Robot spy can survive battlefield damage

[Mister Frag 0]

From http://www.newscientist.com/news/news.jsp?id=ns99994075

Quote[/b] ]Robot spy can survive battlefield damage

19:00 20 August 03

Duncan Graham-Rowe

A military reconnaissance robot being developed at a British lab can keep moving even if it gets damaged on the battlefield. When any of the snake-like robot's "muscle" segments are damaged, clever software "evolves" a different way for it to wriggle across any terrain.

The serpentine spy is a research project funded by aerospace company BAE Systems to make a low-cost military robot that can be dropped out of helicopters to carry out reconnaissance missions. Because it is not wheeled, the low-profile, ground-hugging snakebot should make a versatile battlefield spy. The team behind it has also developed a shape-changing antenna that broadcasts high-quality video and audio.

A self-healing robot has long been a dream of robotics engineers, not least because the machines are notoriously unreliable and absolutely terrible at dealing with unforeseen circumstances.

"When a dog loses a leg it's got a clever enough brain to allow it to adapt," says computer scientist Peter Bentley at University College London. But robots still lack this adaptive ability and so tend to give up the ghost when circumstances change.

Shape-memory alloy

Bentley and his colleague Siavash Haroun Mahdavi borrowed a trick from evolution to allow their robot to adapt to damage. The snakebot is made up of modular vertebral units that "snap" together to form a snake-like body (see graphic).

Each unit contains three separate "muscles" running down its length. The muscles are made out of wires of a shape-memory alloy called nitinol, an alloy of nickel and titanium whose crystal structure shrinks when an electric current is applied to it. Usefully, it regains its original shape and length once the current is removed.

To make the snakebot move in a particular direction, a current is applied to certain wires. When the current is removed, the wires spring back and the robot will jump forward.

The software for making a robot wriggle like a snake is fairly straightforward. But ensuring that the snake will keep moving even if a segment is damaged is trickier, and relies on different segments taking over from the damaged ones.

Genetic algorithm

So Bentley and Mahdavi have created a genetic algorithm (GA) - a software routine that takes a "survival of the fittest" approach - to produce a system that continually evolves to improve itself.

The program starts off with a population of 20 digital chromosomes, with each consisting of an initially random binary digit that corresponds to a muscle wire - where a 1 represents its activation and a 0 its deactivation. Each of these chromosomes forms the basis of a series of movements in the robot.

"You end up with a cyclic pattern of muscle activation," says Bentley. Some may result in the robot moving and some will not. The GA tries them all out and awards them a fitness rating, depending on how far it makes the snake move.

The two best chromosomes are then saved, the remainder are mixed up or randomly mutated and the process is repeated. After a number of generations, the amount of improvement finally tends to taper off, says Mahdavi, indicating that the GA has reached a performance plateau.

Once the robot was mobile, the team disabled some of its segments to see if it could adapt to injury. Initially it was immobilised, says Bentley, but as the GA continued to try to improve the locomotion, it gradually worked out how to move again, albeit more awkwardly and with an ungainly, dragging gait - but it was still good enough to get the robot to its destination.

19:00 20 August 03

[toadlife 3]

very cool.

[Renagade 0]

That reeminds me  of the funny nature program  that was on over here a while  back    that had a robot shark sent out to observe other sharks which  i missed  

I think u can buy those muscle wires urself if u wanted to go off making some crazy droid from Maplin but i can`t find them on the website   .In the catalog it also says that when they`re heated they shrink so the little snake better stay out of the sun

Heres some random  links i found

http://www.nitinol.com/

http://www.imagesco.com/catalog/nitinol/nitinol.html

I wonder how long it will be before we get self healing tinfoil  

[denoir 0]

Hehe, shape-memory wires. Good idea. I saw a demonstration of shape-memory metal a couple of years ago. Basically it's a special alloy that can be "programmed" to have two different forms at two different temperatures. So by running a current through it, they could get it to make it bend and thus moving the snake. It would not be fast as the metal would have to cool off to return to its other form, but it's a nice idea.

[Guest]

Yep...very neat. Nice bit of software too....now if OFP AI were that smart....

[Shashman 0]

There's a shirt made of these 'shape memory wires'...This means you can crumple it up, fold it up and stuff it into a bag, and all you have to do to 'iron' it, is put it on

At more than 2000 Pounds (currency not weight ) it could be regarded as 'haute couture' I'd say

[Mister Frag 0]

In the catalog it also says that when they`re heated they shrink so the little snake better stay out of the sun

I would think that this is precisely the kind of stuff that the genetic algorithm would deal with -- it should allow the worm to overcome such contraction.

[Renagade 0]

If they changegd the shape of it it could probably work underwater too like if they flattened it and added extra commands to the microcontroller so it was on the flat side on land then flipped 90 degrees to swim in water.

[Mister Frag 0]

It seems that it doesn't have a 'side' to it per se and can work in any orientation. It's not going to start flopping around helplessly if you roll it 90 degress to the side.

Speaking of using robots in water, the United States has already tested robotic crabs that are intended to search for mines in coastal waters and along beaches. They work in large numbers and are able to communicate with each other to maximize coverage and avoid obstacles -- I'll see if I can find an article on that.

[denoir 0]

In the catalog it also says that when they`re heated they shrink so the little snake better stay out of the sun

I would think that this is precisely the kind of stuff that the genetic algorithm would deal with -- it should allow the worm to overcome such contraction.

Yes, it probably has a temperature sensor and it was probably trained under different temperature conditions.  As for online training and adaptation once it is deployed, GA has very limited practical use.  Genetic algorithms are good for finding the optimal solution to optimization problems but they are very inefficient and slow.  Basically it starts with a set of random actions and then evaluates how good they are. It then chooses the best actions, combines them and adds mutation so that it has a new set of actions that it has to try out and evaluate. As you can imagine this is a very time-consuming process since all the actions actually have to be performed in real-time.  Not to mention that many actions can lead the system into a destructive state or simply in a state from which it can't recover.  For genetical optimization to work the system must be returned into an initial state before it tries to evaluate a new action.

So, while in theory on-site damage adaption is possible in practice its very unlikely to succeed.

I am actually fairly surprised that they decided to use GA to solve the problem. Genetic algorithms are superb at solving optimization problems for very specific situations. They have however none whatsoever capabilities of generalization and noise reduction. While the basic locomotion patents are suitable to get through genetic optimization, they should have trained a neural network or some other adaptive distributed system to be the actual controller. Such systems, unlike GA, have very powerful generalization capabilities, are very noise tolerant and their distributed nature makes them work even if they're severely damaged.

[Major Fubar 0]

Anyone ever see footage of that tank-killer bot?

Basically it was a wheeled bot with cameras onboard, that searched for tanks. When it finds one it has a big explosive charge onboard behind a big slab of solid copper. I think the test was remote controlled, but the eventual plan was to have it totally AI, able to search out targets on its own.

When the explosive is discharged (the bot was one use I think) this big-ass hunk of copper is blasted with incredible force at the tank. It totally shredded the test tank it hit...punched a massive hole straight through...

------

Anyway, we all know who the ultimate battle bot is...RAZER!

[Shashman 0]

Oh puhhhh-leeeaase!

Robot Wars is so laaaame!

But I gotta agree I love the way Razer just slices through its opposition

[mr burns 132]

omg, i was so happy to see RobotWars banned from German TV

Some of the new inventions in MilitaryTec remind me of C&C:RedAlert (damn i hated those Robot Spiders).

Anyone heard of the self healing minefield ?

Good info, but German

English page #1

English Page #2

I imagine it´s funny to see mines hopping over the battlefield to their new positions..but it´s also frightening cause these mines are already some sort of RobotSpiders.

[Mister Frag 0]

Major Fubar, that sounds like a self-propelled AT mine.

Now all the tank has to do is be on a paved surface and hold still for a few minutes so that the robot can get under it...

[Leveler 0]

Good design, although not the most clever or practical out there. Id rather see creative minds working on something else than war etc etc...

ps: Robot wars! YUCK

[The Frenchman 0]

Id rather see creative minds working on something else than war etc etc...

Here here!