Mini-interview of Garnet Hertz

Yesterday Beverly and i drove to Irvine to meet with Garnet Hertz. We visited his super clean and very big workshop at University of California Irvine and i asked him a few questions (just a few cuz we arrived late and kind of mixed up our schedules.)

Garnet is famous for his cockroach-controlled mobile robot that puts a giant Madagascan cockroach at the control center (on top of the pingpong ball) of a robot.

The idea behind this hybrid biorobotic system is to illustrate that the simple intelligence of a cockroach can provide a novel control center for a mobile robot.

Why did you choose to use a cockroach? Is it because people usually find it repulsive so you can mistreat it and nobody will complain?

There’s something like that, yeah. Everyone has some kind of understanding of roaches. Each one has a history, a reaction, comes with an opinion about roaches. So i was meaning to develop something funny about this past. Another reason why i got interested in roaches is that they are often used in connection with biomimetic robotics, for their networking capacities, to guide autonomous vehicles, etc. I wanted to put these abstractions of the cockroaches into hardware, to invert the biomimetic concept.

Why did you select the Madagascar roaches?

They are big and easy to find. They are raised all over the world to be fed to iguanas and lizards. Besides they don’t have wings to cover their back like other roaches. So i can stick a velcro to the cuticule of their body without hurting them.

They also have a strong emotional response to stimuli. When they are angry, these roaches can hiss very loud.

Are they all “good” performers?

No, i have to go through 20 of them to select two roaches which will do well at demonstrations. They have their own personnalities. Some are too lazy, others are very curious, etc.

If you believe that Garnet doesn’t care for his roaches, check out the lovely dwelling he crafted for them.

Another of Garnet’s works, Experiments in Galvanism: Frog with Implanted Webserver is currently installed at Latitude 53 in Edmonton, Canada. A miniature computer is implanted into a dead frog. The animal is suspended in liquid contained in a glass cube, with a blue ethernet cable leading into its splayed abdomen. The computer stores a website that enables users to trigger physical movement in the corpse. You can view and activate the project online until March 18th 2006.

If you’re in L.A., come and join DORKBOTSOCAL11, “Open Hack” Event (with Hertz “how to solder” workshop). At the Machine Project March 4th 2006, 1pm. i hope to make it there as well when i’m back from UC DARNet.

This month, Garnet will show the roachbot at the STRP festival and will be giving a talk at Dorkbot Eindhoven on Sunday March 26th 2006.

2Comments

Add yours

1

fritz on March 2, 2006 at 6:38 pm

very interesting this mini-interview, especially the suggestive question about ‘why the cockroaches?’: last year at ars electronica I asked him exactly the same question and he was saying something like: ‘oh yea, there are all those japanese guys doing stuff with animals and robots and I thought it would be cool doing something similar with cockroaches…’ :-)
2

Garnet Hertz on March 4, 2006 at 9:28 am

Actually, you’re both basically right. Here’s a link to an article on the Japanese stuff I told you about at Ars:http://www.intercorr.com/roach.htm
A note of clarification:
“Another reason why I got interested in roaches is that they are often used as a model in biomimetic projects: their neural network structure and their physical agility/survivability is admirable. In other words, researchers admire cockroaches (and other insects) because they’re experts at solving specific problems that some computer programs have difficulty with.
Playing with this, I wanted to invert these abstractions of software-cockroach-models, and to use an actual cockroach instead of a piece of software. In other words, if cockroaches are so great, why not use an actual cockroach?”

Share this: