Replicating Robots

Replicating Robots

For humans and other animals, reproduction comes pretty naturally. But for robots, it's a huge challenge. Recently, scientists took a big step by creating simple robots that can make exact copies of themselves.


Replicating robots. I'm Bob Hirshon and this is Science Update.

If you visit Hod Lipson's engineering lab at Cornell University, you'll find robots multiplying like rabbits. He and a team of graduate students, including Victor Zykov, have created robots that can build exact copies of themselves from magnetic cubes. Lipson says the goal was to prove the concept.


Most of the machines, or the robots that you see in industry can make simple things, or parts of complex things, but the things they make are actually simpler than themselves. So the question is, can you make a machine that can make something as complex as itself?

Lipson says robots like these would be ideal for space missions since it's cheaper to launch a box of parts than a fleet of complete machines. I'm Bob Hirshon for AAAS, the Science Society.

Making Sense of the Research

If you think that building something as complex as yourself doesn't sound like a big deal, just think how hard it would be to build a fully functional human being. Luckily, biology takes care of that for us. But robots can't rely on Nature to reproduce – they need someone or something to build more of them.

So what's the problem? For living things, reproduction offers a key advantage: it's self-sustaining. A pair of dogs can make more dogs, which in turn can make more dogs, and so on. (The offspring will not be exact copies, but they will be very similar.) A robot, on the other hand, has to rely on human engineers, or at least a team of more sophisticated robots, to make more of its own kind.

Now imagine that you send a fleet of small robots to study the surface of Mars. What happens when one of the robots breaks down? It would take years to send a replacement. But if the robots could rebuild themselves from a few spare parts, they could solve the problem quickly and cheaply.

The Cornell robots don't look like the kind you see in science fiction movies. They're just a bunch of magnetic cubes that stick to each other in specific shapes. The computer chip inside each cube is a lot like human DNA. Every one of your cells contains the complete genetic code for your entire body (except sperm and egg cells, which contain half of your code). Theoretically, you could use the DNA from any one of your cells to clone an exact copy of you. In the same way, every cube in one of these robots contains complete instructions for building another robot just like it.

These instructions kick in when the robots come to a "feeding station" that delivers new cubes on a conveyor belt. The cube robot picks up the new cubes and puts them together. The new robot even helps to build itself.

The system is simple, and still depends on a steady supply of new cubes in order to work. And right now, the robots don't do anything except make more robots. Still, the research may eventually lead to useful robots that can fix or replace themselves without human supervision.

Now try and answer these questions:

  1. What is the difference between replication and reproduction?
  2. In what ways are these robotic cubes like human cells? In what ways are they different?
  3. Can you think of possible uses for self-replicating robots? Would they be more useful in some situations than in others?
  4. This research "proves a concept" – in other words, it shows that something can be done without resulting in any practical technology. What is the value of research like this?

For Educators

You can see pictures of the Cornell team's robotic cubes on the Self Replication project homepage.

Get a Grip on Robotics, from the Tech Museum of Innovation, is an online exhibit devoted to the robots and the study of robotics.

Related Resources

Curve Balls
6-12 | Audio
Server Naps
6-12 | Audio
Wasted Food Energy
6-12 | Audio

Did you find this resource helpful?

Science Update Details

Grades Themes Project 2061 Benchmarks State Standards