Submitted by Doug Hornig of Casey's Extraordinary Technology
Robotics, Part 1 – Where Are We Today?
In the field of robotics, we have no Newton. No one who, assisted by a falling fruit, cried out, “Eureka, I have it, and it is called a… I know… a robot.”
No, the concept of a robot first occurred to some unknown person in some far distant time, as he or she, engaged in a grinding, repetitive task, dreamed of a mechanical contrivance that could do some of the dirty work. We know that moment was more than five hundred years ago, because we have sketchbooks from the incomparable Leonardo da Vinci, dated 1495, that contain detailed plans for one.
There is no hard evidence that Leonardo actually went on to build his robot. But such was his genius that, five centuries later, engineer Mark Rosheim was able to create a fully functional version faithful to the technology of the time. It works.
Leonardo believed that the human body is basically machine-like in structure, and that he could duplicate its intricate movements through the use of levers and pulleys. A rather simplistic view, of course. Nevertheless, da Vinci was able to use his remarkable knowledge of anatomy and kinetics to design a robot with the capability to walk, stand, and sit, raise its arms, move its head from side to side, and open and close its jaw.
While Leonardo’s creation is often considered the initial introduction of a robot in human form, many believe the honor should go instead to Al-Jazari, the engineering genius of the Islamic world during the Middle Ages. The Kurdish Al-Jazari designed, and apparently built in 1206, a boat that floated on a lake to entertain guests at royal drinking parties. Within the boat were four musical automata – two drummers, a harpist, and a flutist. They appear to have been somewhat programmable, although they lacked the sophisticated articulations of da Vinci’s design.
And some researchers want to push the invention of robots even further back into antiquity. But whatever the robot’s provenance turns out to be, we can date the appearance of the world's first modern humanoid robot, to 1939. Elektro, built by Westinghouse, was placed on exhibit at the 1939 World’s Fair in New York. The robot, standing a robust 6’ 9” tall, could smoke, blow up balloons, and speak more than 700 words. Elektro was still around in 1960, when it landed the role of Thinko in the movie Sex Kittens Go to College. As if that weren’t humiliating enough, it was then decapitated, with its head given to a retiring Westinghouse employee and its body sold for scrap.
Since Elektro, well, electrified audiences at the World’s Fair, the robot has entrenched itself in the popular psyche. It’s been a real love/hate relationship, too. When they’re not depicted as the kindly, human-friendly machines typified by C3PO and R2D2 in Star Wars, they tend to be cast as pure evil, like the malevolent metallic monster lurking beneath the artificial skin of the Terminator.
But so much for fiction. Even as the robot was conjuring up all sorts of creatures in movie screenwriters’ minds, the science of robotics has been progressing out in the real world. So where are we today? Short of the Terminator, but well along in other ways. Let’s have a look.
First, however, we must ask, just what is a robot, anyway? Unfortunately, there’s no consensus answer, although we generally tend to think of clanking metal things constructed more or less along the lines of a human. And most people would probably say they know one when they see one.
The primary definition, according to Webster’s Collegiate Dictionary is “a machine that looks like a human being and performs various complex acts (as walking or talking) of a human being.” Many would expand that to include any machine that does one or more of the following: move around, operate a mechanical limb, sense and manipulate its environment, and exhibit intelligent behavior – especially behavior that mimics humans or other animals. And it’s probably useful to add that a robot can be either physically or mentally anthropomorphic, if not both.
Designing a robot is a matter of thinking from brain to bones, so to speak. We’ve gotten very good at the brain part. Given the giant leaps forward in computer software development, we can write programs to do pretty much anything we want. If we have a device that can execute the instructions.
That’s where the external hardware comes in, and it is a different story. Spectacular CGI effects have become so commonplace in movies, we might be led to think that engineers are constructing the creatures from Avatar in their Bay Area laboratories. They aren’t. Physical robotics hasn’t remotely kept up with advances in CGI. We are, however, slowly closing the gap.
Let’s start with brain. Legendary computer scientist Alan Turing once proposed, in his 1950 paper Computing Machinery and Intelligence, a simple test to determine if a computer had achieved true artificial intelligence. Let any person have a conversation with both a machine and another human, both trying to act human, and see if the person can tell which respondent is the machine, if either. In order to ensure the test wasn’t hampered by the limitations of technologies at the time, he proposed both conversations happen not in person, and not via telephone with synthesized computer voice, but by a simple text chat.
As anyone who’s attempted to use one of those interactive voice telephone agents now popular with banks, airlines, and other soul-draining mega-agencies can attest to, we are far, far away from any system passing “The Turing Test.”
Researchers around the world are hard at work on trying to solve the AI problem, with mixed success. But in the meantime a whole other band of engineers are instead working on creating robots that tackle the challenge Turing chose to conveniently ignore: Can we make a machine that not just communicates via text, but that appears to act and think like a human, with all the nuance of non-verbal communication, the subtlety of speech, and the myriad other factors that (for now) separate us from the machines?
And they are making some impressive, albeit still elementary, progress. Take the Einstein robot, for instance.
This fairly realistic representation of Einstein’s late-life head was built by Hanson Robotics in Texas and programmed by scientists at the University of California, San Diego.
That they were able to get it (him?) to move and exhibit recognizable human facial expressions is remarkable enough. But the head also contains a camera, linked to software that can interpret what it sees. Thus Einstein can learn who you are, including guessing your age and gender, respond to your audio cues, and mimic your own expressions and simple gestures in a sort of faux-empathy. Give him a wireless connection to a decent computer and he could probably beat any of us at chess, not unlike his real-life inspiration. (Well, not quite yet, but he may be getting there.)
For another look down “Creepy Lane,” here’s a full-size, modestly interactive Japanese robot from the 2009 Robotec Expo in Brazil.
As well as demonstrating our present capabilities, these machines also dramatically illustrate our limitations. Giving a robot a brain – the “intelligence” to move, speak, do some rudimentary learning, play chess – is the easy part.
Faithfully recreating even a small part of the human body, that’s tough. Our structure does not readily lend itself to simulation. Muscles, bones, joints, tendons, ligaments are not a series of servos and actuators, but a collection of tens of trillions of individual biological machines all connected in highly intricate ways that allow us to laugh and cry, wrestle with our children, kick a soccer ball, and assume the lotus position. Just imagine how much is involved in the elementary act of stair climbing, as Asimo the robot discovered to his chagrin (if “he” could feel embarrassment).
Speaking of soccer, getting robots to act cooperatively, in addition to kicking a ball, might seem like an insurmountable problem given the state of robotics today. But researchers are working on it. Each year, there is a RoboCup soccer match, and each year the participants get better. The goal (so to speak) is to field a robotic team by 2050 that can defeat the human World Cup champions. For a glimpse of where we stand with that, check out some of the “action” from the 2010 RoboCup.
Neither Einstein’s head, nor the Japanese spokesrobot, nor Asimo, nor our mechanical soccer stars are likely to be confused with real life.