Generating self-organizing collective behavior using separation dynamics from experimental data

TL;DR

This paper introduces a data-driven model based on experimental pigeon flock data that generates diverse self-organizing collective behaviors, surpassing traditional models in complexity and realism.

Contribution

The paper presents a novel approach using averaged behavioral rules derived from real data to produce emergent collective behaviors in simulations.

Findings

Model reproduces flocking, pattern formation, and vortices

Behavioral diversity exceeds standard models

Potential for developing new complex behavior models

Abstract

Mathematical models for systems of interacting agents using simple local rules have been proposed and shown to exhibit emergent swarming behavior. Most of these models are constructed by intuition or manual observations of real phenomena, and later tuned or verified to simulate desired dynamics. In contrast to this approach, we propose using a model that attempts to follow an averaged rule of the essential distance-dependent collective behavior of real pigeon flocks, which was abstracted from experimental data. By using a simple model to follow the behavioral tendencies of real data, we show that our model can exhibit emergent self-organizing dynamics such as flocking, pattern formation, and counter-rotating vortices. The range of behaviors observed in our simulations are richer than the standard models of collective dynamics, and should thereby give potential for new models of complex behavior.