In this post, I'd intended to blather on about one of my favorite topics: the three Ds of robotics (using the technology to do work that's considered too Dull, Dirty or Dangerous for humans). Just before I sat down to write it up, I saw an article in Scientific Computing with an update on a neat little biomimetic robot that mimics the motion of a water insect. I just had to talk about it. If you're not up on the three Ds, you'll just have to wait for another post.

Biomimetry is the technique of designing stuff (especially robots) that recreate the structure and function of a biological organism. The idea is as old as robotics, itself, having figured centrally in Karel Capek's 1920 play R.U.R in which servant androids (biomimetic robots made to look like humans) rebel and destroy the human race. The play's title is an acronym for "Rossum's Universal Robots" and introduced the word "robot" to the English language.

While androids are a staple in science fiction, the most successful biomimetic robots mimic simpler organisms, in this case an insect. Researchers at the Harbin Institute of Technology near Haerbin, Heilongjiang, China have been working on this artificial bug for several years now, and it has demonstrated the ability to walk on water, buoyed by surface tension. The thing was able to walk on water, but sank (breaking through the surface) when it tried to apply enough downward force to get airborne. To make the leap, the researchers had to coat the critter's feet with hydrophobic materials to strengthen the underlying surface tension.

Biomimetry is a good idea. I realized how important it was back in the 1960s when contemplating how one could best drive advances in computer science. Computers of the time had all the intelligence of a box of rocks, and the practical tasks generally proposed for them didn't take much more intelligence. Despite whines by my PHY101 students, solving the equations for projectile motion doesn't take much brain power. I wanted to see computers capable of solving the problem of perception, specifically through stereovision. What I determined was that the best way to drive computers to grow real brains was to give them bodies. The best way to do that was biomimetics: get them to do something that real animals do.

If you care about advances in general computer science, as well as advances in automation, you need to care about biomimetic robot research.

C.G Masi