Red Rover: A Multimodal mmWave Radar Spectrum Ecosystem

The red rover¹ project is an end-to-end system for collecting, loading, and processing mmWave radar time signal data along with lidar and camera data designed for both DART-style reconstruction and GRT-style deep learning.

Info

While packaged together as a mono-repo, each module is designed to be able to be installed and used separately. See each module's documentation for requirements and setup instructions.

Red Rover

• roverc

  ---

  radar + lidar + camera + imu data collection system

• roverd

  ---

  efficient recording and storage format with an adl-compliant dataloader

• roverp

  ---

  data processing and visualization tooling for red rover

• i/q-1m

  ---

  one million i/q frames across indoor, outdoor, and bike-mounted settings

See Also

• abstract_dataloader

  ---

  abstract interface for composable dataloaders and preprocessing pipelines

• xwr

  ---

  python interface for collecting raw time signal data from TI mmWave radars

• nrdk

  ---

  neural radar development kit for deep learning on multimodal radar data

• grt

  ---

  our latest work, GRT: Towards Foundational Models for Single-Chip Radar

• dart

  ---

  our prior work, DART: Implicit Doppler Tomography for Radar Novel View Synthesis

• rover

  ---

  our previous data collection platform for radar time signal

---

1. Red Rover is an evolved version of our previous data collection system, rover, which happens to be red. ↩