// the take
A Python reimplementation of Peter Corke's long-running MATLAB Robotics Toolbox, covering manipulator kinematics/dynamics (DH and the ETS notation), plus lighter support for mobile robot planning, localization and SLAM. Aimed at robotics researchers, grad students and educators who want to prototype arm control or kinematics algorithms without standing up ROS.
Kinematics and Jacobian/Hessian computation are genuinely fast (sub-microsecond FK, microsecond-scale numerical IK) and the toolbox ships four different IK solver families (LM, NR, GN, QP) plus an optional symbolic/analytic solver via SymPy — most hobby kinematics libraries only give you one. It can load any of ~50 bundled models or pull arbitrary robots on demand from the robot_descriptions package, so you're not stuck maintaining your own URDF zoo. Active maintenance is real, not just a stale popular repo: CI, Codecov, Codacy, and a changelog driven by release-please, with a push the day before this was evaluated.
The README admits the tradeoff outright: routines favor readability over computational efficiency, so anything dynamics-heavy (RNE, full dynamics sims) will be slower than a C++ equivalent like Pinocchio. Visualization is second-class on Windows — Swift isn't supported at all out of the box and needs a manual one-line source hack, and the collision extra silently drops `coal` since it has no Windows wheels, so you lose collision checking unless you fall back to conda. The project is really a manipulator-kinematics library with mobile robotics (path planning, EKF SLAM) bolted on as a smaller, less-documented side feature, so if mobile robotics is your actual use case this isn't a strong fit despite being in the topics list.
We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →