Implementation:Haosulab ManiSkill PlaceSphere

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

Overview

Concrete implementation of the sphere placement task environment in ManiSkill.

Description

The PlaceSphereEnv requires a robot to pick up a sphere and place it into a shallow bin on the table. The bin is built from five box shapes (a base and four edge walls) forming a small container. The sphere has radius 0.02m.

Registered as PlaceSphere-v1 with max_episode_steps=50. Supported robots: panda and fetch. Reward modes include "normalized_dense", "dense", "sparse", and "none".

Randomizations: sphere position in [-0.1, -0.05] x [-0.1, 0.1], bin position in [0, 0.1] x [-0.1, 0.1]. Success requires the sphere to be on top of the bin (xy within 0.005m, z within 0.005m of expected height), the sphere to be static, and the gripper to not be grasping the sphere.

Usage

Use this environment for precise placement research. The combination of grasping a sphere and placing it in a small bin tests both grasping and fine-grained positioning skills.

Code Reference

Source Location

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

Signature

@register_env("PlaceSphere-v1", max_episode_steps=50)
class PlaceSphereEnv(BaseEnv):
    SUPPORTED_ROBOTS = ["panda", "fetch"]
    agent: Union[Panda, Fetch]
    radius = 0.02

Import

import gymnasium as gym
import mani_skill.envs
env = gym.make("PlaceSphere-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 the robot

Outputs

Name	Type	Description
obs	dict/array	Observation including TCP pose, sphere pose, bin position
reward	float	Dense reward: reaching + grasping + placing + ungrasping + static
terminated	bool	Whether episode ended by success
truncated	bool	Whether episode hit max steps (50)
info	dict	Contains success, is_obj_grasped, is_obj_on_bin, is_obj_static

Usage Examples

Basic Usage

import gymnasium as gym
import mani_skill.envs

env = gym.make("PlaceSphere-v1", obs_mode="state", render_mode="rgb_array")
obs, info = env.reset()
for _ in range(50):
    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