Friday, 1 August 2014

The Future of Robotics

     Arthur C. Clarke's First Law of Robotics is that a robot must never harm a human being. That would be one of the hardest "laws" to program into a machine, as it would mean teaching it everything which might possibly harm a person. On the other hand, it would be comparatively simple to give a robot a rifle - or, more likely, attach a rifle to it - and program it to shoot any person passing by.
     Most of you probably visualise a robot as a sort of man-shaped mechanical servant, capable of taking orders and doing exactly what you tell it to do. Pause for a minute and try to think it through. If you hire a human valet, you take him around and show him the lay of the house, and explain to him your requirements and timetable, and his duties. But prior to this he has probably been trained in a valet school. Most important of all, he has been born and raised in a human household, so he has learned the basics of household life - admittedly in a lower social class - until it has become second nature to him. All this would have to be programmed into a robot valet. In my previous article I pointed out that, in order to converse with a human being, a computer needs to possess a full complement of human knowledge. The same goes for a robot in order to serve a human being. But don't worry; that's not the future of robotics. Popular, run-of-the-mill science fiction has got it all wrong.
    First of all, let's establish our definitions. Robot, from the Czech word for "forced labour" is a generic term for all those mechanical servants. An android, popularly shorted to 'droid, is a type of robot which can pass, more or less, for a human being. The word means "man-like" in Greek. (Actually, "like a male person"; "like a human being" is anthropoid.) Thus, Data in Star Trek: The Next Generation is an android. Star Wars' C3PO is not, although he is close to it. R2D2 is definitely not.
     Typically, robots have been portrayed as generally humanoid in appearance, ever since the word was coined by Karel Čapek in his 1920 play, R.U.R. Apart from rendering the machine somewhat more familiar and less machine-like to viewers, it also makes it easier to be portrayed by an actor in a special suit. However, industrial robots look nothing like that. And even if robots advance to the stage where they can actually be used as domestic servants, there is absolutely no reason why they should be shaped like that.
      The human body is the result of hundreds of millions of years of evolution from organisms which looked nothing like humans. When a worm-like creature starts crawling through the primeval sludge, it makes sense for its mouth to be placed on the front end, so that it can suck up detritus as it progresses, and for its primitive sense organs to be placed in the same general region. Next, the nerve cells which control them will form a swelling in the same area. The result is a head, with its eyes, ears, nose, mouth, and brain. But there is no reason why a robot should possess a head at all; its sensors could be distributed wherever is most convenient, and the on-board computer placed somewhere in the centre.
     Likewise, there is no reason for a robot to have two legs, and every reason not to, for they are unstable, and present a balance problem. Three would be more stable, and four even better, if they could be supplied with swivel joints. With the addition of multidirectional sensors, they would allow the robot to move in any direction without turning. But even that is far too limited for future robotics.
     What is a robot? Essentially, it is a machine with an in-built computer programmed to allow it to do a specific task, or a range of tasks. Now look at the advances in computers and the miniaturisation of machinery, even nano-technology. Even now we have on the market primitive robotic pool cleaners and carpet cleaners. Even now airliners are heavily computerised. Before every flight the route is programmed in, continuous adjustments are made for wind strength and direction, fuel consumption, and so forth as to make the flight as fast and as economical as possible. When it comes in to land, it follows a beam sent from the airport. Airliners do not yet fly themselves, but we are getting there.
     The engines of advanced automobiles already possess little computers to monitor fuel consumption. They already have sensors to alert drivers of objects to the rear when reversing, and some have automatic parking functions. With the proliferation of sensors, it will not be too difficult for cars to automatically maintain correct spacing on the roads, and to slow down and stop when an obstacle suddenly appears in front of it. GPSs now automatically alert you to changes in the speed limit. It won't be too long before it will be possible to code in the required address, press the ON button, then sit back as the car drives itself. The car will not be driven by a robot; the car will be a robot. Indeed, at the time of writing (2014) Google is already testing the fully automatic driverless car. Science has caught up with science fiction.
     In a similar manner, the fully automatic house is far more likely to eventuate than Rosie the Robot doing the housework in the Jetsons' home. The future of robotics lies, not in large humanoid, multi-purpose robots, but the vast proliferation of everyday intelligent machines with on-board mini-computers.
    In fact, we already have a glimpse into a futuristic high-technology society: UFOs. Quite a few people have been aboard them - in fact, abducted - and many remember the experience without any use of hypnosis. (I intend to write an article about that in a separate blog.) Those extraterrestrials do not appear to use large general purpose robots in the sense that pulp science fiction predicts.
     In my final article, I shall depart from zoological issues and remind people that science fiction writers often forget the limitations of human nature. Otherwise, you may wish to return to the Index.