Designing the Next-Gen Underwater Robot

Massive oil rigs at the middle of the ocean can bravely face its fury for extended periods of time, but the constant clash with the water does leave its rather visible marks on the steel structures. And that's why, especially in parts of the structures that are submerged, robots are of enormous value. They are able to fix damaged cables or sea floor anchors, which would otherwise break loose and allow the entire platform to be carried away by the waves.

The small robots are always controlled by human operators, who find it difficult to see anything on their screens, on account of the fact that, at the level the machines are working, the water is so black that not even the powerful lamps that are attached to the submersibles can breach it. Under these conditions, it takes hours until even the smallest damage is fixed, but it could be days before larger ones are finally taken care of.

Now, scientists are pondering the possibility of having “touch” sensors installed in these robots, which would virtually allow them to guide themselves to where the malfunction has occurred, by “feeling” their way underwater. Naturally, this would also mean that no lights or human operators would be needed.

“One component in this tactile capability is a strain gage. If the robot encounters an obstacle, the strain gage is distorted and the electrical resistance changes. The special feature of our strain gage is that it is not glued but printed on – which means we can apply the sensor to curved surfaces of the robot,” Project Manager Marcus Maiwald, from the Fraunhofer Institute for Manufacturing Technology and Applied Materials Research IFAM, in Bremen, Germany, explained.

Together with colleagues from the Institute and with experts from the German Research Center for Artificial Intelligence DFKI, at the Bremen Laboratory, Maiwald has already developed a prototype of an underwater, “tactile” robot. The printed strips of sensors that allow the machines to “see” where they're going are just about half the width of a human hair in size, or roughly ten microns. Because they are so small, strain gages can be very easily identified, and the robot could know precisely what it's touching.