Aleksandr Khasymski '07 Evolving Languages for Robotic Foraging Tasks

Language is easily one of the most fascinating phenomena, yet very little is known of how it came about. Researchers observe that communication between individuals arises in environments where cooperation and coordination are essential factors for survival. However, the connection between primitive forms of communication, such as that in ant colonies, and the complex structured language of humankind is almost a complete mystery.

In order to shed some light on the processes involved in the creation of a language, two cutting edge fields, evolutionary linguistics and behavioral robotics, have come together. Inspired by nature, researchers from those fields are set out to simulate the emergence and evolution of a communication system between autonomous agents.

The goal of this project was to make a contribution to this ongoing research by better understanding the influence of the environment and properties of a group of autonomous robotic agents on the creation and evolution of a language in that group. We describe experiments in which realistically simulated robots evolve a language for cooperating on a simple foraging task. The "brains" (control programs) for the robots can subsequently be transferred into real robot bodies (Khepera-II platform), which have hardware for performing the foraging/ grasping and communication behaviors. This work improves on an earlier, unsuccessful attempts to simulate language evolution using this approach.

Faculty Advisor: Simon Levy