Radarize: Enhancing Radar SLAM with Generalizable Doppler-Based Odometry [MobiSys'24]

radarize_demo.mp4

Prerequisites

• Ubuntu 20.04
• ROS Noetic
• Conda with Python 3.8+
• CUDA >= 11.3 capable GPU.
• ImageMagick

Setup

1. Install conda environment with conda env create -f env.yaml.
2. Source environment conda activate radarize_ae and then pip install -e ..
3. Install cartographer_ros. Inside cartographer/, configuration_files/ and launch/ into <catkin_ws>/install_isolated/share/cartographer_ros/.
4. Source conda environment (if not already) and cartographer_ros environment.

conda activate radarize_ae
source <catkin_ws>/install_isolated/setup.bash

Dataset Preparation

1. Download the dataset dataset.zip from the link and unzip into this directory.
2. Download the saved models+outputs eval.zip from link and unzip into this directory.

Evaluation

To generate results in the paper, run the top-level script

./run_eval.sh 

Then,

1. Run ./slam_eval.sh to get the SLAM metrics.
2. Run ./odom_eval.sh to get the odometry metrics.

Training from Scratch

./run.sh