Existing research on traffic and pedestrian flows has mainly focused on high-speed mobility data with average speeds above 1 m/s. However, the dynamics of low-speed movement, largely driven by social interactions, remain underexplored. In this study, we collected high-resolution spatiotemporal data on the movement of preschoolers in four distinct classroom and playground settings. Within the low-speed domain (average speed <1 m/s), we identified two novel social phases in children's movement: a gas-like phase characterized by free movement, and a liquid-vapor coexistence-like phase characterized by the formation of small social groups. Furthermore, analysis of peer interactions revealed a transition from ferromagnetic-like alignment to antiferromagnetic-like behavior. We propose a sophisticated potential that models the forces shaping these alignments. Our framework accurately reproduces the observed patterns, providing deep insights into the mechanisms governing collective behavior in low-speed human movement. These findings enhance our understanding of how social interactions influence movement and have broad implications for behavioral science, epidemiology, and active-matter systems.
*This work was partially supported by the National Science Foundation under Grants No. 2150830 and No. IBSS- 1620294, the Institute of Education Sciences under Grant No. R324A180203, the National Institutes of Health under Grant No. R01DC018542, the Marcus Autism Center, Children's Healthcare of Atlanta, and the Simons Foundation for Autism Research Initiative under Grant No. SFI-AR- HUMAN-00004115-01. We would also like to acknowledge the involvement of Emory University's Pediatric Biostatistics Core.
