Awesome magnetic robots

[mikeselectricstuff]

This is the coolest thing I've seen in a long time...
http://www.wired.co.uk/news/archive/2014-04/16/magnetic-microrobot-swarms

Would be fun to figure out how this works - couldn't immediately find the patent mentioned on the SRI site

Observations :
Robots look like they may be clusters of small, presumably neodymium magnets - especially in the first pick/place demo
The flex demo only moves in one axis, and has few connections
There is what looks like metal foil over the PCBs - is this for magnetic reasons, friction reduction or both?
The rigid PCB is sitting on a thick metal plate - magnetic, heatsinking or both?

The PCN top layer appears to have fairly long tracks - far too long for localised control, presumably acting in conjunction with a finer-grained structure on the underside.
I don't know enough about magnetism to know the significance of the "diamagnetic" stuff - I've seen diamagnetic levitation so maybe helps to reduce friction?
I wonder if maybe a lot of the movement actually is along predefined routes in the PCB, as opposed to being fully x/y capable - some of the demos seem to involve unnecessary movements, ane the very first demo seems to have a fairly uninteresting motion route
What really puzzles me is how they get the rotation

[tom66]

One of the ways you could do this is by having the robot in a continuously unfavourable position (i.e. when rotated it wants to snap back.) This would be generated by an opposing magnetic field holding each robot in the unfavourable position. Rotation is then slowly allowing the robot to go back to the position it wants to (under no power.)

[mikeselectricstuff]

How hard can it be...?

[dr.diesel]

This is the coolest thing I've seen in a long time...

Agreed, that is awesome, 

How long till there are bought by Googlenet?

[Dave Turner]

Wow!

I wonder how small they could go.

[Christe4nM]

This is the coolest thing I've seen in a long time...

Definitely! Absolutely cool!

[Kevin.D]

Someones done a  2 D automated movement  version using the same idea that Mike used .

[Rory]

Looks to me like someone has a magnet under the table.

[miguelvp]

Looks to me like someone has a magnet under the table.

If you look at his project link, it looks legit:

http://spritesmods.com/?art=magnmicrobots

[mikeselectricstuff]

Lots of info in the patent  :
https://docs.google.com/viewer?url=patentimages.storage.googleapis.com/pdfs/US8593016.pdf

Couple of interesting things on a quick skim - use of a diamagnetic layer to reduce friction, and conductive (presumably copper) for damping via eddy-currents.
I wonder if bismuth foil would be diamagnetic enough - probably easier then getting a big sheet of pyrolytic graphite.