Past Project : Pedestrian Evacuation

Pedestrian Evacuation

About Pedestrian Evacuation

The goal of this project was to develop a new generation of pedestrian evacuation simulation models and software that would take into account not only the characteristics of the building but also individual and group (social) behaviors of the pedestrians involved in the evacuation.

Participants

  • Natacha Thomas (Civil Engineering, PI)
  • Joan Peckham (CSC)
  • Manbir Sodhi (Industrial Engineering)
  • Charles Collyer (Psychology)
  • Jean-Yves Hervé (CSC)

Publications

J. Peckham, B. E. Aguirre, N. Thomas, Jean-Yves Hervé, Ron Hutt, Angel Castro, Charles Collyer, Kevin Culley, Elizete Fernandes, Fang Han, David Kurowski, Collin Lieberman, Maria del C. Rey, Lisa Ricci, Gabriel Santos, and Katharine Wray, "A Pattern for the integration of conceptual models in support of multidisciplinary efforts to develop software," in Simulation and Modeling VI, Proceedings of the 2005 ASCE International Conference on Computing in Civil Engineering, Cancun, Mexico, July 11-15, 2005.


Simulation of Pedestrian Behavior

Participants

  • Colin Lieberman (B.S. in Mechganical engineering 2007)
  • Elizete Fernandes (REU 2004)
  • Natach Thomas (Pedestrian Evacuation project PI)
  • Jean-Yves Hervé

Tracking of Pedestrians in Video Sequences

We have developed a novel distributed architecture for tracking in real time pedestrians using multiple cameras. We have also developed novel warping-based algorithms for identifying and tracking deformable 2D templates in a video sequence. We define our target as a textured blob, that is, a set of connected horizontal segments. Our tracking system successfully matches the target even in case of partial occlusion by a static object in the scene or occlusion by another pedestrian.

Participants
  • Meng Wan (M.S. Computer Science 2006)
  • Jean-Yves Hervé
Funding

Partial support for this work was provided by the National Science Foundation (Pedestrian EVacuation, NSF award #0331984)

Publications

Meng Wan and Jean-Yves Hervé, "Adaptive Region-Based, Layered Background Model for Target Tracking," in Proceedings of 18th International Conference on Pattern Recognition, Hong Kong, China, August 2006.

M. Wan and Jean-Yves Hervé, "Adaptive Target Detection and Matching for A Pedestrian Tracking System," in Proceedings IEEE Int'l Conf. On Systems, Man, and Cybernetics, Taipei, Taiwan, October 2006.


Visualization

Participants
  • Angel Castro (REU 2004)
  • Jason Carvalho (REU 2006)
  • Jean-Yves Hervé