Implementation:Haosulab ManiSkill AntEnv

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

Overview

Concrete implementation of the Ant locomotion control task environment in ManiSkill, adapted from the DeepMind Control Suite.

Description

The AntEnv defines a quadruped "ant" robot locomotion task where the agent must control an ant-like robot to stand, walk, or run. The ant robot is defined inline via MJCF and has 8 actuated joints. Three registered variants exist:

• MS-AntStand-v1 (max_episode_steps=100): The ant must maintain a standing posture at a target height of 0.55m.
• MS-AntWalk-v1 (max_episode_steps=100): The ant must walk at a target speed of 0.5 m/s.
• MS-AntRun-v1 (max_episode_steps=100): The ant must run at a target speed of 4 m/s.

The robot uses PD joint position control. Rewards are composed of standing height reward, movement speed reward, and a control cost penalty. The supported robot is ant (defined in the same file as AntRobot).

Usage

Use this environment for benchmarking locomotion control algorithms. It provides dense and normalized dense reward modes, making it suitable for reinforcement learning research on multi-legged locomotion tasks.

Code Reference

Source Location

• Repository: Haosulab_ManiSkill
• File: mani_skill/envs/tasks/control/ant.py

Signature

@register_env("MS-AntStand-v1", max_episode_steps=100)
class AntStandEnv(AntEnvBase):
    ...

@register_env("MS-AntWalk-v1", max_episode_steps=100)
class AntWalkEnv(AntEnvBase):
    ...

@register_env("MS-AntRun-v1", max_episode_steps=100)
class AntRunEnv(AntEnvBase):
    ...

Import

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

I/O Contract

Inputs

Name	Type	Required	Description
obs_mode	str	No	Observation mode (default: "state")
reward_mode	str	No	Reward mode: "dense", "normalized_dense", "none"
control_mode	str	No	Control mode for the ant robot

Outputs

Name	Type	Description
obs	dict/array	Observation including joint positions, velocities, and robot state
reward	float	Composed of standing reward, movement reward, and control cost
terminated	bool	Whether episode ended by success/failure
truncated	bool	Whether episode hit max steps (100)
info	dict	Contains evaluation metrics

Usage Examples

Basic Usage

import gymnasium as gym
import mani_skill.envs

env = gym.make("MS-AntWalk-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