Safe and Smooth: Certified Continuous-Time Range-Only Localization

This repository contains the code to reproduce all results of the paper:

F. Dümbgen, C. Holmes and T. D. Barfoot, "Safe and Smooth: Certified Continuous-Time 
Range-Only Localization," in IEEE Robotics and Automation Letters, vol. 8, no. 2, 
pp. 1117-1124, Feb. 2023, doi: 10.1109/LRA.2022.3233232.

A pre-print is available at https://arxiv.org/abs/2209.04266.

Installation

This code was last tested with Ubuntu 20.04.1, using Python 3.10.3.

Local install

Make sure to do a recursive clone:

git clone --recursive [email protected]:utiasASRL/safe_and_smooth

All requirements can be installed by running

conda env create -f environment.yml

To check that the installation was successful, run

conda activate safeandsmooth
pip install pytest
pytest .

You can also check that you can generate some toy example results by running

_scripts/generate_test_results.sh

and then checking the output created in _plots_test.

Docker install

You can also use docker to run the code in this repository. To create the docker image, you can run

make safe-build

and to test that installation was successful, you can run

make safe-test

which runs pytest and generates test data.

Generate results

There are three types of results reported in the paper:

• Noise study: Run _scripts/simulate_noise.py to generate the simulation study (Figures 4 and 7 (appendix)).
• Timing study: Run _scripts/simulate_time.py to generate the runtime comparison (Figure 5)
• Real data: Run _scripts/evaluate_real.py to evaluate the real dataset (Figures 1, 5 and 6).

You can generate all results by running (this will take a while)

_scripts/generate_all_results.sh

After generating, all data can be evaluated, and new figures created, by running python _scripts/plot_results.py. For more evaluations of the real dataset, refer to the notebook _notebooks/DatasetEvaluation.ipynb (you may need to run pip install -r requirements.txt for additional plotting libraries).

Docker instructions

If you are using docker, you can use this one-liner to run the above command in a docker container:

make safe-run

Code references

The code refers to the following papers:

• [1] F. Dümbgen, C. Holmes and T. D. Barfoot, "Safe and Smooth: Certified Continuous-Time Range-Only Localization," in IEEE Robotics and Automation Letters, vol. 8, no. 2, pp. 1117-1124, Feb. 2023. https://doi.org/10.1109/LRA.2022.3233232.
• [2] Barfoot, Tim, Chi Hay Tong, and Simo Sarkka. “Batch Continuous-Time Trajectory Estimation as Exactly Sparse Gaussian Process Regression,” 2014. https://doi.org/10.15607/RSS.2014.X.001.
• [3] Barfoot, Timothy D. State Estimation for Robotics. Cambridge University Press, 2017. https://doi.org/10.1017/9781316671528.