Open-DiffLoco: Open-Source Differentiable Learning for Deployable Blind Quadruped Locomotion

Martin Opat

Developing deployable locomotion policies through conventional reinforcement learning often requires complex reward engineering and expensive training times. While differentiable s…