Abstract
Coherent collective behavior emerges from local interactions between individuals that generate group dynamics. An outstanding question is how to quantify group coordination of non-rhythmic behavior, in order to understand the nature of these dynamics at both a local and global level. We investigate this problem in the context of a small group of four pedestrians walking to a goal, treating their speed, and heading as behavioral variables. To measure the local coordination between pairs of pedestrians, we employ cross-correlation to estimate coupling strength and cross-recurrence quantification (CRQ) analysis to estimate dynamic stability. When compared to reshuffled virtual control groups, the results indicate lower-dimensional behavior and a stronger, more stable coupling of walking speed in real groups. There were no differences in heading alignment observed between the real and virtual groups, due to the common goal. By modeling the local speed coupling, we can simulate coordination at the dyad and group levels. The findings demonstrate spontaneous coordination in pedestrian groups that gives rise to coherent global behavior. They also offer a methodological approach for investigating group dynamics in more complex settings.