search

google-deepmind / dm_robotics

Libraries, tools and tasks created and used at DeepMind Robotics.
Apache License 2.0
332 stars 32 forks source link

How to set joint position for position actuators #24

Open edwardjjj opened 1 month ago

edwardjjj commented 1 month ago

Hi, I'm trying to adapt my robot that utilizes position actuators to the RGB_stacking environment. I implemented the python class for the robot based on the sawyer robot implementation, but it keeps failing with the following error: WARNING:absl:Unable to solve the inverse kinematics for ref_pose: Pose(position=[0.54648661 0.12579726 0.4226723 ], quaternion=[ 6.03993497e-17 9.86396237e-01 -1.64385105e-01 1.00656846e-17])

The following is how I set the joint angles:

def set_joint_angles(self, physics: mjcf.Physics, joint_angles: np.ndarray) -> None:
        physics_joints = models_utils.binding(physics, self._joints)
        physics_actuators = models_utils.binding(physics, self._actuators)
        physics_joints.qpos[:] = joint_angles
        physics_joints.qvel[:] = np.zeros_like(joint_angles)
        physics.forward()

Can anyone help me locate the problem? Thank you very much!