Cláudia Soares

Cláudia Soares

I use optimization, physics, and probability theory to develop robust, interpretable, and trustworthy learning methods to be applied in the real world.

Parameter-free network localization for mobile vehicles by learning from data

Published in 2018 Fourth Underwater Communications and Networking Conference (UComms), 2018

Recommended citation: Claudia Soares, Joao Gomes, "Parameter-free network localization for mobile vehicles by learning from data." 2018 Fourth Underwater Communications and Networking Conference (UComms), 2018. http://dx.doi.org/10.1109/ucomms.2018.8493177

Claudia Soares, Joao Gomes

Abstract

For any operation involving autonomous vehicles, localization of mobile agents is a critical module. Unfortunately, localization methods are still dependant on tweaked parameters, under assumptions of poor trajectories, and that can fail disgracefully under real-world conditions. As our systems evolve to deployment-ready state, we see the looming need for more adaptive responses, where sudden changes of conditions can be accommodated by the localization algorithm. This paper addresses this gap, by proposing a parameter self-tuning scheme based on learning from acquired data.

Access paper here

BibTeX

@inproceedings{Soares_2018,
    author = "Soares, Claudia and Gomes, Joao",
    title = "Parameter-free network localization for mobile vehicles by learning from data",
    url = "http://dx.doi.org/10.1109/ucomms.2018.8493177",
    DOI = "10.1109/ucomms.2018.8493177",
    booktitle = "2018 Fourth Underwater Communications and Networking Conference (UComms)",
    publisher = "IEEE",
    year = "2018",
    month = "Aug",
    pages = "1-5",
    abstract = "For any operation involving autonomous vehicles, localization of mobile agents is a critical module. Unfortunately, localization methods are still dependant on tweaked parameters, under assumptions of poor trajectories, and that can fail disgracefully under real-world conditions. As our systems evolve to deployment-ready state, we see the looming need for more adaptive responses, where sudden changes of conditions can be accommodated by the localization algorithm. This paper addresses this gap, by proposing a parameter self-tuning scheme based on learning from acquired data."
}