He obeys to the word, the new diving robot of the University of Bremen. That is to say: obey on each hand signal. This is real progress under water: because researchers want hands free have, if they work in the depths of the sea.
Seen that in Hollywood movies like Titanic James Cameron in 1997. From a boat, an underwater robot is controlled to explore the sunken wreck and to hide a treasure. Also the science away from the cinema uses diving robots, to obtain information about the nature of the seabed. Diving robots are used for example in research projects of underwater archaeology.
Direct communication under water is difficult
In such research projects, human divers explore the topology, take samples of sediments, create photos, and document the findings. The accompanying diving robot hang so far always on long cables, which allow control from a boat or from land.
Because it is still down there on the ocean floor. Direct communication between popup man and diving machine was impossible. Researchers led by Andreas Birk, Professor of electrical engineering and computer science at Jacobs University have changed, now that in Bremen in the framework of an EU project with six international partners.
Autonomous diving robot as a target
CADDY is called the research project, which promotes the EU with €3.7 million until 2017, with CADDY for “Cognitive Autonomous Diving Buddy”. Loosely translated, CADDY wants to create the autonomous diving partner. Artu is the underwater robot that can interpret gestures of his diving people and act accordingly.
Artu comes from the Italian CADDY project partner Consiglio Nazionale della Ricerche (CNR) in Rome. The diver can signal Artu for example that he can make a photo of an interesting object. He can in the command by gestures, to follow him or to show up with him. Fairly complex gestures statements are possible, the diver can signal the robot to make a map of a specific area.
Artu understands human gestures
Andreas Birk’s team has equipped the underwater robot with a stereo camera with its own processing and evaluation unit. Thus, Artu the sign language of the human Diver can understand and implement. This is not quite trivial, because the visibility under water is often less than optimal. For the software, this means that even with low-quality and sometimes even incorrect data work must.
In addition: Artu must be able process the gestures of his diver in real time and interpret the human-machine communication to work reliably. Artu must that can capture the diver with his whole body and especially his hands and identify despite the difficult underwater situation for man and machine.
This to work reliably, Bremen researchers have equipped the 3D-Systems of robot with a robust low level software components, process the images in real time, identify the relevant parts of the body of the diver, and then track.
“This opens new avenues for the underwater Robotics”
Special high level software components focus on the arrangement of the fingers, hands and arms of the diver. From this, Artu recognizes individual characters, but also longer sequences of commands of his diver. Now, Birk and Mr Artu have tested for the first time in field trials. “The field tests in Biograd were very successful” so Andreas Birk, “we are very proud that our system to detect divers gestures has passed the first tests under real-world conditions as well. This opens entirely new avenues for the underwater robotics.”