Mail: Clementin.Boittiaux (at) ifremer.fr
ESIEE Paris 2020 graduate
Thesis topic: Design of autonomous navigation functionalities for the Coral robot, at a depth of 6000 m, based on bergs detected on the seabed.
- Thesis supervisor: Vincent HUGEL
- Co-supervisor : Claire DUNE
- Co-supervisor : Aurélien ARNAUBEC
- Co-supervisor : Ricard MARXER
Today, the vast majority of the ocean floor remains unexplored. Exploration conditions are difficult due to the many scientific challenges presented by this environment. To better understand the deep sea, many tools have been developed over the years to collect samples, videos and physical data from the abyss. Recently, in 2020, the French oceanographic fleet was expanded with the addition of Ulyx, an autonomous underwater vehicle (AUV) capable of diving to 6,000m.
The objective of such a device is to be able to decide alone on certain actions to be carried out at the bottom to successfully complete the mission previously entrusted to it. To achieve this, Ulyx will need to be able to position itself accurately and recognize the places of scientific interest indicated to it. However, the positioning systems currently in use, such as the USBL, have an uncertainty of the order of 10 m when deployed at great depths.
This thesis therefore aims to reduce this uncertainty by integrating on-board intelligence into the AUV. With the latest major technological advances in deep learning, particularly in the field of vision, our research focuses on the recognition of previously visited points of interest and the localization of the vehicle in these same regions. Our efforts are focused on vision-centric approaches, which present a number of challenges in this environment. One of these, for example, is to make our models robust to topological or physical changes that can occur from one year to the next.
- Nom : Clémentin Boittiaux
- IdHAL : clementin-boittiaux
- ORCID : 0000-0002-1487-3502 ,
Year of production
- Clémentin Boittiaux, Claire Dune, Maxime Ferrera, Aurélien Arnaubec, Ricard Marxer, et al.. Eiffel Tower: A Deep-Sea Underwater Dataset for Long-Term Visual Localization. The International Journal of Robotics Research, In press, ⟨10.1177/02783649231177322⟩. ⟨hal-04089339⟩
- Clémentin Boittiaux, Ricard Marxer, Claire Dune, Aurélien Arnaubec, Vincent Hugel. Homography-Based Loss Function for Camera Pose Regression. IEEE Robotics and Automation Letters, In press, 7 (3), pp.6242-6249. ⟨10.1109/LRA.2022.3168329⟩. ⟨hal-03654445⟩
- Clémentin Boittiaux, Claire Dune, Aurélien Arnaubec, Ricard Marxer, Maxime Ferrera, et al.. Long-term visual localization in deep-sea underwater environment. ORASIS, Thanh Phuong Nguyen, May 2023, Carqueiranne, France. ⟨hal-04108737⟩
- Clémentin Boittiaux, Paul Nguyen Hong Duc, Nicolas Longépé, Sara Pensieri, Roberto Bozzano, et al.. Multi-modal deep learning models for ocean wind speed estimation. 2020 MACLEAN: MAChine Learning for EArth ObservatioN Workshop, MACLEAN 2020, Sep 2020, Virtual online, France. ⟨hal-03104246⟩