kevinzakka/mink

Python inverse kinematics based on MuJoCo.

What it solves

It provides a way to calculate the joint angles needed for a robot to reach a specific target position or orientation (inverse kinematics), while respecting physical constraints like joint limits and avoiding collisions.

How it works

Built on the MuJoCo physics engine, the library uses differential inverse kinematics. It employs a Lie group interface for rigid body transformations and includes native C optimization for the most computationally intensive paths to ensure efficiency.

Who it’s for

Robotics researchers and developers working with a wide range of hardware, including single or dual arms, dexterous hands, humanoids, and legged or mobile manipulators.

Highlights

  • Broad Robot Support: Works with everything from the UR5e and Franka Panda arms to Unitree humanoids and Boston Dynamics Spot.
  • Constraint Handling: Manages joint position and velocity limits, as well as collision avoidance between any pair of geoms.
  • Advanced Kinematics: Supports closed-chain kinematics through equality constraints.
  • Optimized Performance: Uses native C optimization for critical inverse kinematics calculations.

Related

  • Project
    Phylliade/ikpy

    A pure-Python library for computing the inverse kinematics of any robot, supporting URDF and MuJoCo imports and an accelerated JAX backend.

  • Project
    stephane-caron/pink

    A Python library for differential inverse kinematics that uses weighted tasks and quadratic programming to steer articulated robots toward targets while respecting physical limits.

  • Project
    UM-ARM-Lab/pytorch_kinematics

    A parallel and differentiable robotics kinematics library for PyTorch that provides fast forward and inverse kinematics and Jacobian calculations.

  • Project
    open-planning/roboplan

    A modern C++ robot motion planning library based on Pinocchio that provides tools for inverse kinematics, RRT-based path planning, and trajectory timing.

  • Project
    Farama-Foundation/Gymnasium-Robotics

    A collection of Reinforcement Learning robotic environments powered by the MuJoCo physics engine and the Gymnasium API for training AI agents in manipulation and navigation tasks.