Train a balance bot with PPO: Reinforcement Learning for Robotics Part 2

ShawnHymel · x · 2026-09-21

MakerIO teamed up with Shawn Hymel for part 2 of the "Reinforcement Learning for Robotics" series: training a self-balancing bot with PPO. The tutorial is a great entry point for RL — you can watch the agent visibly improve in simulation, and training runs fine on a CPU alone, even though PID loops could solve the task more easily.

Original post →

More from Embodied

Embodied channel →