- The Washington Times - Thursday, May 18, 2006

An autonomous mechanical cockroach may not seem humanity’s most crucial technological need just now, but European scientists have come up with one anyway. And it is a natural leader, able to get the roaches to go where it wants them to.

The robot is a product of the LEURRE Project, started in Europe in 2002. It is thought to be one of the first efforts to have a robot be accepted by animals as one of themselves. The hope is that one day robots based on the same principles may be of use in such fields as herding sheep or raising poultry.

The robot is called “InsBot,” for Insect Robot, and is about the size of a box of matches. It has wheels. On-board microprocessors control it so that it acts like a roach. A tiny camera, a temperature sensor, and 12 infrared proximity sensors allow it to avoid obstacles and find the roaches. Real roaches then follow it, believing that it is one of them.

“Now,” you might say, “You are obviously making this up. Real roaches don’t have wheels. Wouldn’t they notice?” No. As it happens, InsBot is treated with roach pheromones. Pheromones are chemicals secreted by animals, particularly insects, that cause others of the same species to do specific things. Often they are sexual attractants.

In the case of the roaches used in the InsBot project, the pheromones serve only to identify members of the species to each other. Apparently, roaches believe that if it smells like a roach, it must be one, even if looks like a box and rolls around on wheels.

According to the scientists, the insects really do accept it and follow it around. For example, roaches naturally avoid light. Yet they will follow Insbot into lighted areas. This suggests using it to lead them into places where they might die.

This isn’t likely. Using a pricey robot to do what a cheap roach trap will do doesn’t make sense. But if, as a marine biologist I know suggests, similarly convincing robotic fish could lead schools of real fish into nets, there would be considerable commercial utility.

The problems of design were tricky. The InsBot has to be able to detect a roach and distinguish it from a wall. This sounds easy, until you have to do it with minimal hardware. But it seems to work.

It turns out that a lot more is involved with InsBot than a sophisticated fraudulent roach. While roaches seem fooled by the pheromones, they will not follow the robot unless it also acts like a roach. It wasn’t obvious what “acting like a roach” meant to a roach.

So the group also studied the general behavior of roaches, meaning their patterns of walking and stopping, with a camera that watched the insects from above. From the resulting traces they developed a mathematical model, or description, which the Insbot carries in its computers. This is being done in a way that should make it easier to describe the behavior of other species.

To me, the fascinating thing about InsBot is that it is autonomous. In principle, you put it on the kitchen floor and for four hours (that, at any rate, is the battery life) it runs about on its own, making friends with roaches, and deciding by itself what to do. Slick.

If InsBotlike robots had an independent source of power (solar cells for example) they could function until they wore out. The idea of leaving a robot by itself for months to do a job unsupervised approaches practicality.


Copyright © 2018 The Washington Times, LLC. Click here for reprint permission.

The Washington Times Comment Policy

The Washington Times is switching its third-party commenting system from Disqus to Spot.IM. You will need to either create an account with Spot.im or if you wish to use your Disqus account look under the Conversation for the link "Have a Disqus Account?". Please read our Comment Policy before commenting.


Click to Read More

Click to Hide