I was recently graphing the relative angular velocity of a link(applying torque via ROS' Wrench service) and decided to change the "max step size" in Gazebo GUI parameters from 0.0001 to 0.48bc148513887beb0a57bc5dc7a5ba8497fb3af3. When I did that, the magnitude of the angular velocities decrease by a factor of ~4 for some reason.
Seems like this a major bug in how the angular velocity and almost every physics sensor is being compute since dt (time step) is not being calculated properly.
Original report (archived issue) by Anonymous.
I was recently graphing the relative angular velocity of a link(applying torque via ROS' Wrench service) and decided to change the "max step size" in Gazebo GUI parameters from 0.0001 to 0.48bc148513887beb0a57bc5dc7a5ba8497fb3af3. When I did that, the magnitude of the angular velocities decrease by a factor of ~4 for some reason.
Seems like this a major bug in how the angular velocity and almost every physics sensor is being compute since dt (time step) is not being calculated properly.