The Speedy robot sampling a 17th century bottle with its Techno Concept robotic hand (Lune shipwreck, 90 meter deep, Toulon, France). Copyright: DRASMM/Images Explorations – F. Osada / T. Seguin

Underwater robotics for deep-sea archaeology

Since 2013, the Corsaire Concept Project has focused on the development of new tools and methods for deep underwater archaeology (from 50m to 2,000m). This project is led by the DRASSM (Département des Recherches Archéo-logiques Subaquatiques et Sous-Marines, French Ministry of Culture’s Department for Underwater Archaeology). LIRMM coordinates the robotic activities of this project and collaborates with several laboratories (Stanford Robotics, Institut PPrime, Onera DTIM, ENSTA Bretagne) and SMEs (Techno Concept, Becom-d, SIT, Copetech SM, Images Exploration). Several new robotic tools and methods have been introduced and tested under the supervision of the archaeologists.

Most of the tests are conducted at a depth of 90 meters, on the Lune shipwreck, a French warship in Louis XIV’s navy which sank in the Mediterranean Sea in 1664, a few nautical miles from Toulon, France. The robotic concepts that successfully pass the Lune tests are tested on deeper shipwrecks, like antic shipwrecks (500m deep) or the Danton battleship shipwreck (year 1916, 1,025m deep). In this project, we work on manipulation, control, vehicle coordination, and computer vision for localization and modeling. For instance, in 2014, Speedy, a LIRMM robot, demonstrated the ability to grasp fragile artefacts with the Techno Concept underwater robotic hand combined with accurate localization and control. In 2015, Speedy and Leonard, LIRMM’s underwater robots, were coordinated to operate simultaneously on the Lune shipwreck. In 2016, Leonard was used on the Lune shipwreck in collaboration with Stanford’s robotic diver Ocean One for its very first dive at sea. Since then, the two robots have become regular dive companions. Stanford’s Ocean One is the only underwater robot piloted with haptic devices, enabling the pilot to feel what the robot’s arms feel. This project is funded by ANR, CNRS PICS, Région Occitanie, FEDER, Call France-Stanford, DRASSM.