Crowd-Modeling

From TEILab
Jump to: navigation, search

Socially-based Crowd Simulation

Project Team

  • Project Lead: Seung In Park PhD Candidate
  • Dr. Yong Cao Assistant Professor, Center for Human Computer Interaction
  • Dr. Francis Quek Professor, Center for Human Computer Interaction

Note: This is a recent project that is currently dormant

Project Overview

Social interaction and group coordination are important factors in the simulation of human crowd behavior. To date, few crowd simulation methods have been informed by models of human group behavior from the social science studies. In this work we advance a computational model informed by Common Ground (CG) Theory that both inherits the social realism provided by the CG model and is computationally tractable for large numbers of groups and individuals.

Approach

We employ a model from social-psychology and linguistics due to Herbert Clark[1,2] in which members of a group must negotiate and maintain a state of common ground (CG) as a precondition of joint activity. The task of navigation in a group is viewed as performing a joint activity among agents, which requires effective coordination among group members.

Our model includes two kinds of strategies by which behaviors of groups may be embedded into a larger crowd simulation in a space rich with interaction possibilities for the agents. These strategies are macro-coordination and micro-coordination strategies. Macro strategies relate to overall action plans to accomplish the groups' goals, and is dependent on the simulation domain (e.g., a group of soldiers may select a particular strategy influenced by training and doctrine). Micro coordination strategies relate to the dynamics of human communicative behavior, and are used to simulate how group members get each other to understand what they intend and maintain the group cohesion in the navigation. The micro coordination strategy is determined to a degree by the constraints of human perception, the physics of sound in voice communication, and cultural concerns.

We showed our CG-based crowd model generates significantly different crowd interaction and circulation patterns in the simulation. The implication of this result is that models that do not consider the cost of coordination may fail to capture the real crowd effects. Currently we are proceeding with studies to evaluate the realism of crowd behaviors and dynamics by conducting user studies.

References

[1] Clark, H. H. 1996, May. Using Language. Cambridge University Press.

[2] Clark, H. H., and S. Brennan. 1991. “Grounding in communication”. 127–149

Publications

  • Park, S-I., Quek, F., and Cao, Y., “Simulating and Animating Social Dynamics: Embedding Small Pedestrian Groups in Crowds”, in Journal Computer Animation and Virtual Worlds, John-Wiley & Sons.
  • Park, S-I., Quek, F., and Cao, Y., “Simulating and Animating Social Dynamics: Embedding Small Pedestrian Groups in Crowds”, in The 26th Conference on Computer Animation and Social Agents, CASA 2013.
  • Park,S. I., Cao. Y., and Quek, F., "Modeling Small Group Behaviors in Large Crowd Simulation". in Proc of ACM/SIGGRAPH Symposium on Interactive 3D Graphics and Game (I3D '12), March 8-10, 2012
  • Park, S.I., Chao, P., Cao, Y., and Quek, F. “A Crowd Modeling Framework for Socially Plausible Animation Behaviors,” in 5th International Conference on Motion in Games, Rennes, France Nov. 15-17, 2012.
  • Park, S.I., Quek, F., and Cao, Y. “Modeling Social Groups in Crowds using Common Ground Theory”. in Proc. of the 2012 Winter Simulation Conference, Dec. 2012.
  • Park, S.I., Quek, F., and Cao, Y., “Modeling Agent Social Joint Actions via Micro and Macro Coordination Strategies,” in Proc. of the 2012 IEEE/WIC/ACM International Conferences on Web Intelligence and Intelligent Agent Technology, Dec. 2012, (acceptance rate 20%).

Acknowledgement

This research has been partially funded by NSF grants, EAGER: Drummer Game: A Massive Interactive Socially-Enabled Strategy Game, IIS-0940723, and CRI: Interfaces for the embodied mind, IIS-0551610.