Abstract

We quickly go through three different approaches to control robot swarm behaviors. Applying Langevin and Fokker-Planck equations combined allows for a mathematical micro-macro link to model swarm behavior. In a second approach, we program swarms with methods similar to population coding in neurology. Combining the right swarm fractions of specific predefined robot controllers results in desired and predictable swarm behaviors. In a third approach, we model populations of busy or congested robots to obtain a general model of scalability.