Self-organisation phenomena in pedestrian counter flows and its modelling. (November 2022)

• Record Type: Journal Article
• Authors: Xie, Wei; Lee, Eric Wai Ming; Lee, Yiu Yin
• Is Part Of: Safety science. Volume 155(2022)
• Publication Date: 2022-11
• Subjects: Pedestrian counter flow -- Lane formation -- Collision avoidance behaviour -- Following behaviour -- Social force model -- Empirical validation; Industrial accidents -- Periodicals; Accident Prevention -- Periodicals; Safety -- Periodicals; Travail -- Accidents -- Périodiques; 363.11
• DOI: 10.1016/j.ssci.2022.105875
• Languages: English
• ISSNs: 0925-7535

Abstract

Highlights: We implemented realistic avoidance and following behaviour in a force-based model (AFFM) to simulate bidirectional pedestrian motion. The AFFM can reproduce empirically observed microscopic behaviours and lane formation phenomenon. We summarized an accelerate-then-decelerate lateral movement mode of pedestrians for collision avoidance. We used transfer entropy method to determine the spontaneous following behaviour of pedestrians. We introduced a new parameter, lane entropy, to explain the quantitative characteristics of pedestrians' self-organisation. Abstract: We implement avoidance and following behaviour in a force-based model (AFFM) to develop realistic simulations of pedestrian counter flows. Using this innovative approach, we summarise an accelerate-then-decelerate lateral movement mode for pedestrians that minimised their required effort to avoid collisions with oncoming pedestrians travelling in the opposite direction. Then, we use the transfer entropy (TE) method to determine the spontaneous following behaviour of pedestrians to align their motion with a leading pedestrian moving towards the same destination. We also introduce a new parameter, lane entropy, to measure the motion smoothness of pedestrians and explain the quantitative characteristics of the lane formation phenomenon. The performance of the proposed model is compared with that of two other force-based models in terms of their ability to replicate the experimentally observed pedestrian flow patterns in a mock corridor, as designed by Feliciani and Nishinari in 2016. The results show that the AFFM performs well at reproducing microscopic behaviours observed in the experiments while consistently presenting the empirically observed lane formation phenomenon with quantitatively self-organised features (e.g., the order parameter and lane entropy). The proposed model is an efficient tool for future applications that can realistically model pedestrian counter flows under varying conditions.