by Jade Le Maître, Innoecho

The OceanOne diving robot was unveiled to the public during the grand opening of the Mémoire à la Mer (“Memory of the Sea”) exhibition celebrating the 50th anniversary of French underwater archeology. It all began in 1966 when André Malraux created the Underwater Archeology Division, when later became the Department of Subaquatic and Underwater Research, or DRASSM for those in the know.

In 50 years so many treasures have been found and so many advances made in the way of carrying out excavations! While we’re on the topic, you really should see this new exhibition at the Musée d’Histoire de Marseille, set up in record time (16 months to choose 50 objects out of over 5,000 available, and get them to tell the story of the great discoveries of France’s underwater archeology missions).

The latest innovation at DRASSM is a humanoid robot developed through an international collaboration between DRASSM, CNRS-LIRMM and Stanford University – a trio of prestigious names that have given us a major world’s first, the OceanOne underwater archeological robot.

OceanOne is a humanoid robot that can dive to depths of 1,000 meters, while being teleoperated from the surface. Its first official mission? To explore the wreckage of la Lune (‘the Moon’), King Louis XIV’s ship, that lies 100 meters down on the floor of the Mediterranean.

Underwater robotics is at a level of technical prowess just under that of space robotics: the enormous constraints give rise to feats and breakthroughs that will be part of our everyday lives in just a few short years.

OceanOne is not a robot that will replace archeologists.

As André Malraux rightly wished when he founded the DRASSM, it’s better to have archeologists that know how to dive than divers who obey the orders of archeologists that stay at the surface. In the same vein, as the human body cannot withstand underwater depths for long periods of time, the idea is to have a robot that can compensate for the shortcomings of the human body while taking advantage of Man’s intelligence: a teleoperated robot with a human-like appearance, as it is simpler to guide an object that resembles us (with a torso, arms, hands and stereoscopic vision thanks to two eyes). Its only limitation? Its hands have only three fingers – as Michel L’Hour indicated during the press conference, if nature gave humans 5 fingers, it can’t be stupid. Fortunately, Vincent Creuze of LIRMM is just about to give the robot 5…

As a result, this robot has not been designed to replace archeologists, but rather to enable them to focus on their true value added. Research teams wish to give the humanoid greater autonomy, so that the archeologist piloting the device does not have to worry about the way it has to move, but can rather focus on the vestiges themselves. As archeological relics are fragile, it was important to endow the robot with a sense of touch via a haptic interface, which will also enable the archeologist to better feel the material the vestige is made of. OceanOne remains continually connected to the surface via an umbilical cord that transfers all of its data, ensuring both robot and research team latency-free command. To prevent the cord from getting tangled up, the LIRMM’s autonomous submarine robot can be called to the rescue.

The OceanOne team knows how to take risks, like that of transporting the robot from Stanford to Marseille to have it dive onto the la Lune’s wreckage without ever having tested the device in salt water – a crazy level of trust in the Stanford team’s know-how and a tangible amount of stress on board the boat during its first dive in the Mediterranean. From simulated dives in a swimming pool to the Mediterranean in less than a year: what an incredible job they have done! The best part of it all is that after several dives, the robot needed no repairs, even after getting its arm stuck in one of la Lune’s canons.

OceanOne is a great piece of history for French research, sparked by the meeting between archeologists and roboticists.

When I gave my presentation at GDR Robotique, one man told me that a third of what I was asking for had already been done, another third was feasible and the final third was completely crazy. It’s the last third that he was really interested in, and he wanted to work with me. That man was Vincent Creuze! — Michel L’Hour

This meeting took place during a work session of GDR-Robotique (the organization that leads the French robotics research community, bringing together over 60 research teams and 1,000 researchers) organized by Philippe Bidaud, who was then Director of GdR, at ONERA in January 2014.

This work session was designed to review research conducted in France in the field of Underwater Robotics and in Dexterous Handling, with talks by Michel L’Hour (Director of DRASSM)), Vincent Creuze, (Lecturer at LIRMM and leader of the Marine and Underwater Robotics focus area of GDR Robotique) and Oussama Khatib, Professor at Stanford University, invited for the occasion to give a presentation of his work on Humanoid Robotics.

From left to right: Oussama Khatib, Vincent Creuze, Michel L’Hour (Photo: Philippe Bidaud)

For the first time, Oussama Khatib was coming back to the office where he had worked as a PhD student at ONERA 30 years earlier. He discussed his ongoing projects with Michel L’Hour and Vincent Creuze. The 3 colleagues realized they were more or less working on the same project: it was high time to pool their efforts! Thus began the international collaboration, and the three men would work together on this underwater archeology robot project!

During the first Etats Généraux de la Robotique (‘French Robotics General Assembly‘) held at the Innorobo robotics trade show in March 2014, Michel L’Hour, Director of DRASSM, took part in a roundtable discussion on “Fields of French Leadership”. He took advantage of the occasion to make the collaboration official: unable to find a “commercial robot” able to help his teams, he said he planned to join forces with French laboratories to develop his underwater archeology robot.

An underwater archeology mission has highlighted Man’s underuse of the sea floor due to the limits of human diving. Robots could explore these depths. The challenge is to build a robot for hostile environments (corrosive salt water, high pressure, etc.)

Discovery report on the French Robotics General Assembly 2014 – ARDI Rhône-Alpes


A year later, Oussama Khatib was on hand at Innorobo 2015, in a highly interactive conference, demonstrating the future OceanOne robot… albeit via a simulation only. This conference was given on 2 July, which means that in less than a year, research teams had assembled, tested and approved a fully operational robot, capable of diving to depths of 1,000 meters and working there effectively.

Such a feat required mobilizing several research teams at Stanford as well. The haptic technology, for example, was created by Force Dimension, a company founded by one of Oussama Khatib’s former PhD students.

Like with any great story, there is emotion involved: the first object that OceanOne brought to the surface from the la Lune shipwreck was a small 17th-century Catalan jar – the most commonly found object at this depth, an item so commonplace for archeologists that such a discovery stirs little, if any, interest. Yet, for a roboticist, it’s another story! The meeting of two epochs – the first archeological vestige brought up by a robot – sparked true emotion for Oussama Khatib and his team, the result of a year’s work in a race against the clock. The sensation was even stronger as, thanks to the haptic interface, the operator could actually feel what the robot was doing. Thus, the robot becomes an avatar and, in the end, it was a human being, inside the robot, that brought this little 17th-century Catalan jar up to the surface.

One little breaststroke for Man, one giant crawl stroke for Robotics.


Leave a Reply