The standard definition of a robot is an electromechanical device that works automatically. But an artificial jellyfish created by Harvard and Caltech researchers goes a lot further than biomimicry using electromechanical means. It's more like an android: It looks like a real creature, moves like one, and incorporates living cardiac muscle cells.
The artificial jellyfish, dubbed "Medusoid," has cultured rat heart muscle cells that produce the pumping action that propels the artificial creature's silicone muscle structure through water.
Made of silicone and rat heart cells, the Medusoid engineered jellyfish's muscles contract like a real jellyfish when placed in liquid and shocked. (Source: California Institute of Technology/Harvard University)
"As engineers, we are very comfortable with building things out of steel, copper, concrete," said co-researcher Kevin Kit Parker, professor of bioengineering and applied physics at Harvard's School of Engineering and Applied Sciences, in a Harvard Gazette article describing the project. "I think of cells as another kind of building substrate, but we need rigorous quantitative design specs to move tissue engineering to a reproducible type of engineering."
Although this graceful, squishy, robot jellyfish's movements aren't nearly as elegant as those of Festo's AirJelly, they make the action of the silicone robot we reported on that mimics its surrounding look crude in comparison.
Parker, an authority on cell- and tissue-powered actuators, collaborated with Janna Nawroth, a Caltech doctoral student in biology, to reverse engineer the movements of a natural Medusa jellyfish. Nawroth’s adviser, John Dabiri, a professor of aeronautics and bioengineering at Caltech, was consulted for his expertise in biological propulsion. The researchers published their work in an article in Nature Biotechnology (subscription or payment required).
The researchers say that a major goal of creating Medusoid was to advance biological tissue engineering. To date, many of these efforts have focused on copying a tissue or organ without considering the relationship between the components and their function, or analyzing which materials would best suit that function.
Since jellyfish use their muscles to pump their way through the water, and their basic structure is similar to that of a beating human heart, the researchers decided to reverse-engineer that function to advance heart tissue research.
After mapping the alignment of subcellular protein networks in the Medusa jellyfish's muscle cells, they studied the propulsion system's electrophysiological triggering and the propulsive stroke's biomechanics. The team found that a sheet of cultured rat heart muscle tissue contracted when electrically stimulated in liquid. They used a silicone polymer to make the artificial Medusoid's body, a thin membrane with eight armlike appendages, and matched the subcellular, cellular, and supracellular jellyfish muscle architecture with the rat heart muscle cells. When the researchers placed Medusoid in a container of salt water and shocked it, the device began swimming with synchronized muscle contractions.
Is it an android? Maybe not quite yet. The researchers’ next steps will include incorporating simple intelligence so the artificial jellyfish can respond to its environment with more advanced behaviors, such as moving toward a light source, and modifying it so it can move in a particular direction.
Seems like there is some great research potential at the heart of this project. Rat heart muscle cells--curious about that one. Anything about the rat heart muscle that lends itself to this or is it more that rats are the go-to source for research?
Beth, rats are definitely one of, if not the, most common animals used in lab experiments. They are bred specifically for this purpose. And incorporating living bioengineered tissue into robots appears to be a trend. I'll be posting on this subject again soon.
Lou, I think it's unfortunate that the term "android" has been co-opted by a commercial enterprise, and not very accurately, either. Regarding the Medusoid, I agree about the control system--I'm really curious to know what they have in mind. This isn't quite a robot yet, or an android, but with the correct control system, it could be.
Although they are bred for lab experiments, I never really thought of it going much further than rats more than getting injected with drugs that are undergoing testing, or having makeup put on them (wink).
Seriously, though, putting living tissue into robots is a tad bit creepy. More and more, after reading your posts, Ann, am I beginning to understand the term uncanny valley and why it's real.
Lipstick on a rat?? Hadn't heard of that one. I agree, Jenn, it's getting creepy when we start combining engineered living tissue with machines. But also fascinating. I think that uncanny valley may be expanding into more of a continent at this point.
I agree about being creepy and fascinating at the same time. It seem's like a mad scientist movie where life is being created in a lab. I like robots with a mechanized appeal but when they start looking and acting like humans that's where I draw the line. Fascinating article.
Great coverage, and I hate to be the robot dork raining semantics down on the parade, but... This naming of robots/drones/cyborgs/androids issue is really starting to spiral into unmitigated ambiguity, so with all due respect, mitigation: this is certainly a novel robot, but I'm afraid it's not an android - the greek preference "andro," from which the word is derived, distinctly implies "man," and "oid" is of course... well, "of." Sure, meanings of specific words change over time, but this isn't one of them.
As examples, the terminator is an android. It's also a cyborg. ASIMO is an android, but not a cyborg. Both are robots. Predator and Reaper drones aren't robots, they're supertech R/C planes. So what do we call the starfish and things like it? I suppose we might just need a new standardized word for these non-mechanical artifcial moving things!
It's a complicated issue that a dictionary alone won't solve. I've addressed it a bit here: "WarBot Update: What to Call the Drones Now that They're here at Home – Suggestions?" http://goo.gl/Dxhh3
The biggest use for this that comes to mind for me is moving facial features on an android. The "muscles" used would have discrete electrical signals going to them controlled by a cpu or fpga of some sort. When you want the android to smile, particular signals are excited. When you want the android to smile really big, then those signals are excited with a greater amplitude. A frown is just different signals. This is not too far from how our faces actually work.
Obviously there are some hurdles to overcome and refinements to make to get to that point, but the basics of it can be seen in the video.
It wouldn't take much time to create a look-up table for appropriate facial actions (and store that in memory) to make an android have at least basic "emotions".
How many years before your household helping android is able to wink at you when he it cracks a joke?
Glad you liked the post, mrdon. A lot of robot R&D is starting to remind me of science fiction movies. The ones that look like people are really big in Japan, but I agree, they're too weird for my taste. DN did a survey on that subject, asking our Systems & Product Design Engineering and Automation & Control Engineering groups on LinkedIn "Should Robots Look Like People or Machines?" Here are the results: http://www.designnews.com/author.asp?section_id=1381&doc_id=237885
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.