Question about making something "move towards a point" and calculating speed.

Mamoruanime

@Mamoruanime

Question about making something "move towards a ...

« on: November 13, 2010, 07:19:05 am »

^Not actually me.
Reputation: +9/-0
Offline
Gender:
Posts: 9786

I'm no good with math, so this definitely perplexes me.

Lets say I have an object with a position of 128,256 and I want it to move towards another object with a position of 64x32 with a speed of 4 pixels per frame. How exactly do I determine the amount of x it must travel vs the amount of y it must travel to maintain a speed of 4 pixels per frame (in such a way that it's moving exactly the same speed as if it were just moving along the y axis or just the x axis respectively)?

Sure I could cop out and simply subtract from x and y each frame, but that would look really weird, and it would completely ignore the speed factor.

Insight on this crazy "probably needs sqr's" dilemma would be much appreciated ;p

Logged

Zaeranos

Re: Question about making something "move toward...

« Reply #1 on: November 13, 2010, 07:57:42 am »

Reputation: +9/-0
Offline
Gender:
Posts: 3725

ms = movement speed (4)
dx = difference between the x coordinates
dy = difference between the y coordinates

move_x = ms ( dx / (sqrt( dxdx + dydy ) )
move_y = ms ( dy / (sqrt( dxdx + dydy ) )

That is with the Pythagorean theorem. I don't know if there is any easier way to do it, without the use of sqrt. Maybe using matrices, but I'm not to sure about that.

« Last Edit: November 13, 2010, 08:12:18 am by Niek »

Logged

Xiphirx wat	Re: Question about making something "move toward... « Reply #2 on: November 13, 2010, 08:24:32 am »
Xiphirx Reputation: +0/-0 Offline Gender: Posts: 3007	Quote from: Niek on November 13, 2010, 07:57:42 am ms = movement speed (4) dx = difference between the x coordinates dy = difference between the y coordinates move_x = ms ( dx / (sqrt( dxdx + dydy ) ) move_y = ms ( dy / (sqrt( dxdx + dydy ) ) That is with the Pythagorean theorem. I don't know if there is any easier way to do it, without the use of sqrt. Maybe using matrices, but I'm not to sure about that. That only works for 90 degree angles... x += speedcos(atan2(64,32)); y += speedsin(atan2(64,32)); ? D: Probably wrong, I haven't messed with cos and sin in a while D: Logged
Xiphirx Reputation: +0/-0 Offline Gender: Posts: 3007	For The Swarm (StarCraft II Zerg strategy) I whip my hair back n forth

Mamoruanime @Mamoruanime	Re: Question about making something "move toward... « Reply #3 on: November 13, 2010, 08:27:53 am »
^Not actually me. Reputation: +9/-0 Offline Gender: Posts: 9786	Wait; why would Niek's only work for 90 degree angles? You can independently calculate the x and y coords D: Logged

Zaeranos

Re: Question about making something "move toward...

« Reply #4 on: November 13, 2010, 08:33:27 am »

Reputation: +9/-0
Offline
Gender:
Posts: 3725

Quote from: Xiphirx on November 13, 2010, 08:24:32 am

Quote from: Niek on November 13, 2010, 07:57:42 am
ms = movement speed (4)
dx = difference between the x coordinates
dy = difference between the y coordinates

move_x = ms ( dx / (sqrt( dxdx + dydy ) )
move_y = ms ( dy / (sqrt( dxdx + dydy ) )

That is with the Pythagorean theorem. I don't know if there is any easier way to do it, without the use of sqrt. Maybe using matrices, but I'm not to sure about that.

That only works for 90 degree angles...

x += speed*cos(atan2(64,32));
y += speed*sin(atan2(64,32));

? D: Probably wrong, I haven't messed with cos and sin in a while D:

But if you have only two points and you need a third point to get a triangle. Just pick the x from one point and the y from the other and you have your 90 degrees angle. The 90 angle point is not one of the two given points and pretty much irrelevant.

Mammy it works fine. And you would prefer to use only once sqrt, instead of twice atan2 + cos + sin in your game.

« Last Edit: November 13, 2010, 08:37:05 am by Niek »

Logged

Mamoruanime @Mamoruanime	Re: Question about making something "move toward... « Reply #5 on: November 13, 2010, 08:35:06 am »
^Not actually me. Reputation: +9/-0 Offline Gender: Posts: 9786	That's what I figured; since you're defining x and y independently, it wouldn't make sense for it to not work Logged

MG-Zero	Re: Question about making something "move toward... « Reply #6 on: November 13, 2010, 04:49:46 pm »
Doesn't afraid of anything Reputation: +42/-0 Offline Gender: Posts: 7002	Quote from: Mamoruアニメ on November 13, 2010, 08:27:53 am Wait; why would Niek's only work for 90 degree angles? You can independently calculate the x and y coords D: Pythagorean Theorem only works with triangles with a 90 degree angle in them. The same goes for SohCahToa. Otherwise you need to use law of sines and cosines. But yea, Niek's formula looks pretty good...you'll get a dimensionless value for the amount of distance in both x and y directions and then multiply that by the speed to get individual velocities. Logged
	i love big weenies and i cannot lie

Xiphirx wat	Re: Question about making something "move toward... « Reply #7 on: November 13, 2010, 07:32:31 pm »
Xiphirx Reputation: +0/-0 Offline Gender: Posts: 3007	Oh I see, you're creating a triangle with a right angle every time. I never thought of it that way lol thanks for the info. Logged
Xiphirx Reputation: +0/-0 Offline Gender: Posts: 3007	For The Swarm (StarCraft II Zerg strategy) I whip my hair back n forth

MG-Zero	Re: Question about making something "move toward... « Reply #8 on: November 13, 2010, 09:13:40 pm »
Doesn't afraid of anything Reputation: +42/-0 Offline Gender: Posts: 7002	EDIT: Disregard this, I suck !@#$%. « Last Edit: November 13, 2010, 09:16:40 pm by MG-Zero » Logged
	i love big weenies and i cannot lie

Author Topic: Question about making something "move towards a point" and calculating speed. (Read 2084 times)