simulo.FrameTransformer

Computes transformations between reference frames.

Useful for computing relative poses between robot links and objects in the scene, commonly used for manipulation tasks.

Example:

import simulo

# Track end-effector relative to base
frame_tf = simulo.FrameTransformer(
    source_frame="/World/Robot/base",
    target_frames=[
        simulo.FrameTransformerTarget(
            prim_path="/World/Robot/ee_link",
        ),
        simulo.FrameTransformerTarget(
            prim_path="/World/Object",
        ),
    ],
)
scene.add(frame_tf, at="/World/frame_transformer")

# Read transforms
transforms = frame_tf.read_transforms()  # dict of target_name -> (pos, quat)

simulo.FrameTransformer(
    path: Optional[str] = None,
    source_frame: Optional[str] = None,
    target_frames: Optional[List[FrameTransformerTarget]] = None,
    update_period: float = 0.0,
    enabled: bool = True,
    debug_vis: bool = False,
)

read_transforms()

FrameTransformer.read_transforms() -> torch.Tensor

Get transformations from source to all targets.

Returns:

Transforms tensor of shape (num_envs, num_targets, 7). Each transform is (x, y, z, qw, qx, qy, qz).

Raises:

• RuntimeError — If sensor not initialized

read_target_positions()

FrameTransformer.read_target_positions() -> torch.Tensor

Get target positions relative to source frame.

Returns:

Positions tensor of shape (num_envs, num_targets, 3).

Raises:

• RuntimeError — If sensor not initialized

read_target_orientations()

FrameTransformer.read_target_orientations() -> torch.Tensor

Get target orientations relative to source frame.

Returns:

Orientations tensor of shape (num_envs, num_targets, 4). Quaternions in (w, x, y, z) format.

Raises:

• RuntimeError — If sensor not initialized