Stone Soup Goes NUTS: Adding Proposals and the No-U-Turn Sampler to Stone Soup

Tutorial repository for the paper submitted and presented at FUSION 2025 with title: "Stone Soup goes NUTS: adding proposals and The No-U-Turn Sampler to Stone Soup" You can read the paper on IEEE Xplore

Abstract

Particle filters are essential for state estimation in non-linear and non-Gaussian systems, with performance hinging on effective proposal distributions. This paper presents the implementation of Kalman Filter and No-U-Turn Sampler (NUTS) proposals within the Stone Soup Python framework. The Kalman Filter proposal offers computational efficiency for structured systems, while NUTS enables robust exploration of complex, high-dimensional distributions. Benchmark evaluations demonstrate the complementary strengths of these methods, enhancing Stone Soup’s capabilities for diverse state estimation challenges

Authors

Alberto Acuto¹, Lyudmil Vladimirov¹, Alessandro Varsi¹, Paul Horridge¹ and Simon Maskell¹

¹ University of Liverpool, Department of Electrical Engineering and Electronics

Community engagement

The code shared can be run and modifief for various uses. We refer to the Stone Soup community for further examples on how to use the framework, raise issues on the implementation and proposing and implementing improvements.

Installation

To install and test the code provided you can use a virtual environment. How to set up a virtual environemnt:

python3 -m venv <name_of_venv>

Activate the venv

Unix

source <name_of_venv>/bin/activate

Windows

.\<name_of_venv>\Scripts\activate

Then you can install all the dependencies by doing:

python -m pip install -e .[dev]

Please note:

This code is built on top of Stone Soup and when the code will be reviewed and included in the main branch we will update the installation guide for easier running of the tutorial.

Cite this work

@INPROCEEDINGS{11124070,
  author={Acuto, Alberto and Vladimirov, Lyudmil and Varsi, Alessandro and Horridge, Paul and Maskell, Simon},
  booktitle={2025 28th International Conference on Information Fusion (FUSION)}, 
  title={Stone Soup Goes NUTS: Adding Proposals and the No-U-Turn Sampler to Stone Soup}, 
  year={2025},
  volume={},
  number={},
  pages={1-8},
  keywords={Accuracy;Particle filters;Mathematical models;Time measurement;Robustness;Trajectory;State-space methods;Proposals;Kalman filters;State estimation;Particle Filter;proposal distributions;importance sampling;no-u-turn-sampler},
  doi={10.23919/FUSION65864.2025.11124070}
}