`xwr`: Linux-Based Real-Time Raw Data Capture for TI mmWave Radars¶

AWR1642Boost DCA1000EVM AWR1843Boost

PyPI - Python Version PyPI - Types GitHub issues

xwr is a pure-python, linux-based real time raw data capture system for TI mmWave radars, and includes four key components:

xwr: a high-level data capture interface
xwr.rsp: a radar signal processing library with Numpy, Pytorch, and Jax support
xwr.radar: a parameterized python interface for the default radar firmware
xwr.capture: a pure-python, real-time interface for the DCA1000EVM

Fully Typed

xwr is type-annotated where possible¹, and has runtime type checking enabled via beartype.

Requirements¶

xwr assumes a linux-based system and radar hardware which consists of the DCA1000EVM and a supported TI mmWave Radar (XWR) development board.

Supported Devices

Tip

This list of supported radars is expanding, and we may add support for additional radars in the future! Feel free to leave an issue if you have a specific request.

Install¶

The xwr library can be installed from pypi or github:

Direct InstallIntegrated DevelopmentStandalone DevelopmentUsing uv and pyproject.toml

pip install xwr
# or, for the bleeding-edge version:
pip install git+ssh://github.com/RadarML/xwr.git

git clone git@github.com:RadarML/xwr.git
pip install -e ./xwr

git clone git@github.com:RadarML/xwr.git
cd xwr; uv sync --all-extras --frozen

[project]
dependencies = ["xwr"]

[tool.uv.sources]
xwr = { git = "ssh://git@github.com/RadarML/xwr.git" }

See the user guide and hardware setup for instructions on how to configure and use xwr.

Warning

xwr does not include a copy of torch or jax by default! You must specify your own dependency and/or use the xwr[torch] and xwr[jax] extras if you intend to use these backends for the radar signal processing (xwr.rsp) submodule.

xwr: Linux-Based Real-Time Raw Data Capture for TI mmWave Radars¶

Requirements¶

Install¶

See Also¶

`xwr`: Linux-Based Real-Time Raw Data Capture for TI mmWave Radars¶