Simulators

MuJoCo, Gazebo, Isaac Sim, PyBullet, Drake, Webots — six serious simulators, six sweet spots. A decision tree and honest recommendations.

~14 min

URDF import, plugins, sensors, and the headless simulation workflow for CI. The simulator that ships with ROS 2 — slower than MuJoCo, but with the deepest ROS integration in the industry.

~14 min

Open-source since 2021, batched on GPU since 2022 — MuJoCo is now the RL default. Here's the XML format, the Python API, and how MJX makes it 10,000× faster.

~16 min

Zero-install, URDF-native, fastest path from 'I want to simulate a robot' to 'I'm running my first script.' The teaching tool, the prototyping tool, and the lightweight RL sim that punches above its weight.

~11 min

Photorealistic sensors, USD assets, and scaling RL training to thousands of parallel environments. The simulator the industry adopted for production-scale robot learning.

~14 min

More than a simulator — a robotics research toolkit with rigorous physics, first-class optimization, and Lagrangian formulations. The tool MIT's Underactuated Robotics course teaches with.

~12 min

Cross-platform, GUI-first, and perfect for teaching cohorts with mixed hardware. The simulator that doesn't try to be cutting-edge but consistently delivers.

~9 min

The three dialects of 'what is this robot' — what each captures, what each loses, and how to convert between them without losing your mind.

~12 min

The interface every RL framework expects. Wrap any simulator behind reset/step/observation_space, and the entire RL ecosystem can train on it. Five-method protocol; thousands of trainers.

~11 min

Mixing real and simulated components — when it's brilliant and when it's a debugging trap. The patterns for testing real robots against simulated worlds, and vice versa.

~12 min