Sim2Real: Embodied AI with Isaac Sim and Unitree G1

Course Outline Orientation – Overview, outcomes, system requirements Install & Verify – Install Isaac Sim, set up Isaac Lab, project structure USD & PhysX – USD basics, PhysX physics, scene assembly Sensors & Observations – IMU, force sensors, height data, observation terms Commands & Actions – Velocity/waypoint commands, action spaces, scaling MDP – Rewards, terminations, events, randomization hooks Terrain & Curriculum – Terrain generators and difficulty progression Perf & VecEnv – Vectorized environments, async/headless rendering RL (RSL-RL) – PPO training, key hyperparameters, debugging rewards Go1 Task – Go1 model, locomotion env, domain randomization Obstacles & Cues – Goals, markers, obstacles, navigation cues Eval & Logs – TensorBoard/W&B, saving and loading policies Sim2Sim – ROS 2 bridging, packaging policies as ROS nodes Sim2Real (Go1) – Bring-up, safety, deployment, tuning, validation Troubleshooting – Common issues and fixes
Product image for Sim2Real: Embodied AI with Isaac Sim and Unitree G1

Course content

7 sections | 7 lessons