Is this project an undergraduate, graduate, or faculty project?
Graduate
Project Type
individual
Campus
Daytona Beach
Authors' Class Standing
Timothy Mascal, Graduate Student
Lead Presenter's Name
Timothy Mascal
Lead Presenter's College
DB College of Engineering
Faculty Mentor Name
Bryan Watson
Abstract
The problem of multi-agent swarm formation control seeks decision frameworks that exhibit desirable emergent behavior, including distributed sensing and communication, while remaining resilient to environmental noise. Previous work has demonstrated that the Selfish Herd Model is not only resilient to simulated sensor noise but performs better in its presence. This work further investigates the emergent behaviors of the Selfish Herd Model, particularly distributed sensing, by investigating whether the swarm achieves global stability equivalent to that observed in previous work utilizing global knowledge. A parameter variation study was performed across varying agent visual range and maximum neighbor count inputs to simulate agents having access only to local knowledge. Preliminary results show that local agent interactions propagate information across the entire swarm, producing outcomes equivalent to agents having global knowledge, even when each agent relies only on relative local measurements: the bearing and distance to nearby neighbors. This has practical implications for real-world swarm designs, as it eliminates the need for a centralized information-sharing agent, reducing communication overhead and requiring no external global coordinate reference such as GPS.
Did this research project receive funding support (Spark, SURF, Research Abroad, Student Internal Grants, Collaborative, Climbing, or Ignite Grants) from the Office of Undergraduate Research?
No
Included in
Computer-Aided Engineering and Design Commons, Industrial Technology Commons, Systems Engineering Commons
Selfish Herd: Emergence of Distributed Sensing
The problem of multi-agent swarm formation control seeks decision frameworks that exhibit desirable emergent behavior, including distributed sensing and communication, while remaining resilient to environmental noise. Previous work has demonstrated that the Selfish Herd Model is not only resilient to simulated sensor noise but performs better in its presence. This work further investigates the emergent behaviors of the Selfish Herd Model, particularly distributed sensing, by investigating whether the swarm achieves global stability equivalent to that observed in previous work utilizing global knowledge. A parameter variation study was performed across varying agent visual range and maximum neighbor count inputs to simulate agents having access only to local knowledge. Preliminary results show that local agent interactions propagate information across the entire swarm, producing outcomes equivalent to agents having global knowledge, even when each agent relies only on relative local measurements: the bearing and distance to nearby neighbors. This has practical implications for real-world swarm designs, as it eliminates the need for a centralized information-sharing agent, reducing communication overhead and requiring no external global coordinate reference such as GPS.