LAUSR

Abstract Mass events have led to many deadly crowd crushes and crowd disasters. In order to avoid crowd accidents, a number of the studies have been conducted. For the people in public place, even a low density crowd can cause risk due to individual’s local aggregation. The local aggregation process is the crowd distribution of deviating from random distribution process. Therefore, it is particularly important to measure the degree of randomness or degree of aggregation of crowd distribution. Entropy can be used to quantitatively describe the number of microscopic states in a certain macroscopic state. A crowd entropy model is built to reveal the degree of aggregation crowd or the degree of deviation from random distribution based on entropy difference. The fluctuation of the entropy is used to reflect the disturbance caused by random movement of individual. The research shows that we can judge whether the crowd is aggregating by comparing the entropy difference and the fluctuation value under a certain confidence level. The entropy variation regularity shows that the larger the entropy difference is, the more obvious the crowd aggregation is. It will provide a new perspective and tool for crowd managers to monitor the changes in the degree of crowd gathering.