Simulating Flying Insects Using Dynamics and Data-Driven Noise Modeling to Generate Diverse Collective Behaviors.

Area of Science:

• Computational Biology
• Collective Behavior
• Biophysics

Background:

• Simulating large insect swarms presents challenges due to complex interactions and emergent behaviors.
• Existing models often lack biological plausibility or struggle with high densities and realistic noise-induced movements.

Purpose of the Study:

• To develop a biologically plausible dynamics model for simulating large insect swarms.
• To accurately capture individual insect interactions, collision-free trajectories, and collective emergent behaviors.
• To generate realistic noise-induced movements observed in natural insect swarms.

Main Methods:

• A force-based dynamics model incorporating biological insights and experimental data.
• Modeling noise as a constructive force to generate realistic emergent movements.
• A data-driven formulation using pre-recorded insect trajectories.
• A novel evaluation metric and statistical validation approach for swarm motion analysis.

Main Results:

• The model successfully simulates large insect swarms (midges, fruit flies, locusts, moths) with varying densities.
• Generated simulations exhibit realistic emergent behaviors such as aggregation, migration, phase transitions, and escape responses.
• The combination of Curl noise and the dynamics model produces highly realistic swarm simulations.

Conclusions:

• The developed model provides a powerful tool for understanding and simulating complex insect swarm dynamics.
• It offers a biologically grounded approach to generating realistic collective behaviors and noise-induced movements.
• This framework advances the study of collective behavior in biological systems through accurate simulation.