Robots at Work

To accomplish all these tasks and functions, robots must be able to move, maneuver, see, hear, and analyze–or “think through”–their tasks. The goal in industrial, consumer, and scientific applications of robots is to create machines that imitate human movements. But every task–whether large-scale and dependent on brute strength, or small-scale and delicate–is the result of a complicated series of computer commands, electronic impulses, and mechanical responses.

The kind of precision work that gets a specific part onto the assembly line exactly when and where it is needed increasingly requires more and more sophisticated optical and computer systems. These systems must handle the increased flow of information and assist the robot in processing and making use of it. Robots need to hear as well as see; they are even beginning to speak. They interpret spoken words and commands with sophisticated voice recognition programs and respond by performing the requested tasks. Sometimes they respond with preprogrammed statements in voices that are becoming increasingly difficult to distinguish from those of humans.

For most robots, “thinking” involves a fairly simple and repetitive task, like automated welding on an auto production line or moving merchandise around a warehouse. Increasingly, however, there is a need for smarter robots, outfitted with enough processing power to perform more complex functions.

Engineers use the example of tying a shoelace to explain the enormous complexity of teaching a machine to perform even relatively simple actions. Imagine you are explaining to a friend, who is blindfolded and holding a pair of pliers in each hand, how to tie his shoelace. Since he's blindfolded, you can't simply tell him to pick up the ends of the shoelace with the pliers. You first have to instruct him in every single, simple motion required to locate the laces by exact coordinates on a three-dimensional grid, then each motion required to open the pliers and where along the length of each lace to pick up the lace, how high to lift each end, the degree of the arc that your friend's one hand has to make on the three-dimensional grid in order to cross the laces, then the distance along the length of each end of the lace where they should intersect–the process would be endless.

But that's precisely what robotics engineers are attempting to do–not just for a single movement or task, but also for all the tasks performed, consciously or otherwise, by humans. These can range from grasping a hammer or holding an egg without breaking it to walking up a flight of stairs or running through a crowded space without collisions. Many large companies are investing billions of dollars in robotics research, some attempting to create lifelike robots like Honda's Asimo, Fujitsu's HOAP-1, and Sony's QRIO. Each of these robots can, among other things, walk upright on two legs, mimic the human range of motion, navigate autonomously, and avoid obstacles.