Cockroach Controlled Mobile Robot

Garnet Hertz

Last Updated: 31 March 2007







Contents:

Recent Project News

Updated 03 February 2006



General Project Overview

"Cockroach Controlled Mobile Robot" is an experimental mechanism that uses a living Madagascan hissing cockroach atop a modified trackball to control a three-wheeled robot. If the cockroach moves left, the robot moves left. Infrared sensors also provide navigation feedback to the cockroach, striving to create a pseudo-intelligent system with the cockroach as the CPU.

[Listen to an Audio Overview of Cockroach Controlled Mobile Robot 2.6M MP3 File (interview broadcast on ORF from Ars Electronica 2005).]


 ROACH COACH: The creeping contraption isn t remote-controlled, nor does it move on its own, exactly. At its control center, on top of a pingpong ball, sits a giant Madagascan hissing cockroach. Designer Garnet Hertz - a University of California, Irvine, graduate student enrolled in the school s interdisciplinary arts, computer and engineering program - explains that as the roach moves on the ball, she controls the movement of the robot. When sensors detect an immobile object ahead, such as a wall, a bright light flashes at the cockroach. As she scuttles away from the light, the robot moves away from the wall.
Photo: Kevin Sullivan / Text: Tamara Chuang (OC Register: 18 Aug 2004)


Sponsors & Thanks



Cockroach Controlled Mobile Robot #1

(also known as "Control and Communication in the Animal and the Machine")

Description

Development of this system began in January 2004. This robot first made a public appearance at SIGGRAPH 2004 (10 Aug 2004) and was publicly shown until December 2004. The system was built at UCIrvine with the patient assistance of Tom Jennings.

The robot had a lightweight frame and was (from a robotic perspective) a simple mechanism.

Technical summary:

Demo History

Photo Documentation

Press



Cockroach Controlled Mobile Robot #2

(also somewhat still known as "Control and Communication in the Animal and the Machine")

Description

This system was completely redesigned and rebuilt from the ground up in Spring 2005. It operated on many of the same principles as version #1, but was designed to "blow up less". This version had a brand new frame, nice "mag" wheels, better drive system, could go in reverse, had a new harness for the cockroach, and had double the feedback resolution.


Technical summary:

Demo History

Photo Documentation

Press



Cockroach Controlled Mobile Robot #3

(also somewhat still known as "Cockroach Controlled Mobile Robot #2")

Description

This version kept the frame, rear wheel, cockroach harness, distance sensors and D200 motor controllers, but replaced the light-feedback display, drivetrain, wheels and axles. This work was done in August 2005.


Technical summary:

Demo History

Photo Documentation

Press


Manual



Frequently Asked Questions

Robot Operation: How Does it Work?

How does the cockroach move the robot?
The cockroach is placed on top of a modified trackball - a computer-mouse-type device - that is hooked up to some electronics. The cockroach is held gently in place above the ball so that when the cockroach moves it makes the ball spin: a little bit like a two-dimensional treadmill. The electronics (optical encoders) within the trackball send out small electrical pulses to some electronics that make the motors move in a similar direction to where the cockroach is moving on the ball.

What do the lights do?
The lights encircle the front of the insect, and are designed to help the cockroach navigate the object away from obstacles. In theory, cockroaches don't like light: as you may know, cockroaches scuttle quickly underneath the refrigerator if the light is turned on when you catch them in your kitchen at night. The lights around the front of the cockroach control center are each hooked up to a distance sensor. As the robot approaches a wall, for example, the distance sensor will detect that something is close, and will send electricity to the light. The light will turn on, and will alert the cockroach before it crashes into the wall. The idea is that the cockroach will want to scuttle into a "dark" under-the-fridge-type region, and therefore steer the robot into a clear space.

Have you noticed the cockroach change its behavior because it's in the machine? How has it adjusted?
Due to limited controlled-variable tests, this is difficult to know. At some moments, it appears as if the cockroach is "adapting in" to the system. In other words, behaving in a way that would indicate that it thinks that the robot control system is "real". There are other times that it appears the cockroach couldn't care less about the lights shining toward it and seems to enjoy driving the robot directly into walls at full speed.

Does the cockroach control the system better than a computer would?
At this point, likely not - but it depends on how you define "better".

Do you have a computer or microcontroller on the system that is controlling the robot?
No. The system has no microcontrollers or "computers". The electronics are composed of two optical encoders, four analog infrared sensors, six transistors, two 555 timing chips, and some resistors.

The Cockroach On The Robot

Does the cockroach live on the robot?
No. The cockroach is temporarily put into position at the driver's seat of the robot, and is removed after the demonstration is done. The cockroach typically stays on the robot for about 10 to 20 minutes, and then is removed and is returned to its terrarium home.

How does the cockroach stay on top of the trackball?
The cockroach is held in place with a piece of velcro. There is a small patch of velcro on the cockroach's back (cuticle) and a matching piece on the robot. The piece of velcro on the cockroach is held in place with cyanoacrylate adhesive, and is cut in sections to not interfere with normal daily life, mobility or body flexion. The cuticle on the back of the insect is like your fingernail, and having two small pieces of velcro affixed to the back of the insect appears to have no adverse effects to its quality of life. It is important to note that the cockroach is not bolted or surgically affixed on top the trackball.

Is the cockroach in pain?
Madagascan Hissing Cockroaches make a loud "alert hiss" when they are angry. They also enjoy feel safe when crammed into a tight space. Their cuticle has no nerve endings in it. Because of these reasons, and because they do not illustrate a fearful hiss when controlling the robot, it is my opinion that they are in no pain, and do not mind being in the robot system.

It looks like the cockroach is being crushed when it is controlling the robot - is this so?
No. If you get down to the eye level of the insect, you can see that the foot-to-ground relationship of the cockroach to the trackball is very similar to a standard cockroach gait. That is, if you know what a standard cockroach gait looks like. And another thing: cockroaches appear to feel comforted and safe when they are slightly crushed. The native habitat of Madagascan Hissing Cockroaches is the forest floor; as a result, cramming underneath an object/leaf is an instinctively safe place. People tend to like open spaces, cockroaches tend to like confined spaces.

Isn't it cruel to use a cockroach like this?
People are upset if cockroaches are healthy, happy and producing piles of babies in their kitchens. People are also upset if they are being hurt for the sake of entertainment. Both of these concerns are valid: people have a strong desire to control "nature". This project strives to invert the control structure somewhat by having an insect control a mobile robot: without the insect being "too" wild or controlled. The insects I use lead normal, healthy lives: if you don't believe me, send me your address and I'll slip some eggs under your front door.
In comparison, few people entertain the thought that taking a baby dog away from its parents, raising it with a bunch of humans in a house, putting a collar on it and feeding it cold food is cruel. Cockroaches aren't normally pets, and besides the Orkin man, we have few cultural norms regarding how to treat these insects besides crushing them in an emotional panic. It seems a little hypocritical, however, that the same people that swat mosquitos, lay ant traps, or have their houses sprayed for termites would object to a cockroach being used in a research (or creative) context.

Life as a Cockroach

That's a big insect. What kind of insect is it?
It's a Giant Madagascan Hissing Cockroach, also known as Gromphadorhina portentosa. The insect has no wings, moves relatively slowly and grows to a length of 5 to 8 cm. The lifespan of the insect in between two to five years. The females give birth to live nymphs (20-60) after 60 days of gestation, and they reach full maturity at 7 months. The insect is native to Madagascar, and leaf litter on the forest floor is its native habitat. Madagascan Hissing Cockroaches emit a loud hissing sound when agitated: it does this by forcing air out of a pair of modified spiracles in its abdomen.

What does the cockroach eat?
The cockroach eats almost anything, although it is fed organic lettuce and canned Pedigree "Traditional Ground Dinner With Chopped Beef (Improved Recipe)" dog food. A couple of cockroaches have died after being fed Carl's Jr. - although this may have been a coincidence. As a sidenote, it is dangerous to feed cockroaches any fruit with a peel: almost all fruit bought in a grocery store is heavily sprayed with insecticide that will kill the insects.

How do you tell male and female cockroaches apart?
No genitals-under-the-magnifying-glass is required. Madagascan hissing cockroaches can be sexed by looking at their backs: males have two horns on their backs, while females do not. The males use these horns to bash into each other during territorial fights with other males.

Can the cockroach jump up, bite me, and give me a disease?
No. These cockroaches cannot fly, jump, bite, and are as clean as any household pet. As a matter of fact, they find people dirty: after crawling on your hand, the cockroaches will tend to clean their feet from the oily residue natural to human skin.

Does the cockroach have a brain?
No. Cockroaches only have clumps of nerve cells called ganglia, which are positioned around their body. Cockroaches can survive for a number of days without their heads: they eventually starve to death.

Where did you get the cockroaches?
I got my most recent batch of cockroaches from http://www.roachdomain.com/ on the internet. The roaches cost between $5 to $10 each. This supplier has good first-hand knowledge of different kinds of cockroaches, and they specialize in rare and exotic hisser species. I've also bought cockroaches from pet stores in the past: Madagascan Hissing Cockroaches are by far the most common cockroach kept in captivity - they are sometimes kept as pets, educational classroom animals, or as "feeders" for people who keep reptiles.

Are there any legal, research, or technical restrictions with using cockroaches like this?
To the best of my knowledge, Gromphadorhina portentosa are not considered a pest insect in the state of California and there are no restrictions in raising, breeding, or keeping these cockroaches. From a scientific research perspective, cockroaches are not subject to the technical framework of a university animal research committee: animal research committees - who oversee university research involving animals - usually only care if you're doing work with vertebrate animals.

Do you have several cockroaches? Do you have "back up" cockroaches when you give a demonstration?
At this point [Feb 2006], I only about a dozen cockroaches. In November 2004 I only had one - the others were suspected of being killed by Carl's Jr. food (see note pertaining to being killed by Carl's Jr., above).

What is your cockroach named?
As of March 2006, many different cockroaches have used the robot. Some of these have names, some don't. Insects with names freqently have small name-tags on them since the behavior of each insect on the robot is unique: if needing to give a demo, it is best to pick an individual that is active/immersed on/in the robot.

Similar Research

Do you know of other similar art projects that use animals to control mechanisms?
Yes. In particular, two projects come directly to mind: Mark Pauline's Guinea-pig-operated 4-legged walking robot, and Ken Rinaldo's "Delicate Balance": a robotic system controlled by living siamese fighting fish.
Mark Pauline - founder and director of Survival Research Labs - built a guinea-pig-operated 4-legged Walking Machine that was operational between 1983 and 1985. It featured an on-board guinea pig that controlled the leg movements of the robot. From my recent conversations with Mark, he described that the guinea pig was trained to drive the machine by rewarding him ("Stu", as he was named) with lettuce. Although Stu was trained to enjoy entering the robot, during operation Stu would tend to position himself in a way that the machine wouldn't move. Video of this machine can be seen in the Survival Research Labs video "A Scenic Harvest from the Kingdom of Pain" available for purchase at http://srl.org/video/
Ken Rinaldo's "Delicate Balance" is a device with four fish bowls each containing one siamese fighting fish. These bowls were hung from two trusses that were controlled by the fish's body movements, and played with the natural response of the fish attempting to threaten or attack fish in adjacent (separate) bowls. More information about this project can be seen at http://accad.osu.edu/~rinaldo/works/mediated/

Are there other similar things you've heard of on a "folk" level... just ordinary people making stuff?
Although I haven't seen this in person, I've heard of people constructing "bug kites" which consist of attaching several June Bugs (Phyllophaga crinita) onto a paper or balsa airplane. Some more details about this can be seen at http://www.craftygal.com/archives/november/table1100.htm. I have also seen fly-powered airplanes - you can actually buy them at http://www.flypower.com/. It is important to note a distinction between this project and these june bug / housefly projects: these projects involve permanently attaching the insects to the device, and therefore killing them.

Are you familiar with similar projects done in a scientific context?
Yes. This work is directly inspired by scientific research in biorobotics: in particular the remote-controlled cockroach work at the Shimoyama Lab at the University of Tokyo in collaboration with Raphael Holzer. Within this project, the team was attempting to control the movement of a cockroach by implanting electrodes into their antennae. An Associated Press story that gives an overview of the research can be found at http://www.intercorr.com/roach.htm with the published paper titled "Behavioural Response of Insects to Electric Stimulation (Control of Insect Motor Function)" available at http://dmtwww.epfl.ch/~rholzer/papers/cbs.article.pdf with diagrams at http://dmtwww.epfl.ch/~rholzer/papers/cbs.ohp.pdf.

Are there other similar scientific projects funded by the military?
Yes. A large majority of hybrid insect/robot systems are funded by DARPA - the American Department of Defence's Defense Advanced Research Projects Agency (DARPA). In a survey of published cockroach/robot research within America, I was unable to find a lab that was not funded in some way by DARPA.

Project Funding

How is this project funded?
I fund this project myself. I get some money from different grant agencies that allow me to attend university, but most of my funding for the materials for this project come from me doing commercial design work for companies that manufacture toys and electronics.

How much did this project cost to build?
The materials for this project cost about $3,000 USD. Although it's an arbitrary calculation, I value my time invested in this project to be worth about $20,000.

Does this project have DARPA or military funding?
No.

Rationale and Goals of this project

Why are you doing this project?
I designed this particular project for many reasons. Some of these reasons are as follows: Why was this project initially called "Control and Communication in the Animal and the Machine"?
This is a little bit of a long story.
In 1948, Norbert Wiener published a book titled "Cybernetics: on Control and Communication in the Animal and the Machine" in which he coined the term "cybernetics". He used this term to describe the science of transmitting messages between humans and machines, or from machine to machine. Wiener saw human communication as a model for human-machine and machine-to-machine interactions, and that the quality of these communications influenced one's inner well-being: especially in a society in which people increasingly interact with and are reliant on machines. As the term cybernetics stems from the Greek kybernetis meaning "steersman", the process of designing machines that effectively respond to us is important and has direct impact on our social condition.
All of this theory has a large influence on the history of computing, communications, robotics control theory, and many other fields. This project somewhat takes Wiener's title literally: it's interested in control and communication in the animal and the machine. It's also interested in the history and social impacts of cybernetic theory within contemporary culture: especially hybrid bio/machine systems.

What does the robot solve?
The success of this project should be measured by the questions that it raises, not the solutions it gives. Building a system with superior task competence, objective measurement, generality or realism is something best left to the experts.

Is this project just for "fun"?
No. Although this system is built to be interesting and engaging, it is not "just for fun". Because this project engages with cultural theory doesn't mean that its sole purpose is entertainment. Laughing at something is usually good, anyhow: it can indicate that your existing mental framework is being stretched by something that doesn't quite fit it.

Is this an "artwork"?
Not really. The term "art" can be problematic, and conjures up concepts of things that reside on a rich person's coffee table or wall. In that sense, this isn't an artwork. In the sense of attempting to engage in cultural discourse, yes: this project could be considered an artwork.

Is this project "science"?
Not really. It's not really "scientific" unless it leads to a published paper in a peer-reviewed scientific journal. (Insert Latourian or Kuhnian rant here.)

What do you think about artificial intelligence or artificial life?
In one word: boring. The quest for artificial intelligence or artificial life might be more interesting if it was less artificial.

Project History & Development

Where have you shown this project?
I have shown this project in several different contexts. Check out the list above.

I am a professor and think that a live demo of this device would be an interesting addition to a conference I'm organizing. Can we talk?
Of course - that's a good idea. Everybody knows that most conferences are too boring, anyhow. Please contact me and let's scheme something up. I'm especially interested in demonstrating this system within robotics, biorobotics, artificial intelligence/life, body/technology, and media/critical/cultural theory contexts.

I am organizing a conference and I'm wondering if you have papers associated with this project that you could present. Can you present a paper?
Of course. I can also present a paper and do a demo if you like. Please contact me and let's figure something out. I'm especially interested in demonstrating this system within robotics, biorobotics, AI/ALife, body/technology, and media/critical/cultural theory contexts.

I am a scientist and am interested in seeing this work and am interested in the research applications of this device. Can we talk?
Of course.

I am an organizer/curator. Can I get you to come and give a lecture/demo at an upcoming event?
Of course. Covering travel/shipping expenses and an honorarium is required, however. I also don't fly with the insects, so we need to co-ordinate pre-shipping or sourcing the insects at your location. Usually this doesn't prove to be very difficult. We also need to come up with a plan about what to do with the insects after the event: often shipping them back to the source where they were bought is best.

I am a curator. Can this piece be in a three-month exhibition at my museum?
Let's talk. I prefer to give live demonstrations of this project when possible, but I'm interested in hearing about your show.

I am a member of the press and would like to do a story on this project. Can I come and interview you or shoot video?
Let's talk - I encourage dissemination of this project through the mass media, however, I'd prefer if it wasn't only billed as a "weird science" sidebar.

I am just a person that is interested in seeing this project. Can I get a demo?
Yes: If you come to my studio/office in Irvine California, I will gladly give you a 10 minute one-on-one demo. In some instances, the machine may be disassembled, travelling or unavailable, however.

I phoned you at 949-824-8876 and there is no answer and no answering machine. Did I get the right number?
No: that's an old number that I'm no longer at. Go to conceptlab.com and track down my contact information there.





Garnet Hertz - http://www.conceptlab.com