I saw an interesting post on the website for The Atlantic a few weeks ago about moth-driven robots. The idea is that nature has evolved elegant and effective solutions to problems that science and engineering are still struggling mightily with; in this case, the ability to track smells. Researchers at the University of Tokyo built a wheeled robot that could be driven by moths walking on a crude trackball (picture what was in your computer mouse in the dark ages before optical mice became common). The (male) moth-bots were placed in an obstacle course with female moth sex pheromones at the opposite end, towards which they made their way surprisingly quickly and effectively (even when the researchers biased the steering to always pull in one direction).

The moral of the story, other than the fact that moths have a very healthy libido, is that it may be possible to harness features of nature that cannot yet easily be replicated artificially. For instance, tracking environmental spills to their source is one potential immediate application in the tracking realm, and the concept can easily be stretched to any number of other fields. Godzilla would be wise to beware – perhaps next time Mothra will bring some new toys to the fight.

Swarm Behavior

I recently saw this video of a starling flock murmuration (a collective swarming behavior). In addition to being an great example of how cool nature is, I found it fascinating as an example of emergent behavior.

Emergent or complex systems are those in which complex (and hard to predict) patterns arise out of interactions based on simple rules, without a central entity controlling or affecting the system properties. The sophistication and complexity of these systems is often unintuitive based on the simplicity of the governing rules. In nature, examples other than starling flocks include ant colonies or the weather. In the sphere of human interaction, the stock market or the development of cities can be interesting case studies (though both can be affected by central planning in some cases). Complexity, by Melanie Mitchell, offers an engrossing but still layman-oriented look into emergent systems in more depth.

Though I find emergent systems interesting in and of themselves, they are also tied to some potentially groundbreaking mechatronics research. Understanding how to model and eventually even control the collective behavior of decentralized, self-organizing systems could be instrumental in the development of more sophisticated artificial intelligence and other complex systems. The capability and robustness of systems that are not dependent on the existence of or communication with any centralized intelligence authority can far exceed more traditionally designed systems, whether it is a robot with a microcontroller “brain” or an army with a concentrated command team.

Robotic swarm behavior is a very active area of research; groups worth checking out include the Orb Swarm and Swarmanoid projects. It will be fascinating to watch progress in this field going forward – maybe the next remake of The Birds will use robots instead.