Science is fascinating in theory, but it can also be dreadfully dull.
Some experiments need hundreds or thousands of repetitions or tests, an exceptional opportunity to automate.
That is precisely what MIT scientists have done, building a robot that does a particular experiment, examines the results, and designs a follow-up… and has presently done so 100,000 times in the year it has been working.
The range of fluid dynamics requires a lot of complicated and unpredictable forces, and sometimes the safest way to understand them is to redo things over and over until patterns appear.
One of the remarks that need to be done is of “vortex-induced vibration,” a type of change that means a lot to create ships that travel through the water efficiently.
It includes a close inspection of an object passing through water… over, and over, and over.
It turns out it is even a super duty for a robot to take over.
What the Intelligent Tow Tank Does
Although the Intelligent Tow Tank, as they describe this robotic experimentation platform, is created not only to do the mechanical work of pulling something through the water.
But to effectively see the results, adjust the setup accordingly to seek more information, and remain to do that until it has something worth reporting.
The ITT has previously conducted around 100,000 experiments, basically making the equivalent of all of a Ph.D. student’s tests every two weeks.
The tricky part, of course, was not creating the robot, though that was undoubtedly challenging as well, but the logic that allows it to understand, at a surface level, so to talk, the currents and flows of the fluid system and plan follow-up experiments that generate useful results.
The Need of The Robot
Usually, a human (presumably a grad student) would need to observe every action — the parameters of which might be quite random — and determine how to move ahead.
However, this is rote work — not the type of thing an aspiring researcher would like to waste their time doing.
So it is a blessing that this robot, and others like it, can eventually take over the grunt task while humans center on high-level concepts and ideas.
The paper records different robots at CMU and elsewhere that have shown how automation of such work could advance.
For more updates and the latest tech news, keep reading iTMunch.