Cutting-edge technology for underwater archaeology

Who would believe that the increasing scarcity of fish poses a threat to the three million or so shipwrecks dotting the world's seabed?

Vincent Creuze, University of Montpellier

Yet it is industrial trawling, often practised at depths of over 1,000 metres and now partially limited to 800 metres in Europe, that constitutes one of the main threats to the underwater heritage of the abyss. In the space of a few seconds, a site that is thousands of years old can be turned upside down, or even destroyed, even though it has been remarkably well-preserved until now, thanks in particular to darkness and low temperatures, far from violent tidal currents or surface weather phenomena.

The remarkable state of preservation of deep-sea wrecks makes them of major archaeological interest, but unfortunately also attracts the interest of a few private "treasure-hunting" companies, who salvage cargoes for resale, without any concern for archaeological study and often in violation of the 2001 Unesco Convention on the Protection of the Underwater Cultural Heritage.

Preserving underwater heritage

Faced with these threats, states urgently need to locate and appraise their submerged cultural heritage in order to preserve its inestimable scientific and cultural value. In recent years, the underwater archaeology departments of many countries have set out to conquer the abyss.

France, the world's second largest maritime area, is in a special position in this race. As early as 1966, André Malraux, then Minister of Culture, created the DRASSM, the Department of Underwater and Subaquatic Archaeological Research. For over 50 years, French underwater archaeologists have forged their expertise and know-how, now recognized worldwide. Until recently, most archaeological campaigns were limited to human diving, with occasional forays into deeper waters as part of oceanographic(Ifremer) or industrial(Comex, in particular) operations.

In the 2010s, given the urgent need to protect our deep-sea heritage, Michel L'Hour, underwater archaeologist and Director of DRASSM, launched an ambitious program to develop innovative resources dedicated to deep-sea archaeology. The first phase involved the design of a vessel suited to human archaeological excavation at depths of between 5 and 50 meters, but also to the deployment of robots to survey wrecks at depths of over 1,000 meters. This 36-metre vessel was launched in 2012 and rightly bears the nameAndré Malraux.

Wreck detection

Since electromagnetic waves are virtually unable to penetrate the sea, most wreck detection technologies rely on the use of acoustic waves. This is the case, for example, withsidescan sonar, an acoustic antenna towed at a distance of just a few meters from the seabed. It can be used to map the seabed and detect relief anomalies. When mapping depths beyond 300 or 400 meters, sidescan sonar is installed on an autonomous robot called an AUV (Autonomous Underwater Vehicle). Side scan sonar is often combined with a magnetometer to detect magnetic anomalies, possibly caused by the metallic masses of a wreck.

These operations, known as "surveys", are carried out either as part of preventive archaeological operations (e.g. prior to laying a submarine cable), or as part of the search for specific wrecks. For example, the wrecks of the Cordelière and the Regent, which have been lying at the mouth of the Brest Narrows since 1512, are currently the subject of an extensive search campaign. For this type of clearly identified wreck, prospecting areas are narrowed down by a meticulous study of archives (accounts from witnesses or survivors, meteorological archives, study of ocean currents, morphogeology, old nautical charts, period newspapers, etc.), as well as by analysis of oceanographic data, such as currents, tides or prevailing winds.

The appraisal phase

The survey is followed by inspection, to visually determine whether magnetic or acoustic anomalies correspond to shipwrecks. Beyond the limits of human diving, this inspection is most often carried out by a remotely operated vehicle(ROV). Thanks to its umbilical, the cable that connects it to the surface, the ROV transfers live video to the pilot on board the ship stationed above the wreck.

If the site is of interest, hundreds or even thousands of photos are taken to build a 3D model of the wreck. This technique, known as photogrammetry, has now been mastered to perfection. It can be used with a reflex camera(Ifremer, ipso facto), or with several digital cameras(Comex). The models are then used scientifically, or delivered to the general public for virtual reality visits, as DRASSM did for the wreck of the Moon or that of the battleship Dantonwhich has lain at a depth of 1025 meters since 1917.

In 2019,Onera (Office national d'études et de recherches aérospatiales), LIRMM (Laboratoire d'Informatique de robotique et de microélectronique de Montpellier) and DRASSM have designed a new single-camera system, the size of a mineral water bottle, capable of producing a 3D model in real time, while precisely calculating the robot's position. This technology facilitates piloting and speeds up interpretation of the sites visited.

Picking without breaking

When surveying a wreck, it is sometimes necessary to take samples. Existing ROVs were developed primarily for the oil industry and are unsuited to archaeological work. The slowness and poor dexterity of their hydraulic manipulator arms force pilots to place the robots on the seabed, i.e. on the wreck itself. What's more, the arms' hydraulic clamps are often incompatible with the fragility and shape of the most delicate archaeological objects (glass, wood, leather, ropes, fabrics, etc.). For these reasons, since 2012, DRASSM has been developing specialized robotic tools, with the support of LIRMM.

A laboratory site at a depth of 91 metres

Most of the tests in this program are carried out near Toulon, on the mythical wreck of the Luneat a depth of 91 meters. This ship of Louis XIV, wrecked in 1664, meets most of the constraints encountered on deep-sea wrecks, and is the ideal site for testing new robotic tools, once they have been validated in the laboratory.

So, in 2014, archaeological samples were taken using a robotized hand, designed by Techno Concept (Loupian, Hérault), and then a claw.

Thanks to a computer-assisted control algorithm developed at LIRMM, the carrier robot achieves horizontal and vertical precision of the order of 1 to 2 cm, making it possible to do away with manipulator arms and work "on the fly", without touching the bottom. In fact, it's the robot itself that pivots and moves to gently bring the claw or hand to the object to be picked. The absence of arms considerably reduces the size of the robots, enabling them to access tight or complex areas of wrecks more easily.

Building on these initial results, French archaeologists have expanded their collaborations. In 2016, the underwater humanoid Ocean Onefully developed by Professor Oussama Khatib 's team at Stanford University, made its first dive on the Moon wreck, accompanied by DRASSM and LIRMM. The robot has two innovative arms, driven quickly and precisely by electric motors and equipped with force sensors. Efforts are transmitted via haptic interfaces, a kind of motorized joystick used to control the robot in translation and rotation, similar to those used to control surgical robots.

Still in the field of manipulation, this time as part of the ANR SeaHand project, theInstitut PPrime finalized, in early 2020, a robotized hand specifically designed for underwater archaeology. Measuring the forces perceived by each finger, the SeaHand opens the way to "touch" excavation in turbid environments. This was unthinkable just a few years ago.

All these advances meet many of the needs of deep archaeology, but there are still many challenges to be met, such as the delicate removal of large volumes of sediment during a methodical excavation, or the automatic analysis of the vast quantities of data generated during survey operations.

Vincent Creuze, Lecturer in underwater robotics, University of Montpellier

This article is republished from The Conversation under a Creative Commons license. Read theoriginal article.