First step to partially biologic vehicles?

According to an article under www.wissenschaft.de scientists have managed to make a brain-like net of 300,000 neurons learn by the activities of a robot.

They let them grow on an array of 80 electrodes. The signals of the neurons make a robot roll within a wooden box. Each time the robot hits an obstacle it sends a signal back to the net of neurons.

The net of neurons indeed learns this way and intensifies some connections and links between neurons.

The article refers to Kevin Warwick and Ben Whalley, University of Reading, and to New Scientist ( www.newscientist.com ), 16th of August, page 22

What about applying this to small experimental space vehicles. After this has matured it might be applied to a manned rocket finally - where the net of neurons could receive commands and instructions of humans.



Dipl.-Volkswirt (bdvb) Augustin (Political Economist)

Nice video here
http://news.bbc.co.uk/1/hi/technology/7559150.stm

For unmanned craft; yes. For manned craft, however, we'll probably stick to the old method of brain to computer interface: via a computer screen and hands.

Hello, Terraformer,

regarding this point I explicitly need to disagree, because the brain-like net of 300,000 neurons seems to be very specialised while the human brain is not.

The brains of humans never should be constrained to that special activities since they are are capable of a universe of activities that are required, healthy or simply making it all worth the "pain". Because of this Humans and their brains allways should be supervisors for such nets and their doings.

And there are three more reasons why I think so:

1. The net is busy only controlling special single equipments - there can be other such simple nets doing other tasks. Since those tasks are routine humans don't need to pay permanent attention to them.

2. Humans and their brains are extremely flexible regarding changing tasks whle those simple nets are not. So humans should do things where the tasks are changing often or/and intensively.

3. In between I read about it under www.welt.de also and there it is said that three such nets exist the robot can be switched between - and they are very different in the level of activity at least. And that without any impact or influence of humans. So it may be required to supervise all such nets indeed and simply to switch between the nets - which humans would have to do.



What about that?



Dipl.-Volkswirt (bdvb) Augustin (Political Economist)

Most spacecraft are not engaged in tasks that require learning. Nearly everything that will happen to them is determined on launch day.

Rovers, mining equipment etc however may all benefit from grafted neural nets.

According to a recent article under www.wissenschaft.de scientists have constructed electrical switches using neurons of brains. It is a logical switch computers or CPUs apply. The neurons were human.

At the moment this makes me think of a vehicle being biologic in so far as at least one of its computers might apply such neurons.


...



Which thoughts?



Dipl.-Volkswirt (bdvb) Augustin (Political Economist)

I think biological computers will be the oposite of what is needed in space because of their need of nutrients.
Because of their sensitivity of radiation
Because you cant turn them off to save energy
etc.
Because you cant make a reliable backup.

just my 5 ¤