Understanding the effects and capabilities of multi-agent crowd modelling when designing for human variability

Open Access
Author:
Krupski, Adi Atar
Area of Honors:
Engineering Science
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
  • Matthew B Parkinson, Thesis Supervisor
  • Corina Stefania Drapaca, Honors Advisor
  • Judith A Todd, Faculty Reader
Keywords:
  • Netlogo
  • Human variability
  • Waist Size
  • Crowd modelling
  • Agent-based modelling
  • MATLAB
  • CapSim
  • Aircraft egress simulation
Abstract:
This thesis provides a foundation for crowd modelling methods that consider the body size and shape of crowd members. Specifically, it seeks to understand the possibilities of crowd modelling software to incorporate waist size in order to create more realistic models. The OpenDesign Lab at the Pennsylvania State University has resources including anthropometric data that can be utilized with previous crowd modelling techniques in order to provide more detailed and accurate models. This thesis considers several possibilities of implementing refined crowd modelling techniques with OpenDesign Lab tools and preexisting models. It will provide a background of crowd modelling by providing an analysis of past and current research. Furthermore, this thesis proposes two crowd modelling techniques that enhance preexisting crowd modelling research. The techniques were created using Netlogo and MATLAB and investigate the possibilities and enhancements that anthropometric data can add to crowd modelling. The similarities and differences of these two programming languages are discussed and the crowd modelling methods are presented. This thesis also proposes a procedure using the Netlogo "breed" function and explores its applications.