silvery107/rl-mpc-locomotion

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1020 stars · 92 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A Python framework combining deep reinforcement learning with model predictive control (MPC) for quadruped robot locomotion, where a policy network dynamically predicts MPC weight parameters. It supports Unitree Go1, A1, and Aliengo robots with parallel training in NVIDIA Isaac Gym.

Use cases

• train RL policies for quadruped locomotion
• simulate and control a quadruped robot with MPC
• run a hierarchical RL+MPC controller on a Unitree robot
• learn gait control for legged robots in Isaac Gym
• port an MPC locomotion controller between simulators

When to choose

• you need RL-trained MPC weight prediction for quadruped locomotion
• you work with Unitree A1, Go1, or Aliengo robots
• you want a Python MPC controller with open sensor/motor interfaces
• you need fast parallel RL training in Isaac Gym

When to avoid

• you need out-of-the-box sim2real deployment (not included)
• you work with biped or non-quadruped robots
• you need a non-Python MPC implementation
• you cannot use NVIDIA Isaac Gym or CUDA GPUs

Facets

library · maturity active

reinforcement-learning simulation robotics machine-learning robotics reinforcement-learning machine-learning simulation python quadruped mpc isaac-gym sim2real locomotion unitree pytorch linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="silvery107/rl-mpc-locomotion")