Installation

Requirements

Python 3.10 or later. No GPU required; GPU acceleration is optional via JAX.

Install from PyPI

You can install drone-models directly from PyPI using pip:

pip install drone-models

This brings in NumPy, SciPy, CasADi, and array-api-compat as core dependencies. The models are immediately usable with NumPy arrays.

For system identification — fitting model parameters from your own flight data — you need the sysid extra, which adds JAX (used for auto-differentiation during optimization) and matplotlib:

pip install "drone-models[sysid]"

Developer install with pixi

For development work, we use pixi as the environment manager. Pixi is a conda-based tool that locks the full dependency tree — including native libraries like CasADi — into a reproducible environment. It also defines named feature environments so you can activate exactly the dependencies you need for a given task.

Clone the repository and install all environments in one step:

git clone https://github.com/utiasDSL/drone-models.git
cd drone-models
pixi install

The available environments are:

Environment	Activate with	Includes
default	pixi shell	Core package + ruff linter
sysid	pixi shell -e sysid	Core + JAX + matplotlib for system identification
tests	pixi shell -e tests	Core + JAX + pytest + array-api-strict for running the test suite
docs	pixi shell -e docs	Core + MkDocs + mkdocstrings for building the documentation

Each environment has pixi tasks associated with it:

# Run the test suite
pixi run -e tests tests

# Serve the documentation locally (live reload)
pixi run -e docs docs-serve

# Build the documentation
pixi run -e docs docs-build

SCIPY_ARRAY_API

drone-models requires SciPy to be loaded with its experimental Array API support enabled. The package sets the SCIPY_ARRAY_API=1 environment variable automatically at import time — but only if SciPy has not been imported yet. If your environment or another library imports SciPy before drone_models, you will see a RuntimeError.

The fix is straightforward: make sure drone_models is imported first.

import drone_models   # must precede any scipy import
import scipy

Or set the variable in your shell before starting Python:

export SCIPY_ARRAY_API=1

Supported drone configurations

The following drone configurations ship with pre-fitted parameters, covering both the brushed Crazyflie 2.x series and the brushless Crazyflie 2.1:

drone_model	Platform	Mass
cf2x_L250	Crazyflie 2.x — L250 props	31.9 g
cf2x_P250	Crazyflie 2.x — P250 props	31.8 g
cf2x_T350	Crazyflie 2.x — T350 props	37.9 g
cf21B_500	Crazyflie 2.1 Brushless — 500 props	43.4 g

If your drone is not listed, you can fit parameters from your own flight data using the system identification pipeline, then inject the identified values into any model.

Verify

from drone_models import available_models
list(available_models)  # ['first_principles', 'so_rpy', 'so_rpy_rotor', 'so_rpy_rotor_drag']