Wheel with high torque causes slipping with highest friction coefficients
I am trying to model a car in Gazebo. The mass is about 2000kg and the wheels have a diameter of .356 meters.
Based on the max acceleration of the real car 3 m/s^2 it would require a force of 6000 newtons or a total torque on the wheels of 2136 N*M. Split among its two front tires that is a max of 1068 each.
I modeled the car (ford escape) in Gazebo using realistic dimensions and weight, when I apply 760 N*M of torque each (Using AddTorque on a Link with a plugin) it works up until about 3.5 m/s then the wheels begin slipping.
My model is using:
<collision name='front_right_wheel_link_geom'>
<pose>0 0 0 0 0 0</pose>
<geometry>
<cylinder>
<length>0.229</length>
<radius>0.356</radius>
</cylinder>
</geometry>
<surface>
<bounce>
<restitution_coefficient>.2</restitution_coefficient>
</bounce>
<friction>
<ode>
<mu>1</mu>
<mu2>1</mu2>
<fdir1>1 0 0</fdir1>
<slip1> 0</slip1>
<slip2>0</slip2>
</ode>
</friction>
<contact>
<ode/>
</contact>
</surface>
</collision>
As well as placing it on a plane with max friction and no slip.
Is there any other parameters I can try to get this working?
(Using Gazebo 1.5 that ships with groovy)
Asked by FiliusFortunae on 2013-08-01 13:50:57 UTC
Answers
To get better friction solution, try increasing number of inner iteration in the physics engine setting. For example:
...
<physics type="ode">
<ode>
<solver>
<iters>10</iters>
</solver>
</ode>
</physics>
...
Note also gazebo 1.9 does the iterations on friction directions last, resulting in better solution than gazebo 1.5.
Asked by hsu on 2013-08-01 16:44:13 UTC
Comments
Alright, thank you for the reply. I am going to experiment with this. I was wondering what the default is and also what version of sdf does 1.5 use? (I plan on migrating to the new version of gazebo at some point)
Asked by FiliusFortunae on 2013-08-01 20:54:41 UTC
Comments