In November 1664, La Lune , the flagship of the French King Louis XIV’s navy, sank 30 kilometers from Toulon, in southeastern France. 352 years later, OceanOne, a robotic humanoid created at Stanford University (USA), has dived among the ship’s treasures and objects.
The android, piloted from a ship by Oussama Khatib, a computer professor at Stanford, picked up a vase, felt its contours, weighed its size, and inserted a finger through its mouth to get a better grip. He dived into a basket, deposited it there, and closed the lid of the container before returning to the surface, where Khatib and a dozen archaeologists and engineers waited anxiously.
The expedition to the site of the sinking of La Lune has been the first mission of this diving robot with integrated cameras, an electronic brain that takes advantage of the latest advances in Artificial Intelligence and a haptic control system: the sensors integrated in its hands send information to the joysticks with which the robot is operated, so that its pilot feels the same as the machine; for example, if you hold something heavy and solid, or light and fragile. And that’s just the first step, because soon there will be sensors on each of your fingers.
The robotic brain processes information from cameras and sensors and manages to manage its strength: it is able to firmly grasp a delicate object, without smashing it. This ability makes OceanOne a perfect ally for recovering parts from shipwrecks, working with living things such as coral, or depositing sensors in hard-to-reach places on the seabed.
Virtual diver
According to Professor Khatib, OceanOne “will be like an avatar. The intention is to use it to dive virtually and safely. You can precisely feel what the robot is doing and feel almost as if you are there; its touch sensors create a new dimension of perception “.
The head of this artificial Cousteau integrates stereoscopic vision cameras (like that of the human being) that show the pilot exactly what the robot sees. It also has two articulated arms, and in the rear it carries batteries, computers and eight multidirectional thrusters. But it is his hands that distinguish him and make him such a special ingenuity.
Each of his fully articulated wrists is equipped with force sensors that transmit haptic information to the pilot’s controls. That is to say, that he feels in his hands how the robot is holding.
An independent machine
The pilot is in command whenever he wishes, but OceanOne manages well alone. Sensors spread across its body pick up currents and turbulence, and automatically activate thrusters so the robot stays where it should. When moving, motors adjust the movement of the arms so that the hands are balanced while they work.
Navigation is based on the perception of the environment through sensors and cameras, whose data is processed with algorithms that prevent the android from colliding with something. Its sophistication is such that if it detects that its thrusters are not braking enough, it protects itself from the impending impact with the arms.
OceanOne, created at Stanford University with the help of the private company Meka Robotics and the King Abdullah University of Science and Technology (Saudi Arabia), was born to replace humans in dangerous underwater tasks (mining, pipeline maintenance, rescue …) or that challenge our physical ability. Still in the process of improvement, it can also be of great help in scientific research and archeology.
As Khatib says, “We create a connection between human and robot. The person brings intuition, experience and cognitive abilities; and the android goes where the person cannot, takes them to places that are too dangerous for them. It is an amazing synergy.”
Photo: Stanford University
