Spontaneous gait synchronisation in the wild

This research presents the first systematic analysis of gait synchronisation within uninstructed socially interacting pedestrian groups using ecological data. The study examines over 5000 trajectories captured via laser sensors and video cameras in an underground pedestrian network in Osaka, complemented by annotations of group membership and social interaction levels. A novel method for extracting gait signals from tracked trajectories is introduced. To assess gait synchronisation, analytical tools such as Cross Wavelet Coherence, determinism, maximal Lyapunov exponent, and cross-recurrence analysis are employed. Results consistently show that pedestrians within social groups exhibit significantly higher synchronisation compared to incidental pairs (unrelated individuals in proximity) which, in turn, demonstrate greater synchronisation than random pairs. Further analysis reveals a positive correlation between social engagement levels and gait synchronisation. However, since social interaction and spatial proximity are interrelated, the primary driver of increased synchronisation is investigated. Findings suggest a nuanced interplay: both physical closeness and emotional engagement contribute to synchronisation, with social interaction amplifying the effect beyond mere proximity. These insights deepen understanding of the mechanisms underlying gait synchronisation in naturalistic settings and highlight the importance of social context in collective movement. Empirically, the study demonstrates a co-variation among gait synchronisation, social interaction, and proximity. Practically, this knowledge has the potential to inform better the design of safer, more resilient infrastructure, and enhance crowd simulation and urban planning. At a broader context, it sheds light on unconscious cooperative behaviours, emphasising social factors’ influence on collective dynamics.