aaucns_rio

This software contains an implementation of Radar-Inertial Odometry using multi-state EKF and factor graphs.

Compilation

catkin_make --force-cmake -DCMAKE_BUILD_TYPE={Release, Debug}.

--force-cmake is there only to force copying the yaml config file into the ~/.ros folder. If you copy it manually you do not need that option.

Use

The software can be used either as a ROS node or to replay ROS bag files. For the latter, modify the file in nodes/rio_(fg)_replay_node.cpp by giving the bag filename (file must be in ~/.ros) and modify the topic names. Acronym fg stands for "factor graph". The default version uses EKF as the estimation backbone.

Within src/rio_(fg)_replay.cpp you might also have to modify the message types.

Then, run the software using one of the launch files in the launch/ folder.

RIO estimator must be configured through a yaml config file. An example is in config/ folder. The software reads this config from the ~/.ros folder (or your alternative ROS runtine folder if you changed it).

Dependencies

Users are expected to have Eigen, PCL and GTSAM installed system-wide.

Cite (BibTeX)

If you use this software please cite:

@INPROCEEDINGS{Michalczyk2024aaucns_rio,
  author={Michalczyk, Jan and Quell, Julius and Steidle, Florian and Müller, Marcus G. and Weiss, Stephan},
  booktitle={2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  title={Tightly-Coupled Factor Graph Formulation For Radar-Inertial Odometry},
  year={2024},
  volume={},
  number={},
  pages={3364-3370},
  doi={10.1109/IROS58592.2024.10801945}}

Example dataset

You can download here and here example bagfiles to use with the rio. Remember to place them in your ~/.ros folder.

Parameters

In the file config/config.yaml there are several parameters to control the behavior of the estimator. Some parameters require at the moment the code to be recompiled. They are in the include/aaucns_rio/config.h file.

Future work

It is planned to provide a docker container with the code.