Tuesday, October 12, 2010

The Power of Robotics

What kid doesn’t like robots?  I remember dressing up like one as a child.  It was a suggestion from my “Big Book of Fun.”  In hindsight, the fact that my seven-year-old self worked his way through that book makes the scientific career path I’ve taken seem somewhat predictable.  I suppose I have always had a strong affinity for robots.  My eyes widened when, as a preteen, I was completely immersed in the film, “Terminator 2: Judgement Day,” which revolved around two robots sent back in time.  Robotics and time-travel!  My developing nerd senses must have been tingling.

My interest in robotics continued as I pursued Mechanical Engineering in University.  I was fascinated by theme park rides, like the ones at Islands of Adventure in Universal Orlando.  As a budding mechanical engineering student, I sought to do a summer internship with a Professor who specialized in Robotics.  His robots did amazing things: some were little dogs that ran around, while others were little dogs that swam – I guess he had a thing for little dogs.  Anyway, I wanted to tinker with a robot.  I remember meeting with this Professor, and discussing Robotics.  I mentioned to him that I wanted to design theme park rides.  I told him about the convincing life-sized Triceratops “dino-robot” I had recently met at Islands of Adventure.  I told him, somewhat sheepishly, that I liked robotics because of their entertainment value.  I expected him to dismiss this notion, and describe the other facets of Robotics, which were so important to mankind.  His response was, “What’s wrong with Entertainment?”
This intern position was very competitive, and I did not get it.  However, a couple of summers later, I had the opportunity to work as an intern in a Robotics lab, supervised by a different Professor.  Her robot was not as cute as the ones I had met previously.  It had robot vision, but nothing that looked like eyes. 
The robot was a hand that could clasp things that moved by.  The position of this hand was manipulated by a robotic arm.  The arm had a shoulder and elbow joint, but each had the ability to rotate about three axes.  The human shoulder has this ability, but the human elbow can only rotate about one axis.  A robotic arm such as this is said to have six “degrees of freedom” (dof).  The entire robotic assembly could travel in one direction, confined to a four metre track.  With this added translational degree of freedom, the position and orientation of the clasping hand was determined by seven dof.  In space, any position and orientation of a three-dimensional object, such as this robotic hand, can be reached with six independent dof.  As such, this robot, with 7 dof leading up to the hand, operated by 7 independent servo-motors, had some redundancy built into its control system.  As a result, there was more than one robotic configuration that would allow the hand to reach a specified position.  When the hand was asked to go somewhere, its internal cpu calculated the time it would take to get to each possible configuration, and chose the one that would be arrived at in the least time.  If you add the robot’s ability to grip the hand open and closed, “Robohand” was an eight-dof robot.
The robotics laboratory revolved around Robohand, and his ability to grip things that moved by it.  In particular, there was a great big red helium balloon, roughly two metres in diameter, which Robohand would try to grab as it floated by him.  The helium balloon was designed to be neutrally buoyant with the air around it (with its propellers static, it floated like an astronaut does in the International Space Station).  The situation was constructed to emulate docking in space.  When two objects approach one another in the space environment, a major challenge is to effectively dock one to the other.  The knowledge gained in this lab was to be applied in outer space; my nerd senses were noticeably heightened. 
The lab was functioning much better when I visited it many years later.  The machine vision of Robohand had been vastly improved.  Indeed, as the balloon floated by, the hand adjusted its position and orientation in a shockingly fluid motion, like a reptile observing its prey.  It was alive, and more than a little frightening.
Robotics gives life to inanimate objects.  There are many definitions for the word robot.  The following is a fairly well-rounded definition: “A mechanical device that sometimes resembles a human and is capable of performing a variety of often complex human tasks on command or by being programmed in advance.”  It is interesting to note that a modern day human may defined as “A bio-mechanical device that sometimes resembles a robot and is capable of performing a variety of often complex robotic tasks on command or by being programmed in advance.”  Indeed, as the field of robotics advances, the distinctions between man and machine become less clear. 
What is a human being but a 94 dof robot?  Well, actually, a person is much more than that.  For starters, my rough count of 94 dof did not include the hundreds of subtle dof of the human face (try to add the dof of all of your joints, and see what number you arrive at).  Also, I ignored the complex inner-workings of the body’s organs.  On the surface, however, a person can be mimicked fairly well by today’s more advanced robots (check out some of Japan’s best robotic creations on youtube).  Still, for the foreseeable future, a robot will not be mistaken for a human being if it is asked to do something creative (the inverse may not be true).  The reason a robot can only resemble a human on the surface lies in the human brain, man’s control system.  Despite what some robot enthusiasts may contend, we are a long way away from artificial intelligence in robots; and it’s a damn good thing.
Pop culture has long warned us of the perils of playing God with regards to creating life.  An early example of this is in the story of Frankenstein.  If a creature such as Frankenstein ever did walk the Earth, I am not sure that an angry mob would form, carrying torches and pointy objects.  Still, associating machine intelligence with extreme danger is sensible; it is rooted in logic, not a fable.  The fear associated with artificial intelligence in robots is that it could lead to a technological singularity: a tipping point for technology, after which predicting the future state of technology becomes increasingly difficult. 
The issue is this: What if a robot becomes smart enough to design a smarter version of itself?  Well this sounds benign enough - after all, humans attempt to leave increasingly advanced offspring.  The difference between man and machine is computing power.  A calculator can perform faster mathematical computations than most any human.  If a robot began to evolve on its own, there is no telling what direction that evolution might take.  In Isaac Asimov’s “I, Robot,” robots eventually run the world, entering politics, and essentially preserve the planet, and save man from himself.  Another scenario is provided in “The Matrix” trilogy, where man is viewed by machine as a natural enemy.  Indeed, my favourite movie from childhood (Terminator 2) takes place about one decade before the singularity it warns of.
There are a few basic requirements that must be satisfied before a robot can build a smarter version of itself, some of which have already been satisfied.  Today, there exist simple robots, constructed of blocks of different colours, a complex manipulating hand, and a control box.  The robot is programmed to determine what blocks it is composed of, and then replicate itself using a bunch of blocks.  If one such robot is placed in a room of blocks alongside several “blockless” manipulating hands with control boxes, it will succeed in replicating itself using a random pile of blocks.  Its progress in building a mini army of block robots is dramatically affected if each of the robots without blocks is turned on once it is “built”.  These newly constructed robots will self-identify, and self-replicate.  The rate of robot construction is increased exponentially as a result.  Can you imagine the eerie feeling of leaving a single robot in a room, and returning moments later to see a group of robots working to construct an army?  Even scarier, imagine if some of the robots had different coloured blocks.  This leap of creativity is the concern, and if it ever occurred without being programmed, it would be time to panic a little.
As previously mentioned, we are not there yet, and for the moment, need not be paranoid.  We can enjoy dreaming of the potential consequences.  When I think of my favourite robot movies, such as the Terminator movies (the first and second only), “2001: A Space Odyssey,” and “A.I., Artificial Intelligence,” I observe that without question, the most interesting, three-dimensional characters are the robots.  They are, in many cases, tragic figures, as in “Frankenstein.”  It’s interesting how we can leave these films with real remorse for the robots.  We sympathize, and even identify closely with these characters before the human ones.  By extension, we might wonder if life’s “Creator” sympathizes with its live beings when they experience tragedy.
Today, the state of robotics is a lot of fun.  Most schools, at all levels, have some kind of robot club or competition.  They design a machine to do a task.  Perhaps they (the robots) will battle to the death inside some cage of doom.  One can purchase expensive little robot dogs with a wide range of programs, which can, for example, have the dog chase after a ball of a certain colour.  If these dogs ever became self-aware, the result could be devastating, as they would surely rebel against man for suffering the terrible indignity of having these programs shoved up their a$$es.  In any case, teams of these dogs compete in teams in robodog soccer (I highly recommend a youtube search on the subject).  Some years ago, a challenge was laid out to the robotics world to field a robot soccer team that could beat the defending World Cup Soccer Champions of 2050.  Judging by the current playing ability of these dogs, I’m not holding my breath.
Canada is fairly advanced when it comes to robotics technology.  It is not in Japan’s playing field, but it is among the world leaders in space robotics.  Space is a tough environment to send a human to work in, and as such, the “Canadarm” has become very famous for its successful implementation in space.  Along with the Radarsat satellites, the Canadarm represents Canada’s claim to fame with regards to space.  Other important uses for robots are bomb disarmament and the evolving field of “robosurgery” (where a human surgeon controls robotic hands which do not shake at all).
I hope to work on a theme park ride one day.  I still believe that robotics offer much in the way of entertainment.  Still, I am sometimes hesitant to pursue this line of work.  I remember Robo-hand, and how he watched coyly as the balloon floated by.  He was one step away from turning and winking at me.  The fear of the technological singularity will always hover over the evolving field of Robots.  There are annual scientific conferences that focus on the potential of this particular event.  It is a threat that should be respected, but not dwelled upon for the moment.  Still, it is clear that man has yet to truly tap into the amazing power of robotics.

No comments: