|
 Science fiction, especially science fiction movies, have given us many examples of robots of the future. C-3PO and R2-D2 (Star Wars), Johnny 5 (Short-Circuit), T-800 (Terminator series), Wall-E and Eve (Wall-E), and the list goes on. These examples have molded our perceptions of what advanced robots "should" be like. They should be dexterous - able to lift, grab, hold and manipulate objects; they should be mobile - able to move at human-like speeds (or faster!) and along various grades and terrain types; and they should be autonomous - largely able to carry out tasks on their own.
 Or should they? Manuela Veloso, a professor at Carnegie Mellon, has challenged the expectation that robots be largely independent in order to be deemed useful. Her robot receptionist, called a Cobot, looks like little more than a laptop PC and a bunch of sensors on a barstool with wheels. But this humble machine, despite having no arms and a limited vocabulary, can effectively navigate the halls and elevators of the computer science department with just a little bit of help from people.
This "help" may be in the form of pushing an elevator button, placing or removing a package, or removing an obstacle in its path. Whenever a Cobot needs this help, it simply asks the nearest person, or, if nobody is nearby, it sends an email to office staff.
The concept behind human-assistance is simple: to remove the most difficult barriers to programming and designing robots. It may be relatively easy to design a robot to perform a certain task (e.g. carry something). But designing the robot to independently recognize, grab, and manipulate things with any precision is very challenging. The same goes for comprehensive speech recognition or response.
Consider a task like lifting heavy objects off the floor and moving them somewhere. It would be feasible to build a robot with the strength and mobility to perform this repetitive task all day long. But designing and programming a robot with the ability to see, and find the objects on its own would be a much more complex (and expensive) endeavor. One solution would be to design a control that allows a person to drive and/or place the robot into position when needed.
 Robot assistance is already commonplace in many industrial settings, but the hope of the human-assisted robot concept is that robots will eventually be accepted in more day-to-day applications. iRobot (makers of the Roomba, pictured left) is one company trying to bring robots into the home. Veloso argues that "...if people embraced robots with limitations we would have them in our homes as we speak." Certainly Veloso's Cobots have proved their worth in providing thousands of hours of useful service.
It will be interesting to see where the development of the human-robot interface will take us. We may not ever have a true C-3PO in our midst (and maybe that's a good thing, if you're afraid of an eventual robots-controlling-humans doomsday scenario...). But regardless, that shouldn't stop us from utilizing robots where we can, and helping them help us.
| 
Re: Should Robots Require Our Help?
