Traffic dynamics of uni- and bidirectional pedestrian flows including dyad social groups in a ring-shaped corridor

In this paper, we introduce dyad social groups into the experiment to mimic uni- and bidirectional pedestrian flows that are closer to real life. According to the experimental videos, different strategies of collision avoidance for dyads are observed and classified. Moreover, we observe an interesting lane-merging phenomenon in bidirectional scenarios. Fundamental diagrams are calculated based on two measurement methods, and further compared with previous individual experiments (without dyads), confirming the negative impact of social groups due to their inherent cohesion during the movement process. Then, group characteristics such as group distance and spatial alignment are analyzed under the influence of different global densities. It is interesting to note that the effect of density on spatial alignment distribution is opposite in uni- and bidirectional scenarios, i.e. the increasing density will make the distribution less scattered in the unidirectional flow, which is contrary to the phenomenon in the bidirectional flow. Finally, based on the time evolution of congestion level and crowd danger, they can be good indicators of transitions from disorder to order, especially in bidirectional scenarios. Furthermore, the differences between uni- and bidirectional flows can be well distinguished through the relations between density and congestion level, as well as crowd danger.