Add OperationSpaceController to docs and tests and implement corresponding action/action_cfg classes
#913
Conversation
ozhanozen
requested review from
jsmith-bdai,
Dhoeller19,
kellyguo11 and
Mayankm96
as code owners
source/extensions/omni.isaac.lab/omni/isaac/lab/envs/mdp/actions/task_space_actions.py
Outdated
Show resolved
Hide resolved
source/extensions/omni.isaac.lab/omni/isaac/lab/envs/mdp/actions/task_space_actions.py
Outdated
Show resolved
Hide resolved
|
@Dhoeller19, I have created this PR to replace #879 but haven't worked much on it yet. It is far from complete at the moment. I am currently working on it and will ping once it reaches a state worth reviewing. Sorry for the confusion. |
|
Converted to a draft util ready for full review. @ozhanozen i will continue to work with you on this branch. Thanks for doing this! |
|
@jtigue-bdai, I have finished the following:
From my side, the PR is ready to be merged in terms of features, and I will be happy to integrate your feedback/reviews. Once it is merged, I will create other PRs for the other potential features as discussed. |
…this is more correct
There was a problem hiding this comment.
@OOmotuyi this looks great. Just a few docstring/formatting stuff. If you are ready it seems like you can pull it out of draft and make it ready for review from the wider team.
source/extensions/omni.isaac.lab/omni/isaac/lab/envs/mdp/actions/actions_cfg.py
Outdated
Show resolved
Hide resolved
| self._task_frame_transformer.data.target_quat_w[:, 0, :], | ||
| ) | ||
|
|
||
| def _preprocess_actions(self, actions: torch.Tensor): |
There was a problem hiding this comment.
Please add google style docstrings for arguments
There was a problem hiding this comment.
def _preprocess_actions(self, actions: torch.Tensor):
"""Pre-processes the raw actions for operational space control.
Args:
actions (torch.Tensor): The raw actions for operational space control. It is a tensor of
shape (num_envs, action_dim).
"""
| return rigid_child_prim_path | ||
|
|
||
| def _resolve_command_indexes(self): | ||
| """Resolves the indexes for the various command elements within the command tensor.""" |
There was a problem hiding this comment.
please add google style docstrings for the exceptions raised
There was a problem hiding this comment.
def _resolve_command_indexes(self):
"""Resolves the indexes for the various command elements within the command tensor.
Raises:
ValueError: If any command index is left unresolved.
"""
| """ | ||
|
|
||
| def _first_RigidObject_child_path(self): | ||
| """Finds the first RigidObject child under the articulation asset.""" |
There was a problem hiding this comment.
please add google style docstrings for the exceptions raised
There was a problem hiding this comment.
def _first_RigidObject_child_path(self):
"""Finds the first RigidObject child under the articulation asset.
Raises:
ValueError: If no child RigidObject is found under the articulation asset.
Returns:
str: The path to the first RigidObject child under the articulation asset.
"""
| def _run_op_space_controller( | ||
| self, | ||
| robot: Articulation, | ||
| osc: OperationalSpaceController, | ||
| ee_frame_name: str, | ||
| arm_joint_names: list[str], | ||
| target_set: torch.tensor, | ||
| ): |
There was a problem hiding this comment.
please add docstrings for this function
There was a problem hiding this comment.
"""Run the operational space controller with the given parameters.
Args:
robot (Articulation): The robot to control.
osc (OperationalSpaceController): The operational space controller.
ee_frame_name (str): The name of the end-effector frame.
arm_joint_names (list[str]): The names of the arm joints.
target_set (torch.tensor): The target set to track.
"""
| def _update_target( | ||
| self, | ||
| osc: OperationalSpaceController, | ||
| root_pose_w: torch.tensor, | ||
| ee_pose_b: torch.tensor, | ||
| target_set: torch.tensor, | ||
| current_goal_idx: int, | ||
| ): |
There was a problem hiding this comment.
please add docstrings for this function
There was a problem hiding this comment.
"""Update the target for the operational space controller.
Args:
osc (OperationalSpaceController): The operational space controller.
root_pose_w (torch.tensor): The root pose in the world frame.
ee_pose_b (torch.tensor): The end-effector pose in the body frame.
target_set (torch.tensor): The target set to track.
current_goal_idx (int): The current goal index.
Returns:
command (torch.tensor): The target command.
ee_target_pose_b (torch.tensor): The end-effector target pose in the body frame.
ee_target_pose_w (torch.tensor): The end-effector target pose in the world frame.
next_goal_idx (int): The next goal index.
Raises:
ValueError: If the target type is undefined.
"""
| def _convert_to_task_frame( | ||
| self, osc: OperationalSpaceController, command: torch.tensor, ee_target_pose_b: torch.tensor | ||
| ): | ||
| command = command.clone() |
There was a problem hiding this comment.
please add docstrings for this function
There was a problem hiding this comment.
"""Convert the target command to the task frame if required.
Args:
osc (OperationalSpaceController): The operational space controller.
command (torch.tensor): The target command to convert.
ee_target_pose_b (torch.tensor): The end-effector target pose in the body frame.
Returns:
command (torch.tensor): The converted target command.
task_frame_pose_b (torch.tensor): The task frame pose in the body frame.
Raises:
ValueError: If the frame is invalid.
"""
| def _check_convergence( | ||
| self, | ||
| osc: OperationalSpaceController, | ||
| ee_pose_b: torch.tensor, | ||
| ee_target_pose_b: torch.tensor, | ||
| ee_force_b: torch.tensor, | ||
| ee_target_b: torch.tensor, | ||
| ): |
There was a problem hiding this comment.
please add docstrings for this function
There was a problem hiding this comment.
"""Check the convergence to the target.
Args:
osc (OperationalSpaceController): The operational space controller.
ee_pose_b (torch.tensor): The end-effector pose in the body frame.
ee_target_pose_b (torch.tensor): The end-effector target pose in the body frame.
ee_force_b (torch.tensor): The end-effector force in the body frame.
ee_target_b (torch.tensor): The end-effector target in the body frame.
Raises:
AssertionError: If the convergence is not achieved.
ValueError: If the target type is undefined.
"""
| def _update_states( | ||
| self, | ||
| robot: Articulation, | ||
| ee_frame_idx: int, | ||
| arm_joint_ids: list[int], | ||
| ): |
There was a problem hiding this comment.
please add docstrings for this function
There was a problem hiding this comment.
"""Update the states of the robot and obtain the relevant quantities for the operational space controller.
Args:
robot (Articulation): The robot to control.
ee_frame_idx (int): The index of the end-effector frame.
arm_joint_ids (list[int]): The indices of the arm joints.
Returns:
jacobian_b (torch.tensor): The Jacobian in the root frame.
mass_matrix (torch.tensor): The mass matrix.
gravity (torch.tensor): The gravity vector.
ee_pose_b (torch.tensor): The end-effector pose in the root frame.
ee_vel_b (torch.tensor): The end-effector velocity in the root frame.
root_pose_w (torch.tensor): The root pose in the world frame.
ee_pose_w (torch.tensor): The end-effector pose in the world frame.
ee_force_b (torch.tensor): The end-effector force in the root frame.
"""
Co-authored-by: jtigue-bdai <[email protected]> Signed-off-by: Özhan Özen <[email protected]>
@jtigue-bdai, Thanks a lot! I have addressed all your comments, updated the changelog, and marked the PR as ready for review. |
|
Hi @jtigue-bdai, I’ve addressed the comments (primarily regarding docstrings) currently marked as unresolved. Would it be okay if I go ahead and mark them as resolved? I’ve also merged the latest changes from main and updated the relevant documentation. The CHANGELOG entries remain as [unreleased], assuming the version/date will be finalized when merging to main. On a separate branch, I’ve implemented the nullspace controller feature for OSC and successfully deployed a trained policy on a real robot to test it. It’s working well, and I observed it significantly improves learning. I’ll create a new PR for it once this one is merged. Thanks again for your review and feedback! 😊 |
Description
This PR adds the
OperationalSpaceControllerto the docs and provides some tests for its parametric features. Moreover, it implements the correspondingOperationalSpaceControllerActionandOperationalSpaceControllerActionCfgclasses so they can be used with manager-based environments.Fixes #873
Type of change
Checklist
pre-commitchecks with./isaaclab.sh --formatconfig/extension.tomlfileCONTRIBUTORS.mdor my name already exists there``