My colleague Nils Napp from Radhika Nagpal‘s lab at Harvard attended the DARS symposium this year and agreed to share some highlights with us.

This time around, the biannual Symposium on Distributed Autonomous Robotic Systems (DARS) was at Johns Hopkins University in Baltimore Maryland. The single track featured a great lineup of invited and technical talks. For example, Prof. Gracias gave a fascinating keynote talk about his group’s work in self-folding microstructures and progress toward making swarms of untethered micro-robots. They designed tiny metal grippers that can perform biopsies in difficult-to-reach areas and be retrieved in bulk via a magnet.

A striking aspect of the diverse (and generally fantastic) technical program was that the exotic Brazil-nut effect kept surfacing. With random agitation granular mixtures of different sized particles spontaneously separate, even against gravity! Larger particles usually end up on top: a fact that can readily be exploited for picking out large brazil-nuts from a mixture that includes less desirable varieties. Applied to robots, this effect can be used to design shockingly simple motion strategies to transport relatively large objects or sort robots.

K. Sugawara, D. Reishus, N. Correll (2012): Object Transportation by Granular Convection Using Swarm Robots. (download paper)

Gauci M., Chen J., Dodd T.J., and Groß R. Evolving Aggregation Behaviors in Multi-Robot Systems with Binary Sensors. (download paper)

As a special treat during the poster session, students from Sarah Bergbreiter’s group at the University of Maryland demoed some really neat, really small, and (comparatively) really fast mini-robots called TinyTeRPs.

Also, mark your calendars for 2014 when DARS will be in Korea!

More on the author: Nils Napp works on distributed robotic construction using iterative amorphous depositions. You might have seen his foam-depositing robot on the Automaton Blog or Robohub. At DARS he presented a paper to reliably build ramps in unstructured environments. The idea is to exploit the low-level mechanical feedback of compliant building materials to make construction more robust. As a result, relatively simple local strategies for interacting with irregularly shaped, partially built structures can give rise to robust, adaptive global properties.

Sabine Hauert is lecturer at the Bristol Robotics Laboratory and co-founder of Robohub, the Robots Podcast and the Autonomous Robots blog... read more

Follow @sabinehauert