Implementation:Haosulab ManiSkill TransportBox

Knowledge Sources	Haosulab_ManiSkill
Domains	Robotics, Simulation, Humanoid_Control
Last Updated	2026-02-15 08:00 GMT

Overview

Concrete implementation of the box transport task for the Unitree G1 humanoid robot in ManiSkill.

Description

The TransportBoxEnv requires a Unitree G1 humanoid robot (upper body with head camera) to find a box on one table, pick it up, and transport it to another table. The robot is initialized in a standing pose with arms positioned for grasping.

Registered as UnitreeG1TransportBox-v1 with max_episode_steps=100. The supported robot is unitree_g1_simplified_upper_body_with_head_camera.

Randomizations include the box's xy position in the region [-0.05, -0.05] x [0.2, 0.05] and z-axis rotation in [0, pi/6]. Success is determined by the box resting on top of the target table. Reward modes include "normalized_dense", "dense", "sparse", and "none".

Usage

Use this environment for humanoid manipulation research involving whole-body coordination for pick-and-place tasks across different surfaces.

Code Reference

Source Location

• Repository: Haosulab_ManiSkill
• File: mani_skill/envs/tasks/humanoid/transport_box.py

Signature

@register_env("UnitreeG1TransportBox-v1", max_episode_steps=100)
class TransportBoxEnv(BaseEnv):
    SUPPORTED_ROBOTS = ["unitree_g1_simplified_upper_body_with_head_camera"]
    agent: UnitreeG1UpperBodyWithHeadCamera

Import

import gymnasium as gym
import mani_skill.envs
env = gym.make("UnitreeG1TransportBox-v1")

I/O Contract

Inputs

Name	Type	Required	Description
obs_mode	str	No	Observation mode
reward_mode	str	No	Reward mode: "normalized_dense", "dense", "sparse", "none"
control_mode	str	No	Control mode for G1 humanoid

Outputs

Name	Type	Description
obs	dict/array	Observation including robot state, box pose, goal table position
reward	float	Reward based on grasping, transporting, and placing the box
terminated	bool	Whether episode ended by success/failure
truncated	bool	Whether episode hit max steps (100)
info	dict	Contains success flag

Usage Examples

Basic Usage

import gymnasium as gym
import mani_skill.envs

env = gym.make("UnitreeG1TransportBox-v1", obs_mode="state", render_mode="rgb_array")
obs, info = env.reset()
for _ in range(100):
    action = env.action_space.sample()
    obs, reward, terminated, truncated, info = env.step(action)
    if terminated or truncated:
        obs, info = env.reset()
env.close()

Related Pages

• Environment:Haosulab_ManiSkill_Python_SAPIEN_Core