We've reported on some research being done to help humans and robots work together by helping robots recognize human gestures. Researchers at the Massachusetts Institute of Technology have developed a different algorithm to help robots adapt to humans in factory settings by predicting what they will do next.
The researchers envision robots and humans working side by side, instead of walling off robots in metal cages while humans work on fine-tuned tasks in less dangerous areas. Robots that work in isolation perform repetitive tasks by following a set of preprogrammed instructions. Working with people is much more complicated -- it requires robots to predict each individual's actions.
The MIT team, headed by Julie Shah, Boeing Career Development assistant professor of aeronautics and astronautics, devised an algorithm that lets robots adapt quickly to individual styles and preferences for completing a certain task.
Professor Julie Shah with graduate students Ron Wilcox, center, and Matthew Gombolay coordinate human-robotic interaction while developing an algorithm that lets factory robots predict humans' movements. (Source: William Litant/MIT)
In aircraft manufacturing, human workers do a lot of walking around and picking up parts to carry to the plane. That may be good exercise, but from an efficiency standpoint, it's idle time. If robots could deliver the parts to the humans, that idle time would be reduced considerably. "It's really hard to make robots do careful refinishing tasks that people do really well," Shah said in an MIT press release. "But providing robotic assistants to do the non-value-added work can actually increase the productivity of the overall factory."
Shah leads the Interactive Robotics Group in MIT's Computer Science and Artificial Intelligence Laboratory. She chose spar assembly (building the main structural element of an aircraft's wing) as a test case. The steps involved in this process seem predictable. Two pieces of the wing are aligned. A mechanic applies sealant to predrilled holes, hammers bolts into the holes, and wipes off any excess sealant.
Ann, this is a very interesting use of robots. It is encouraging that this research is looking at ways for robots to cooperate with humans. Machines are meant to be an extension of ourselves, enabling us to do more in the same amount of time.
I imagine this kind of technology would be particularly useful and important in medical applications where the mindmeld, so to speak, between a robotic surgical tool and the actual human surgeon would ensure the best outcome from a patient standpoint.
Interesting research. There is significant work being done pursuing robots working with humans, and we've featured robots being used as "robotic assistants" in surgery. For use in the factory, I'm sure there are major challenges with safety and other concerns. Thanks.
Nice story, Ann. Yes, working with humans is tough for robots because humans are so unpredictable. Developers of autonomous cars refer to human-driven vehicles as "rogue vehicles." Some suggest that autonomous vehicles could take over the roads today if not for those unpredictable rogues.
Another fascinating story, Ann. There's seems to be a real escalation of robots research in just the last few years. It's interesting the different organizations that are supporting the research, from the military to universities and industry. It's good to see Boeing and ABB contributing to MIT's research.
The isolation described in the article is for safety. The weakness in the the technology described is safety.
The image shows a worker wearing a glove with what one can assume is transmitters which the robot can use to track the worker. Let's stipulate that Human Safety will be designed into the system from the start, and that such safety technology is accepted by the governing bodies (EN 13849). That takes into account the operator, wearing the transmitter (or RFID chip, or whatever). The operator is protected, but what about people not wearing the device?
This seems not so much taking the robot out of the cell, but putting a human inside with it. The cell would still need protective barriers (physical or light curtain) for the non-operators in the area.
The term "cell" has more than one definition in this conversation...
Robot "surgeons" are actually sophisticated, precision instruments working as an extension of the human surgeon's hands, guided by optics/machine vision. The robots in this article are standalone, separate industrial one- or two-armed robots "observing" a disconnected human. I can see this research being useful for other types of medical robots, such as assistants of various kinds. The main purpose, at present, is for assisting humans in relatively routine tasks that can yet be done in a non-routine, individualized way.
In the picture, I see the potential for huge cost savings and increased efficiency. The supervisor holding the clipboard would be trivial to robotise. This would replace the most expensive and inefficient component. As the supervisor can be programmed not to speak, there will be even greater savings in efficiency from the workers, human or robot.
Are they robots or androids? We're not exactly sure. Each talking, gesturing Geminoid looks exactly like a real individual, starting with their creator, professor Hiroshi Ishiguro of Osaka University in Japan.
Truchard will be presented the award at the 2014 Golden Mousetrap Awards ceremony during the co-located events Pacific Design & Manufacturing, MD&M West, WestPack, PLASTEC West, Electronics West, ATX West, and AeroCon.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.