Steering (actuators) issue


from 2 day I try to using the Steering actuators , this is really fast and work well

but I found some issue

how move one object dynamic while using actuators Steering?

in this demo there 2 cube which make the same (inverse) path
when come to contact there a infinite collision and both cube not moving more
how move a bit to the right the cube?

I have try to using the linearVelocity , but this seem which not work when run actuator,(not sure 100%)

the best solution which I found for now is using the applyMovement()

there better solution?


pathfinder.blend (75.8 KB)

Hmm. I havent messed around with the steering actuator a whole lot, but there is a chance this may work… Try setting the accelleration and velocity to 0. Then the steering actuator will only steer the object, instead of moving it. Then use a different motion actuator or python to control the movement. I dont know if this will work. It might just keep the object stationary if you set the velocity and accelleration to 0. Try it and tell me if it works, because I was thinking of doing this for the game I’m working on

I had already tried, but if you choose the velocity (of Steering) to 0.0
does not return even rotation.

to avoid the abstacle on the path, I am using now the Force(applyForce() ), instead of applyMovement () which can cause problems with collisions …
(this way allow to avoid the dipendences to actuators motion wich is very cool :slight_smile: )

at moment all seem work well


not… applyForce not work anything well (in some cases the obj can make a biiig jump)
bad workaround

Steering can be using also with velocity 0.0 (i see this now,different to what I have say first ) by the “steeringVec” value ,
but is totally usenless!!
because Steering write a new value on the linearVelocity in ALL CASE (!!!) ,so to move the obj must REWRITE the linearVelocity with another actuators. (motion)

ok , nothing of new , now anyway ,if not wrong this is written in C+ , so this mean which for other 100 year this remain the same because anyone want put the nose in the C+.


How about making them obstacles? With that they should avoid each other.

you mean in the physic panel?(i just see this option but not know if this is for pathfinder)

never tried , and parent it to the object dynamic?

I played a little bit with it.
You need to enable the obstacle check box at the physics tab (another non physics related UI element) for both the obstacle and the steered object.

Then you need to choose an obstacle simulation type in the world tab.

For me that makes no real sense. I think I simply miss the concept.

Yes, I find that there is a little too much intrusiveness of Steering on the physical .(but not only as design)
I mean, the pathfinder should be concerned with the logic, not with the physical.
this is a limit.
seems like a “game” part. closed
I tried to do such “drifting” car: seems impossible.

the solution “would be” very simple.
Steering must only find the nodes, or at least should be an option, to final movement is not his task

anyway thanks for the hints.
with with RVO (ceils) + flags to obstacle, you solve many problems

only 1 hour to make this games!
but no way to make improvement!

if someone want try to make a drifting…try. (press SPACE to start):wink:


GP.blend (67 KB)

I found another issue…

world -> obstacle simulation -> RVO(ceils)

this solve many issue with dynamic obj (because the object become more “elastic” without this the collision of physic engine due often a error and object cannot more follow the path)

but with RVO(ceil) enabled , the RAM continue to increase until the crash .
not only , also when quit the game , the RAM stilll hight , (so , is a “continue accumulation”)
to normalize the ram must close and restart blender.

this can be a issue of my PC anyway (never seen first this kind of error)
or someone can confirm this issue?

PS: to see the RAM in windows O.S. click RMB on the “bar” at the bottom and choose “task manager” or something