High Tech That Thinks For Itself

As I was growing up, robots were limited to the car production line.  Adverts abounded that showed robotic arms welding chassis with amazing speed and precision.  Slowly, though, robots have moved out of the factory and, on their caterpillar tracks, have trundled into a world where they make our lives safer by working in dangerous situations on our behalf.

Now, I am a technophile.  I love new technology, and have worked with it for my whole career.  Some may have ethical worries about what we may “allow” robots to do, but in this month’s blog I want to focus on the positives of how they allow us to do tasks we otherwise couldn’t, and what else they can do for us with the aid of artificial intelligence.

The power of robots comes from their ability to operate in hostile environments, where they can act as their handler’s remote arm. Robots can examine suspicious objects and give their operators the ability to disarm bombs without being so close that an explosion would kill them.  They can enter nuclear reactors and check for cracks or leaking radiation. Robots can work beneath the waves at depths where the pressure would crush our feeble bodies, and stay there longer than our puny lungs would permit.  They freely enter tunnels without being distracted by the risk that the roof might collapse or that gas levels could become dangerous.

Let’s explore what these beautiful bits of technology can do for us.

I was fascinated recently by an MIT press release that told of how a robot it had designed to search the walls of nuclear reactors for cracks, had been converted to an anti-smuggling robot. These football-sized machines can now swim underwater and secretly scan a ship’s hull to find the secret compartments used by smugglers.  It’s envisaged that a shoal of twenty such robots would work together to scan a complete ship’s hull without those on board ever realising.

And while we’re thinking about the fight against smugglers, other robots are being used to continuously search the maze of tunnels near the U.S border that are so frequently used to smuggle drugs into America.

TheBlaze.com reported at the start of this year that over 170 tunnels have been found beneath the U.S. – Mexico border, linking into Arizona’s storm drainage system, through which the smugglers can enter the U.S. These tunnels are frequently no larger than a man can crawl through, so securing them can be dangerous and time-consuming, thus making it a perfect application for a robot. These machines are really little more than remote-controlled cars with a camera, which feed back what they see to a Border Patrol agent at the surface.  You can watch a great video of one of these at work on theblaze.com [link].

robot bomb disposal

This is very similar technology, of- course, to that used in bomb-disposal robots such as the British Army’s Dragon Runner that was deployed in Afghanistan.  This version is far more rugged than those sent after drugs smugglers, and has arms that allow its operator to manipulate what it finds – for example to open a bag or to carry suspicious objects to a safer location for destruction. Essentially, though, they are still just remote controlled vehicles, and a long way short of the science-fiction robot that can think for itself.

Scientists are, however, working hard to give robots intelligence.  Researchers in artificial intelligence (A.I.) have been striving for over fifteen years to give robots the ability to work things out for themselves, and to allow the tether between robot and its human operator to be severed. Rather than such robots running on software that’s been written to control what the machine does when it encounters a pre-defined set of conditions, robots endowed with artificial intelligence work in a different way.

They are given many sensors, and their software allows the robot to use them to learn the effect of its actions. For example, consider a model car with ultrasonic sensors: if programmed with A.I. to have the objective of reaching the far wall of a room with barriers placed in the way; the first few times it tries to cross the space, it hits the barriers and can’t progress.  However, it soon learns that its sensor gives a high output just before it fails its mission, and realises there is a correlation.  From then on, whenever its sensor output goes high, the robot stops and tries a different route.  Before long, it is able to manoeuvre around the obstacles to reach the far wall.  Like a baby experimenting with arms and legs and fingers, it learns by itself the best way to succeed.

The advantage of a robot that can think for itself is the speed with which it can process information.  Take, for example, the automated missile defence systems that have been so successful at protecting countries from incoming rockets. A human operator has sufficient time to recognize an in-coming missile and start to say something, but can manage nothing more in that time.  A robot with A.I., however, thinks sufficiently quickly that it can identify the threat, aim its own weapons, and shoot it down before it strikes.

The downside, though, is that it is, after all, just a computer controlled robot. Some years ago, a system that was being demonstrated at a defence show incorrectly locked-on to a group of bystanders and had killed them before a human operator could de-activate the robot.


Such events, and memories of such films as Robocop and The Terminator, make the public wary of letting robots lose when they have integrated weapons. There is still a very strong pull, however, for weaponized robots to be developed that can be used in frontline fighting.

One major obstacle is of course batteries. We all know how quickly a power-pack runs flat, especially when powering powerful motors.  In order to prolong the amount of time robots can spend on a mission, scientists have been developing what are being called scavenger robots.  One research machine knows whimsically as EATR (Energetically Autonomous Tactical Robot) has been developed to re-charge itself by consuming organic matter. While out on a mission, just like a wild animal it consumes vegetation to keep itself powered, giving it the ability to roam freely without the need to return for a re-charge.

Worries exist here for its battlefield use, though. Proponents say it is programmed to identify and eat only vegetation.  Those who oppose such a system, on the other hand, are worried that it may still start of feed on fallen soldiers.  Shudder.

Technology marches on ever faster, with robots gaining more intelligence, speed, and the ability to operate with increasing freedom from their human handlers. They can outthink us; the worry is, will the ever outsmart us?


Coming soon: Eavesdrop by Ian Coates.

Ian Coates is the author of the thriller, Eavesdrop, to be published  shortly by Bad Day Books, an imprint of Assent Publishing.

Smuggling and industrial sabotage leave Customs Investigator James Winter’s career in tatters.  Who arranged his dismissal, and what’s their link to Middle East assassins?



Twitter: http://www.twitter.com/@ian_coates_

Facebook: www.facebook.com/ThrillersByIanCoates