Robotic Behavior In 3D Space: The Next Advance?

I’ve been following the impact of automation of human employment for a while now. There’s a range of thinking on the matter. Some people think artificial intelligence will become so good that large amounts of the population will be left unemployed. The other end of the spectrum thinks new types of jobs based on new technologies will emerge. Thus, those displaced from redundant work will have better work to do. My thinking wavers along the spectrum, according to what I know. So, I read all I can, when I can.

I came across a study published by the consulting company, McKinsey. The gist of the study is that automation will replace jobs that involve tasks that have a high degree of repetition and limited range of motion: for example, putting a windshield on a car in an auto assembly plant or painting guitars in a musical instrument factory. Jobs based on random tasks and troubleshooting – such as hotel maids -- are harder to automate and thus will be exempt from robotic replacement, according to the report. A robot does not have the visual granularity in 3D space necessary to scan a hotel room and distinguish random occurrences of pillows and return them to the bed. Nor does a robot have the fingers, a.k.a. the end of arm tooling (EOAT), required to make the bed. Such chores are no problem for a human, thus, humans will continue to work as maids.

After reading the report, I went to the ATX Automation Technology Convention, where I had the good fortune to meet Mike Fair of Rethink Robotics. Mike and I got to talking about the McKinsey report and the problem of the chambermaid. Mike agreed that distinguishing objects in 3D space is difficult for a robot. He demonstrated by calibrating a Rethink’s robot to a tabletop plane. If the robot’s arm is not calibrated properly, instead of picking the part up off the table, the arm will try to punch through the table. Mike agreed that a critical problem to solve is making it so a robot can identify and manipulate objects in 3D space without human assistance.

automation.jpg

I was intrigued by the 3D space problem as I drove back from the convention. My thoughts wandered to the history of technological innovation. Consider this: It took mankind more than 10,000 years to get to the point when the Wright Brothers first flew an airplane in 1903. However, it took humans just another 66 years to get to the moon. The telephone followed the same trend. Alexander Graham Bell made the first call in 1876; 41 years later people made calls over the Atlantic. What if the same rate of technological achievement is applicable to robotic vision? I could not help by wonder, how much longer will it take until robotic maids cleaning hotel rooms are commonplace? Maybe Five years? Or 10 years?

The analysis made in the McKinsey report is based on “currently demonstrated technologies.”  Yet, as I’ve described above in the flight and telephone analogies, technology advances at dramatically increasing rates of acceleration. I think it’s only a matter of time before the problem of robots working unaided in 3D space is solved. Once robots can negotiate 3D space, another type of human labor, cleaning hotel rooms, will be replaced by automation.

In terms of human employment, this is either good news or bad news depending on which end of the spectrum comes to fruition—new work or no work. Either the human maid, now freed from the labor of performing redundant chores, will have the opportunity to make money using his or her knowledge of hotel rooms to design rooms that are easier to clean. Or, maids will need to find other work in order to earn a living, provided there is work to do. I leave it to you to imagine what that future will be.