Implementation:Haosulab ManiSkill PickClutterYCB

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

Overview

Concrete implementation of the pick-from-clutter task environment using YCB objects in ManiSkill.

Description

The PickClutterYCBEnv extends PickClutterEnv to use objects from the YCB dataset in cluttered arrangements. The robot must pick up a target object from a cluttered table and move it to a goal position.

Registered as PickClutterYCB-v1 with max_episode_steps=100 and asset_download_ids=["ycb", "pick_clutter_ycb_configs"]. Supported robots are panda and fetch. The reward mode is "none" only.

Clutter configurations are loaded from JSON episode files (default: ycb_train_5k.json.gz). Multiple objects are placed on the table per configuration, and a visible target object is randomly selected. The goal position is randomized in a 3D region above the table.

Usage

Use this environment for cluttered scene manipulation research. It tests the ability to identify and extract a specific object from among multiple objects on a table.

Code Reference

Source Location

• Repository: Haosulab_ManiSkill
• File: mani_skill/envs/tasks/tabletop/pick_clutter_ycb.py

Signature

@register_env(
    "PickClutterYCB-v1",
    asset_download_ids=["ycb", "pick_clutter_ycb_configs"],
    max_episode_steps=100,
)
class PickClutterYCBEnv(PickClutterEnv):
    DEFAULT_EPISODE_JSON = f"{ASSET_DIR}/tasks/pick_clutter/ycb_train_5k.json.gz"

Import

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

I/O Contract

Inputs

Name	Type	Required	Description
obs_mode	str	No	Observation mode
reward_mode	str	No	Reward mode: "none"
control_mode	str	No	Control mode for the robot

Outputs

Name	Type	Description
obs	dict/array	Observation based on obs_mode
reward	float	No reward (none mode only)
terminated	bool	Whether episode ended
truncated	bool	Whether episode hit max steps (100)
info	dict	Contains success, fail flags

Usage Examples

Basic Usage

import gymnasium as gym
import mani_skill.envs

env = gym.make("PickClutterYCB-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