Open ended modular robotic learning system

Open ended modular robotic learning system are being developed in the UK. Evolutionary Modular Neural Nets uses their animat robots as a test bed. Basically it is way to develop some flexibility and learning in a small but manageable neural net when you are unable to evolve a larger neural net to handle all of the desired tasks.

They aim to grow a network which can control all aspects of the robot’s behaviour. This is accomplished by both growing the robot’s body and its mind (neural network) at the same time. The approach is inspired by the evolution of animals from simple to complex forms over evolutionary time periods. After all, if you want to develop a complex robot why not take the same route as biology did? Start with a simple machine and slowly build up its brain, body and environment together. In our system it is done by starting with a very simple robot and gradually adding small neural networks (modules) to its “brain” until it can function well. Then the robot (and the environment it interacts with) is allowed to become slightly more complex and the process of adding further network modules is repeated. This process continues until the robot can fulfil all its desired tasks. The previously added functionallity stays (it does not evolve further, only the newly added modules evolve) and the newer parts build up like the layers of an onion. This means that, on each iteration, the search space remains small. The idea is shown in the illustrations below.

Existing robots cannot usually cope with physical changes – the addition of a sensor or new type of limb, say – without a complete redesign of their control software, which can be time-consuming and expensive.

So artificial intelligence engineer Christopher MacLeod and his colleagues at the Robert Gordon University in Aberdeen, UK, created a robot that adapts to such changes by mimicking biological evolution. “If we want to make really complex humanoid robots with ever more sensors and more complex behaviours, it is critical that they are able to grow in complexity over time – just like biological creatures did,” he says.

As animals evolved, additions of small groups of neurons on top of existing neural structures are thought to have allowed their brain complexity to increase steadily, he says, keeping pace with the development of new limbs and senses. In the same way, Macleod’s robot’s brain assigns new clusters of “neurons” to adapt to new additions to its body.

About The Author

Add comment

E-mail is already registered on the site. Please use the Login form or enter another.

You entered an incorrect username or password

Sorry, you must be logged in to post a comment.

8 comments

by Newest
by Best by Newest by Oldest
1

It makes sense to improve on efficiency and life extension for existing and proven reactor designs. Given the amount of work needed to bring a new reactor design into service, which can take years, the fastest path to deployment will be with current technologies.
http://djysrv.blogspot.com/2008/02/idaho-lab-epri-craft-new-lwr-reactor.html

The Idaho lab is also home to work on the Generation IV R&D reactor program and plans to build the "Next Generation Nuclear Plant." Like all demonstration plants the expected costs are always higher than anticipated, but that doesn't mean they won't prove out commercially in the long run. This is what's happening with South Africa's PBMR.

http://djysrv.blogspot.com/2007/12/ngnp-costs-said-to-be-high-than.html

http://djysrv.blogspot.com/2007/12/pbmr-seeks-mitsubishi-investment-in-new.html

2

Ambient air temperature has a considerable effect on parasitic losses. Dry coolings systems in cold climates work much better than dry cooling systems in hot deserts.

Heller dry cooling systems have much less parasitic loads, and cost about the same as a wet cooling tower.

There are also more promising options for heat exchanger materials, such as carbon foam that could reduce the cost while improving the performance of dry cooling systems.

3

For the analogy of car fuel mileage the flouride fuel would be a 530 mpg car.

4

Hi Gary,

thanks. I fixed up some HTML in one of my posts and it is ok now.

You should be good either right now or with a reload.

Thanks again

Brian

5

Kirk,
Looking at your link from your own diagram I think it is 914 GWd/t your link says 914,000 MWd/t.

So that would be 10.75 times better than the 85 GWd/t.

We definitely should go for that and fund the molten salt and flouride fuel reactors. Fuji Molten salt etc... $100 million for one good demo - $1-2 billion for a crash program.

6

Hi Gary,

the site is run through blogger.com. I am not sure what is causing the issue and have limited means to address it. I will check the html of my posts and see if something is there.

Kirk.
The 85 GWd/t would be a 70% improvement for the installed base of light water reactors. If we are keeping those running for another 40 years then it would be good that they run 70% better. Just as if I am not going to junk my current car, getting it upgrade to 170% of current mileage (30mpg up to 51mpg) is good. Even if we can get new flouride cars in ten years with 30000 mpg. The plan is only talking about the current breed of nuclear reactors.

7

Brian. Your site is having tech issues. Only four articles are loading up and the side bar isn't lodaing up either.
I've restarted my comp. and everything, but it is still not working

8

b. Develop high-burnup (HBU) fuel [85 Gwd/t target]

It's pretty hard to get excited about this burnup level when uranium-233 in fluoride fuel lets you hit http://www.energyfromthorium.com/images/thoriumVsUranium.jpg" REL="nofollow">~900,000 GWd/t, the theoretical limit.