February 2004

The Robots Are Here

Continued from page 1

By Rodney Brooks

But we're also just a couple of research advances away from even bigger growth in a whole new set of markets-growth that will look like what happened in the computer industry over the last 25 years. Take farming. Agriculture in Western Europe and North America relies on vast numbers of migrant laborers to manipulate individual plants. Polish laborers flock to Germany to push dirt up around asparagus to make the spears white. North Africans travel to Italy to pick grapes. And workers from Latin America toil on farms across the United States. All these workers use their eyes to identify the plants and their locations and their hands to manipulate them. In short, a multibillion-dollar market awaits robots that perform these kinds of tasks. Such markets will likely emerge first in Japan, where far higher labor costs make the economics more compelling. Same goes for a broad array of manufacturing jobs. Fabricating home appliances, toys, clothes, and electrical goods requires visual perception coupled with manual dexterity. This could be yet another multibillion-dollar robot market.

True, robots today are still not very good at either recognizing generic objects or readily manipulating them. But Moore's Law has been very, very helpful in chipping away at both problems. Computer vision, while still lagging far behind the average two-year-old, is getting less fuzzy. For example, programs in the lab have gotten very good at tracking motion and recognizing faces. And right now our robots are not particularly adept at grasping objects with varying sizes, shapes, and surface properties. But new sensors enabled by microelectromechanical systems and nanotechnologies-and fueled by plenty of embedded computational horsepower-make the time ripe for researchers to tackle robot dexterity, too. It will help that military funding for advanced information-systems research is likely to shift from navigation to logistics and resupply-which will require better robotic vision and dexterity.

Robots with the vision capabilities of a two-year-old and the manipulation capabilities of a six-year-old will be more disruptive to our way of life than any robot portrayed by the governor of California. They will reorder the world labor markets that have developed over the last 50 years. They will change immigration patterns and the massive shift of labor from developed to developing countries. But the most important impact might well be on elder care: caregiving robots could help us weather the tsunami of aging baby-boomers about to submerge the economies of Europe, North America, and Japan. But more on that in a later column. For now, suffice it to say: the robots are here.