depend on the timestep.

FIXED

[Charlie_Fulton]

It's always fascinating reading conversations between you guys

Scott is the tutorial you mentioned for cocos2d posted somewhere?