Implementation:Isaac sim IsaacGymEnvs AllegroKukaTwoArmsReorientation

Knowledge Sources	IsaacGymEnvs
Domains	Robotic_Manipulation, Bimanual_Manipulation
Last Updated	2026-02-15 11:00 GMT

Overview

AllegroKukaTwoArmsReorientation is a dual-arm object reorientation task environment where two Kuka arms with Allegro hands must cooperate to rotate an object to a target pose.

Description

The AllegroKukaTwoArmsReorientation class extends AllegroKukaTwoArmsBase to implement a bimanual object reorientation task. This task requires two robotic arm-hand systems to coordinate in order to rotate a manipulated object to match both a randomly sampled target position and a randomly sampled target orientation (quaternion). It uses four corner keypoints ([1,1,1], [1,1,-1], [-1,-1,1], [-1,-1,-1]) to fully capture the object's rotational state.

Goal assets corresponding to each object type in the training set are loaded with gravity disabled. A translucent goal object is placed at the target pose to visually indicate the desired configuration. For non-block objects, the goal object is colored in a light blue shade. During target resets, the position is randomized within a configurable volume with an x-offset of +/-0.75 (randomly near either arm), and the orientation is set using get_random_quat to sample a uniformly random rotation. The velocities of the goal object are zeroed out on each reset.

Curriculum learning is applied through tolerance_curriculum to progressively tighten the success tolerance. The _true_objective method provides a PBT-compatible objective that prioritizes tolerance improvements over raw success counts, following the same strategy as the single-arm reorientation task.

Usage

Use this class when training a reinforcement learning policy for bimanual in-hand object reorientation tasks. It is suitable for research into coordinated dual-arm manipulation where both position and orientation of the object must be controlled simultaneously. Configure it through the IsaacGymEnvs Hydra configuration system with dual-arm environment settings.

Code Reference

Source Location

• Repository: IsaacGymEnvs
• File: isaacgymenvs/tasks/allegro_kuka/allegro_kuka_two_arms_reorientation.py
• Lines: 42-158

Signature

class AllegroKukaTwoArmsReorientation(AllegroKukaTwoArmsBase):
    def __init__(self, cfg, rl_device, sim_device, graphics_device_id, headless, virtual_screen_capture, force_render):
        ...

    def _object_keypoint_offsets(self):
        ...

    def _load_additional_assets(self, object_asset_root, arm_pose):
        ...

    def _create_additional_objects(self, env_ptr, env_idx, object_asset_idx):
        ...

    def _after_envs_created(self):
        ...

    def _reset_target(self, env_ids: Tensor) -> None:
        ...

    def _extra_object_indices(self, env_ids: Tensor) -> List[Tensor]:
        ...

    def _extra_curriculum(self):
        ...

    def _true_objective(self) -> Tensor:
        ...

Import

from isaacgymenvs.tasks.allegro_kuka.allegro_kuka_two_arms_reorientation import AllegroKukaTwoArmsReorientation

I/O Contract

Inputs

Name	Type	Required	Description
cfg	dict	Yes	Hydra configuration dictionary containing environment parameters, object types, dual-arm settings, and curriculum parameters
rl_device	str	Yes	Device string for the RL algorithm (e.g., "cuda:0")
sim_device	str	Yes	Device string for the simulation (e.g., "cuda:0")
graphics_device_id	int	Yes	GPU device ID for rendering
headless	bool	Yes	Whether to run without a display
virtual_screen_capture	bool	Yes	Whether to capture virtual screen output
force_render	bool	Yes	Whether to force rendering even in headless mode

Outputs

Name	Type	Description
rew_buf	Tensor	Per-environment reward buffer incorporating position and orientation error from both arms
is_success	Tensor	Boolean tensor indicating which environments achieved the target pose
true_objective	Tensor	PBT objective combining tolerance progress and success count

Usage Examples

# Register and instantiate the AllegroKukaTwoArmsReorientation task
from isaacgymenvs.tasks.allegro_kuka.allegro_kuka_two_arms_reorientation import AllegroKukaTwoArmsReorientation

# Typically created through the task registry with Hydra config
env = AllegroKukaTwoArmsReorientation(
    cfg=task_cfg,
    rl_device="cuda:0",
    sim_device="cuda:0",
    graphics_device_id=0,
    headless=True,
    virtual_screen_capture=False,
    force_render=False,
)

# Run a step in the environment
obs = env.reset()
actions = policy(obs)
obs, rewards, dones, info = env.step(actions)