Smarticle swarm: In active matter, spontaneous robot dances show a new kind of order


In science and engineering, predicting when and how sets of particles, robots, or animals will be organized remains a challenge.

The discipline of statistical mechanics was founded by scientists and engineers in the 19th century, predicting how groups of simple particles transition between order and disorder, such as when a set of randomly colliding atoms solidifies into a uniform crystal lattice.

More difficult to foresee are the social actions that can be done as particles become more complex so that they can travel under their own control.

This type of mechanism – found in bird flocks, bacterial colonies, and robot swarms – is called “active matter.”

A team of physicists and engineers has proposed a new theory, as stated in the Jan. 1, 2021 issue of the journal Science, through which active-matter systems can spontaneously order themselves without the need for higher-level instructions or even programmed interaction between agents.

And in a number of systems, including groups of periodically shape-changing robots named ‘smarticles’ – intelligent, active particles – they have demonstrated this theory.

The theory, founded at the Massachusetts Institute of Technology by postdoctoral researcher Pavel Chvykov while he was a student of Prof. Jeremy England, now a researcher at the Georgia Institute of Technology School of Physics, postulates that certain forms of active matter with sufficiently chaotic dynamics spontaneously find states that researchers call “low rattling.”

“Rattling is when matter takes the energy flowing into it and converts it into random motion,” England said. When the motion is more aggressive, rattling may either be stronger, or more random.

Low rattling, on the other hand, is either very light or highly structured – or both. Because the theory is that the mechanism will spontaneously rearrange until it reaches that state and then get trapped there if your matter and energy supply allows for the possibility of a low-rattling state.

If you supply energy with a specific pattern by powers, that means the chosen state can find a way for matter to move that fits that pattern exactly.

England and Chvykov drew inspiration to establish their theory from a phenomenon discovered in the late 19th century by Swiss physicist Charles Soret, which he called thermophoresis.

In the 19th century, Soret discovered it. In his experiments, Soret discovered that the salt concentration in the colder area spontaneously increases when an initially uniform salt solution in a tube is exposed to a temperature difference – leading to an increase in the order of the solution.
To illustrate the low-rattling theory, Chvykov and England developed various mathematical models, but it was not until they sat down with Daniel Goldman, Dunn Family Professor of Physics at the Georgia Institute of Technology, that they were able to test their forecasts.

Goldman said, “A few years ago, I saw England give a seminar and thought that some of our smarticle robots might prove useful in testing this theory.”

Graduate students William Savoie and Akash Vardhan, working with Chvykov, who visited Goldman’s lab, used three fluttering Smarticles enclosed in a ring to equate the experiment to the theory.

The students found that the robots spontaneously self-organized into a series of dances instead of demonstrating complicated dynamics and exploring the container entirely – for example, one dance consists of three robots slapping their arms in succession.

These dances were able to last for hundreds of flaps, but then unexpectedly lost stability and were replaced by a new patterned dance.

After first demonstrating that these basic dances were actually low clatter states, Chvykov worked at Northwestern University with engineers, Prof.

Todd Murphey and graduate student Thomas Berrueta, who built Smarticles that are more refined and controllable.

The improved smarticles helped the researchers to test the limits of the theory, including how for different arm flapping patterns the types and number of dances differed and how those dances could be regulated. “By controlling sequences of low rattle states, we were able to get the system to achieve configurations that do useful work,” said Berrueta.

Researchers at Northwestern University claim these results have far-reaching practical consequences for micro-robotics.


Leave A Reply