This is National Robotics Week. You didn't know that? Don't feel bad. It seems to be passing unnoticed by the mass media, too. I wouldn't have known about it if it weren't for my contacts in the high-tech business-to-business media.
It's funny that a submersible robot (automated submersible vehicle, or ASV) is mentioned in nearly every article about the search for Malaysia Airlines Flight 370, which went down in the Indian Ocean about a month ago. You'd think that would trigger something in the mass media, but no.
One trend that does seem to be getting some notice is the effort to make robots more humanoid, and National Robotics Week seems like an appropriate time for me to weigh in on that subject again. Folks who've been following my comments about technology development over the past couple of years are aware that I'm not a big fan of anthropomorphic robots, or androids. That said, I entirely support efforts to develop dual-arm robots, which are a little bit anthropomorphic.
Let me explain:
Androids are robots made to look and act as much like humans as possible. The theory behind them is that human morphology is highly adaptable -- able to jump in to perform almost any task. So, taking that page from Nature's notebook, it seems it would pay to copy human morphology to produce the most hightly adaptable general-purpose robot possible.
The problem I have with that is it amounts to reinventing the wheel. We've already got six or seven billion examples of the most highly adaptable general-purpose mobile automated systems possible. We call them "people." In fact, entirely too many of the things are going spare on the unemployment roles, already. We don't need to knock ourselves out trying to develop artificial ones.
What we need are special purpose automated systems designed specifically to do specific tasks. One of those tasks is assembly of all the bits and pieces that make things work out better for those six or seven billion general-purpose mobile automated systems (people).
The basic paradigm for an assembly task is "insert tab A into slot B." Generally, that requires two hands. Assembly robots usually have only one. We'd like that to change.
Assembly robots usually have only one hand because coordinating the actions of two hands is extraordinarily difficult. It's well beyond the capacity of most computers traditionally employed as controllers for automation systems.
To work, it really requires three computers, each with its own separate talents and responsibilities. First, of course, you need two computers, each to guide the motion of one of the arms. Then, you need a machine-vision system (with its own separate computer) to keep tabs on where the separate arms (hands are just end-effectors on the ends of arms) are, and what they're doing. Actually, you need a fourth computer to supervise and coordinate the other three computers.
A number of robotics companies have taken on the challenge. Notably, ABB Robotics last month hosted a Technology Days event at their North American headquarters in Auburn Hills, MI at which they showed off their FRIDA dual-arm concept robot for assembly-line work.
Assembly lines have been at the heart of the debate over technology's role in society practically from day one. Fritz Lang's 1927 film Metropolis dealt with the dehumanizing effects of assembly line work. Let's face it, mass manufacturing tasks on an assembly line are dull, dirty and dangerous, making assembly line work the poster child for tasks crying out for automation. We want robots to do it, and we want humans to go do something else. Anything else!
So, building dual-arm robots with an aptitude for putting tab A into slot B is a very, very good thing.