A Python package to solve robot arm inverse kinematics symbolically or hybrid.
Generate closed-form inverse kinematics solutions as LaTeX, Python, and C++ code using behavior trees.
IKBT is a Python package that automates closed-form solutions to manipulator inverse kinematics problems using behavior trees for action selection. It suits robotics engineers and researchers needing explicit symbolic or hybrid solutions rather than purely numerical approximations.
The software applies behavior trees to try symbolic, one-variable, and hybrid solvers in order to minimize numerical portions. It outputs native C++17 and Python code alongside LaTeX documentation, and includes automated closed-loop validation scripts.
It suits robotics researchers and engineers working on manipulator kinematics who require explicit closed-form or hybrid solutions. It does not suit those looking for generic, out-of-the-box neural network planners without kinematic formulations.
Automatically sequences symbolic, one-variable, and hybrid solvers to find closed-form kinematic solutions.
Outputs complete, self-contained Python and C++17 code using Eigen types directly from the solver.
Generates SVG vector graphics visualizing solution sets as a transit map via tube_map.py.
Runs rigorous closed-loop numerical validation scripts to verify that generated solutions match target end-effector poses.
Free to use as source code, with the only cost being your own time and compute resources to run and integrate the solver.
879 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
IKBT outputs solutions as LaTeX documents, Python modules, and C++17 source files.
It finds a solvable arm close to the true arm, gets exact solutions, and seeds a numerical damped least squares method.
Yes, scripts like numerical_closed_loop_sol_check.py and cpp_closed_loop_check.py automate solution validation.
See Zekai first in Google