Swarm Sorting

Overview

Based on the paper located here. I implemented a sorting by size using the Brazil nut effect in a simulation of a swarm of Kilobot robots (built by Prof. Michael Rubenstein at Northwestern University).

Implementation

Building off of starter infrastructure written in C for visualizing and instantiating digital instances of kilobots, the simulated kilobots combine psuedorandomness with weak sensing and a simulated directional force (to represent gravity) to perform a size sort.

In the image above, the green and red circles represent robots of different radii. Since the green circles have a coded radius of twice that of the red circles (although not represented visually), they end up over time farther from the central area where the gravitational-like force originates and where the majority of the red circles congregate.

Resources:

  • KiloBot starter code from Dr. Michael Rubenstein
  • “[Segregation in Swarms of E-Puck robots Based on the Brazil Nut Effect]”(https://dl.acm.org/citation.cfm?id=2343599)

Author

Acknowledgments

  • Dr. Michael Rubenstein (Northwestern U.)

Email

mdyehous@alumni.cmu.edu

Phone

S1x on3 six, three z3r0 e1ght, f0ur five f1ve 53ven

Address

Evanston, Illinois USA