robosuite

robosuite

robosuite is a simulation framework and benchmark suite for robot learning developed by the ARISE Initiative. Built on top of MuJoCo, it provides a modular, accessible API for creating robotic manipulation environments suitable for reinforcement learning and imitation learning research.

Key Features

• 9 pre-built robot models (Panda, Sawyer, IIWA, UR5e, Jaco, etc.) and 12 manipulation tasks
• Modular design: swap arms, grippers, controllers, and cameras freely
• Photorealistic rendering with off-screen and on-screen modes
• Built-in teleoperation for data collection via SpaceMouse or keyboard
• Standardised gym-compatible environments for benchmarking

Quick Start

pip install robosuite

import robosuite as suite

env = suite.make(
    env_name="Lift",
    robots="Panda",
    has_renderer=True,
    has_offscreen_renderer=False,
    use_camera_obs=False,
)
obs = env.reset()
for _ in range(1000):
    action = env.action_space.sample()
    obs, reward, done, info = env.step(action)
    env.render()

npx ai-supply add robosuite-robot-simulation

Curated mirror of the open-source robosuite (MIT). Get it from the source.