Getting Started¶

Prerequisites¶

Linux x86_64 with AVX2 (Haswell / 2013 or newer)
Python 3.12–3.15

Installation¶

Pre-built wheels are available for Python 3.12–3.15 on Linux x86_64. No local compilation required:

pip install wheelchair-planning

The core install pulls in only numpy, scipy and casadi: motion planning (OMPL + VAMP), FLASK kinodynamic planning, TOPP-RA time parameterization and constraint / cost planning all work out of the box. Optional extras:

Extra	Adds	Enables
`kinematics`	`pin`, `pin-pink`, `osqp`	Pinocchio FK, collision model, Pink IK
`examples`	`pybullet`, `fire`, `matplotlib`, `xmltodict`	PyBullet environments and the example scripts
`dev`	`pytest`, `trimesh`, `pre-commit`	The test suite

pip install "wheelchair-planning[kinematics,examples]"

The TRAC-IK solver needs the compiled pytracik extension, which is only built when orocos-kdl and NLopt are available — use the pixi development environment below for it.

Verify installation¶

from wheelchair_planning.planning import create_planner
from wheelchair_planning.wheelchair import HOME_JOINTS

planner = create_planner("wheelchair_arm")
start = planner.extract_config(HOME_JOINTS)
result = planner.plan(start, planner.sample_valid())
print(f"Planning {'succeeded' if result.success else 'failed'}")

Building Wheels from Source¶

Release wheels are built with cibuildwheel using the configuration in pyproject.toml:

pipx run cibuildwheel --platform linux

This builds manylinux x86_64 wheels for every supported Python version into wheelhouse/. Docker must be installed and running.

Development Setup¶

For contributing or rebuilding C++ dependencies from source, use pixi:

git clone --recursive https://github.com/H-tr/Wheel-Chair-Planning.git
cd Wheel-Chair-Planning
bash scripts/setup.sh

Or manually:

git clone --recursive https://github.com/H-tr/Wheel-Chair-Planning.git
cd Wheel-Chair-Planning
pixi install
pixi run cricket-build
pixi run foam-build
bash scripts/download_assets.sh