The five most important robotics trends of 2011 enable volume manufacturing and aim at further integration of robotics with machine vision and automated systems. Some of the trends outlined below discuss very targeted applications, from bomb sniffing to baggage handling, but the developments in each are relevant to other, sometimes very different, types of applications. These developments concern both the design of robots themselves and the systems in which they work. (Separately, take a look at our Slideshow: Humanoid Robots Get Real.)
One of the most significant trends in robotics is the integration of multiple technologies
that will help speed up and automate the factory.
1) Integration. A key trend is the integration of multiple technologies to speed up and automate the factory. Computing Advances Permit New Integration Answers looks at how vision systems, robotics, and 3D laser scanning are being used in a wider range of applications as prices come down and performance goes up. One example is a vision system that verifies box size and the placement of labels on cases coming off a packaging line. Ten years ago, the system would have required cameras and processors costing more than $100,000, but today it can be built using a single camera for around $30,000.
2) Bomb-sniffing robots. Integration is also a major theme in military robots that sniff out bombs. These sophisticated robots integrate sensors, cameras, processors, and various feedback devices. Though they offer a great deal of autonomy and multiple capabilities, their growing complexity is increasing their weight, size, and cost. A new modular design platform, described in Simplifying the Complex in High-Risk Duty Robots, departs from onboard processors and other devices and substitutes direct user control. The robot’s path is determined via images transmitted by its wide-angle infrared camera. These and other improvements have reduced costs to the point where 10 robots can be purchased for the cost of one previous model.
There definitely is a marketing impediment to consumer robot uptake, along the lines of "What can these things actually do?" With the Roomba, it's not clear that this can vacuum your floors better than a normal vacuum. With Honda's Asimo, it's a curiousity. With the Japanese humanoid robots, well, the less said the better. I think it's going to take a robotics industry Steve Jobs to "create" a need for these things. Maybe Heather Knight is that person.
I'd love to see how the robots pull luggage. Since luggage comes is so many shapes it seem the robot would have to be very flexible. There there's the question of whether the robot might damage the luggage. Also, I would imagine the robot would need to identify the bags. Perhaps they use RFID so that line-of-sight is not a problem.
Thanks, Rob. Right now, we may all have to go to the Amsterdam airport if we want our luggage tracked correctly! Kidding aside, I, too, was surprised, in fact dumfounded, that such an enormous system works at all, regardless of it's underpinnings. And I was even more surprised that it's done with robotics. Part of the fun in investigating this area is that the technology is much more advanced than I realized. I don't think we're quite yet to the point of a War with the Machines (Terminator? Frank Herbert? Dan Simmons' AI net?), but progress in robotics is certainly proceeding apace.
The crossover I had in mind was not making industrial robots cuter or more human-looking--and I agree, Beth, a cute-looking ET-like prison guard sounds like a very scary idea.
What I do think possible is that some of the movement emulation work described in #5 could be used to influence how consumer robots move, making them even more human, and could also be used in surgical robots (larger than the one described in this article), which require extraordinary precision (if it's not already). At the same time, some work like this (motion emulation) may have already been done in surgical robots which might then translate back insto consumer robots.
The canonical example of robots run amok (that sounds like a title for a Star Trek episode) was what happened at GM under Roger Smith when they were first implemented. Of course, that's a period in the U.S. auto industry that everyone would rather forget (paging Chevy Vega). As you say, jhankwitz, things have thankfully progressed a great deal. The interesting development now is that we have solid tech progress simultaneously on TWO robotics fronts: industrial, which is relatively mature, and the newer consumer oriented robotics, like Roomba and the Japanese attempts to create humanoid-looking machines (to which I say, ick).
I was involve with ASEA Robotics way back in the late 70's. Their primary focus was getting auto manufacturers up and running with them. Their biggest problem was that the robots demanded far less variation between parts being assembled, which got them involved with advancing vision technology to compensate for some variation. It's amazing how this industry has expanded and progressed over the years.
Ann, I think there is a small amount of crossover starting to happen. One example: The VGo robot, from VGo Communications, which plays a mildly human role but does not have any human attributes. VGo Communications said that they deliberately used a non-human form, so it "wouldn't be intimidating."
Nice trend piece, Ann. I was little surprised to see luggage tracking on the list. It makes sense simply because of the complexity. I'd love to see how automatic luggage tracking is improving the process of keeping travelers moving parallel to their bags.
Being in an incubator can be analogous to shopping in a “big box retailer.” You can find many things you need under one roof along with moral support to sustain and move your startup to a successful launch.
Scientists at four major universities in Europe have released a joint paper describing the use of light to put active materials into motion and to control that motion, producing lifelike mechanisms that may or may not contain living organisms, but can produce useful work.
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.