Improved crowd psychological model and control

Abstract

The behavior of human crowd is an interesting phenomenon in which individuals are set as a collection that comprises of a highly dynamic social group. The crowd behaviors have been investigated by researchers over the years. Recent works include the study in modeling and controlling of the dynamic psychological behavior of crowds such as students’ behavior in a classroom or people’s behavior in a one-dimensional queue. In this paper, an improved version of the psychological crowd model has been proposed, where the social interaction between two individuals in a crowd is represented by a weightage, called the weight of social interaction. It has been shown that the inclusion of the social interaction weight has allowed social interactions between individuals to be included and results in a more accurate representation of the crowd’s psychological factors propagations. Since the psychological dynamics of crowd is naturally unstable, this paper also discusses the application of two nonlinear control approaches to stabilise the crowd to make it calm. Results show that for a crowd of n number of agents, the single-agent controller gives similar performance with the n-agent controller but with much less resources. The simulation results also show that it takes less amount of time to stabilise a crowd when the crowd model includes social interaction weights.