There is no great stagnation (remote-controlled cockroach edition)

Built-in power supply? Check. Ability to survive anything? Check. Easy to control? Okay, anyone who’s had a cockroach as an uninvited houseguest knows that’s not the case. So, rather than re-inventing the biological wheel with a robotic version, North Carolina State university researchers have figured out a way to remotely control a real Madagascar hissing cockroach. They used an off-the-shelf microcontroller to tap in to the roach’s antennae and abdomen, then sent commands that fooled the insect into thinking danger was near, or that an object was blocking it. That let the scientists wirelessly prod the insect into action, then guide it precisely along a curved path, as shown in the video below the break. The addition of a sensor could allow the insects to one day perform tasks, liking searching for trapped disaster victims — something to think about the next time you put a shoe to one.

What’s it like trying to climb the IQ gradient with this device?  There are videos at the link, and for the pointer I thank magilson.


Amazing. Imagine what it would do with larger animals. Good lord!

Could be useful for making politicians do intelligent things.

Fascinating. What I'd like to know is can you electro-tickle and fool the cockroach for ever? Or do they quickly learn that the danger is illusory?

OTOH, maybe you can just get rid of smart Mister Roach after mission's over and find a new gullible one.

I would hypothesize that it never knows that it was an illusion, so it can never habituate. Even if it could, you might be able to increase the electrical stimulus to overcome any habituation effects.

That is the most disgusting thing I've seen in a while.

I wonder if an approach like this will have a higher manufacturing cost than a real robot. After all, you probably need to raise the bugs, hold them still in some machine, and wire up the circuits. Maybe manufacturing everyone electronically is cheaper? Certainly, the economics are not obvious either way. A bug like this may have so many intricate moving parts that reproducing them in silicon or whatever *is not* cheaper.

You think "cockroach raising" is a significant cost? Visit a sewer......

No, I'm sure that part is cheap. I'm thinking of the cost of plucking one of them and holding it still while the microcontroller is attached. It's possible that they did this demo bot by hand. I don't see how that can scale up. Maybe machines can assemble these in the future? Not sure.

Grad students: cheaper than cockroaches.


You can put them in a freezer to keep them still for a while. Physiology labs that work with roaches do it all the time.

I've bought and used a Backyard Brains kit...not only is it labor intensive, it's skilled labor, too. My first attempt took almost three hours and resulted in a partially crippled cockroach.

The backpack is about $100. It's not clear that we can even make a robot as mobile as a cockroach for any cost, but the backpack sharply cuts into the roach's mobility, so it's hard to say. Also, the roach gets desensitized after being driven for half an hour or so, but it would probably be difficult to put a large enough battery in a roach-sized robot to get thirty minutes of aggressive use (how long does it take to run down a toy car?).

Back in the 1960s, some researchers at Sandia Labs put electrodes in the putative pleasure and pain centers of a mule. It was stimulated by a pack that had a photosensor which rewarded keeping a constant angle to the Sun. They sent the mule into the desert, and it followed a straight line.

This was a follow-on to research by Jose Delgado, who was famous for having implanted similar electrodes in the brain of a bull which he could control remotely using a modified walkie-talkie. Delgado actually stepped into the ring with the bull, and when it charged he stopped the charge by transmitting a signal to the receiver that controlled stimulation of the electrodes.

Note that SUNY-Brooklyn demonstrated a similar system with rats over 10 years ago:


I wonder what happened to those things? I did a quick google search and found a wikipedia page on them, but no real info. All the articles are years old. Did it not pan out? Or is it locked up by DARPA?

There are brain implants for electrical stimulation of the brain being done today, in people. See the bottom of page 7.

If nobody is making ratbots anymore, I'd guess it's for practical reasons like maybe their grant ran out.

People have been making these things with cockroaches or flies for a while. I mean, hell, you can mail order these things: Typically, they steer them through antennal or muscle nerves. It'd be like 'controlling a person' through activating their leg muscles. Although I bet you could do it by activating vestibular afferents.

The antenna nerves is more like steering you by creating burning sensations in your hands that you turn to avoid than it is like controlling muscles more directly. The latter is a harder problem because you need many more attachment points and code to fire them in the correct sequence. Actually I suppose it's a lot like controlling a horse with a bridle.

"Oh calm down, Mr. Samsa, we were only kidding!"

This seems like torture.

Since they're cockroaches, keep it up.

And the first thing it will be used for? Something useful like rescuing earth-quake victims? Or.... spying on people! I bet a year of my current earnings against a year of anybody else's that the military-intelligence cockroach applications are the ones that get explored first and everything else a long, long way behind. I bet MI6 have already set up a special entomological section..

Gee--any chance the U.S. military is funding this (I'm not judging)--but I am guessing they have thought of quite a few military applications. I see that this line of research is happening at many universities.

It would be suited for the military of poorer countries. For example, implant a brain stimulator, GPS, and an atomic bomb in a whale. That would be a lot cheaper than developing ICBM technology, and it wouldn't leave a trail pointing back to the perpetrator. DC, New York, Boston, LA, San Francisco, etc. are all near the ocean and/or large rivers.

Comments for this post are closed