Ornstein–Uhlenbeck Process

Saw this at Dyna Robotics. And a bunch of papers use this.

For example:

• What Matters for Batch Online Reinforcement Learning in Robotics
• HuB Learning Extreme Humanoid Balance