Science and Robotics

Robots are sneaking into our lives faster than anyone really thought (unless you say, lived in the 70’s, you probably believed we’d have robot butlers by now).

No robot butlers just yet, but the US Military and large technology corporations are pushing the boundaries.

Legged Squad Support System (2014)
Just what every soldier has always wanted, a robot that will carry their packs! It’s mission is to “Go where dismounts go, do what dismounts do, work among dismounts”. That means it needs to be rugged, adaptable and long-lived for a robot. The software is the real innovation here, complex dynamic balancing systems mean it keeps itself stable like humans and animals do, rather than relying on rigid logic like the primitive robots of ages past. You can literally kick the LS3 in it’s side, on ice, and it will flail around wildly but not fall over. It will also be able to follow an operator at a distance, responding to visual hand signals for commands just like other soldiers do, making integration effortless.

It’s not hard to imagine the future of the LS3. How long before commanders start using them for reconnaissance in a firefight? It can handle the terrain, would only have to go a short distance, and if it gets shot where’s the problem, soldiers can simply carry their packs.

How long before it gets a weapon for it’s own defence? Considering it will already have the software to recognize human shapes (by virtue of needing to find the soldier it’s following), it wouldn’t be hard to mount a heavy machine gun, point it in a direction and tell it to brass up any human shape it see’s. This concept would obviously need some fine tuning to comply with the Laws of Armed Conflict. What if it had a mortar or fully automatic 40mm grenade launcher on it? It already has the sensors to properly adjust itself relative to the ground, which could potentially make it more accurate than any mortar plate. A camera  near the front of the battle could be all it needs to automatically adjust fire onto the target while the battle rages.

LS3 will be based on the big dog, but quieter:
http://www.youtube.com/watch?v=cNZPRsrwumQ

Solar Eagle (2013)
The solar eagle’s ultimate goal is to stay in the air for 5 years at a time.  Used much like a satellite except much closer to the earth, the Solar Eagle will save solar energy by day, and using efficient motors and a 120m glider wingspan, will fly through the night without losing altitude. While perhaps not quite as revolutionary as the LS3, the Solar Eagle will bring surveillance costs down significantly as satellite launches will no longer be required to gain an asset in the sky. While I’m not sure it would be a suitable platform to hold kinetic weapons (it’s supposed to stay up there for 5 years which would limit reloading), I wonder whether the LSR could be the first platform for sky based electronic attack? With a seriously powerful directional antenna it might one day be possible to interact with ground based WiFi signals. It’s designed to fly at 18km, but I’m assuming it could be brought down to 10km for a period. With the longest unamplified WiFi signal currently being 304km, I wouldn’t have thought a 10km link with direct line of sight would be out of the question. Obviously the recipient wouldn’t be conveniently aiming a dish toward the sky, which would make the link that much more difficult, but with the right kind of equipment who knows. Once the WPA key has been sniffed (I won’t even go into WEP..), an FPGA based super computer (a la the one just built by JP Morgan) back in the USA could conceivably crunch some seriously long WPA keys, opening the doors to institutions who previously thought themselves secure.

IBM Watson (Now)
Link: http://www.youtube.com/watch?v=xm8iUjzgPTg
This, I personally think is pretty amazing, provided no short cuts have been taken. The idea behind the Watson is it’s a natural language processor. During the Jeopardy series it had no internet connection, but it did have all of Wikipedia to reference. Watson actually understands the questions, works out what it needs to be looking for, then sifts through all the data quicker than the human competitors. I’ve said for a long time that the problem between human’s and computers isn’t their capabilities (they can do a -ton- of things better than we can), it’s their flexibility and the interface. If you were able to simply ask complex questions like ‘Computer, get me the maiden names of the wives of the ten last Australian Prime Ministers’, you’d have so much more information at your finger tips than ever before. Maybe a more useful question would be ‘Get me the US government R&D budgets for the last twenty years versus Chinese R&D budgets, sorted by administration and adjusted for inflation’. How long would it take a human to get this information out of public documents? By being able to simply ask these questions, I think it will vastly, vastly improve the number of informed decisions being made, by giving them the answers they need to confirm their intuition, fast enough to make the decision timely and without even losing their train of thought. The Watson is revolutionary.

Google and MIT Self Driving Car (2020 estimate)
Basically, Google are funding development of a car that has clocked up 140,000 miles around San Francisco driven completely by computer. It has a whole bunch of sensors around it, it knows where humans, cars, obstacles and everything are. The only accident it had was when stopped at a red light, someone rear ended it. And it can do this:

http://www.youtube.com/watch?v=t9Fxp3HK6DI

Sweet? It does pose an interesting question though; if everyone adopted these cars, and car accidents were reduced to 5% of their original number, would you take them? There were 18,000 car accidents in 2006-2007, so 5% of that would be 900 accidents, many of them fatal, caused by bugs, glitches, oversights in the logic of the software. Do you still want Google’s miraculous self-driving car? Probably not, because most people think they’re above average drivers, even though we can say with 100% certainty than half of them are below average. So then, would you rather take your chances driving the car yourself rather than risk an accident due to a software bug, even though your chances of dying may be increased by up to 2000%? Time will tell what society says.

Me, I’d take the cars, and I’d support roads that only these cars can drive on. No more drunk drivers, no more T-bone’s at intersections, no more running red lights, and you can actually do things during your commute to work, hooray! Robotic cars will actually protect you from all the other shoddy drivers, there’s anyone will agree with. But all it’s going to take is one bad accident followed by the press attention to bring the whole thing crumbling down for many, many years.

If you liked this post, please consider a Bitcoin donation. Even a few cents would be huge! My address is: 1HcgS9GCPRQ1APv2TcqZMMskDxLjBaAKsG

Kind Regards,
Nick Cooper

Comments are closed.