基于相关性的上海世界博览会行人流分析

Correlation Analysis on Pedestrian Flow in Shanghai World Expo

拥挤环境下的行人流疏导与高效排队方法的研究对于应急疏导与社会安全有着重要意义。对上海世界博览会期间40个场馆的排队长度、时间数据的自相关特性,及场馆之间的相关特性进行了全面分析。研究发现,在访问量较高、平均排队长度较长的场馆,其客流一般呈现出较明显的正相关特性,特别是客流量达1 000左右的场馆,其表现出非常强的正相关性,而小流量场馆则有着很大的不确定性。访问量和自相关性之间的关系大致满足非中心卡方分布。此外,一些临近场馆之间的客流相关性斑图有着高度的相似性。这种基于相关性的讨论,有助于构建整个景点的客流斑图和行人流预测。

The studies on the dispersion of crowd pedestrian flow and queuing methods are important for emergency evacuation and the safety of society. Using the dataset of World Expo （Shanghai, 2010）, the autocorrelations between the series of queuing length and waiting time of pavilions as well as the correlations among different pavilions wereanalysed. It is found that the pavilions with higher flow usually show positive autocorrelations, especially those pavilions with flow of guests up to 1 000 per day, but the autocorrelations of others generally display much uncertainty. The relation between the traffic of visitors and the autocorrelation follows non-central chi-squared distribution. In addition, neighboring pavilions often demonstrate similar visiting patterns. These correlation-based discussions would be much helpful to the construction of global patterns and prediction of pedestrian flow.