Multi-agent Reinforcement Learning-based Pedestrian Dynamics Models for Emergency Evacuation

While there are numerous successful models like social force model and agent-based models that address high-density crowds, there is glaring lack of effective modeling techniques targeted at low- to medium-density pedestrian situations. Furthermore, previous studies have focused on either pedestrians’ route planning or pedestrians’ physical movements without considering the interactions between these two levels. This project will integrate these two levels to dynamically plan routes and control pedestrian’s movements during plan execution. The information of local environment and human behavioral characteristics is formulated into a reward matrix to re-plan pedestrians’ path or adapt to the changes in the environments. This facilitates modeling of the nonlinear characteristics of the human decision-making processes beyond simple rule-based models. The research team will develop the computational modeling framework and simulate emergency evacuation of a midsize airport. The research team will develop reinforcement leaning model to learn local navigation behaviors and simulate dynamic pedestrian behaviors. This model will be utilized to determine intermediate goals for each pedestrian particle, which is a key input for the time evolution of pedestrian trajectories. The project outcome will lead to a multidisciplinary computational framework for understanding and modeling the human decision-making process and resulting actions in emergency evacuations.

Language

• English

Project

• Status: Active
• Funding: $196908
• Sponsor Organizations:

  Center for Advanced Transportation Mobility

  North Carolina A&T State University
  1601 E. Market Street
  Greensboro, NC  United States  27411

  Office of the Assistant Secretary for Research and Technology

  University Transportation Centers Program
  Department of Transportation
  Washington, DC  United States  20590
• Project Managers:

  Park, Hyoshin

• Performing Organizations:

  North Carolina A&T State University

  1601 E. Market Street
  Greensboro, NC  United States  27411

  Embry-Riddle Aeronautical University

  600 S. Clyde Morris Boulevard
  Daytona Beach, Fl  United States  32114
• Principal Investigators:

  Park, Hyoshin

  Liu, Dahai

  Namilae, Sirish

• Start Date: 20190401
• Expected Completion Date: 20210331
• Actual Completion Date: 0
• USDOT Program: University Transportation Centers Program

Subject/Index Terms

• TRT Terms: Behavior; Crowds; Decision making; Evacuation; Machine learning; Pedestrian movement; Pedestrian traffic; Route choice; Simulation
• Subject Areas: Data and Information Technology; Pedestrians and Bicyclists; Planning and Forecasting; Security and Emergencies;

Filing Info

• Accession Number: 01708121
• Record Type: Research project
• Source Agency: Center for Advanced Transportation Mobility
• Files: UTC, RiP
• Created Date: Jun 19 2019 9:07AM