Spontaneous gait synchronisation in the wild: exploring the effect of distance and level of interaction

Adrien Gregorj, Zeynep Yücel, Francesco Zanlungo, Takayuki Kanda · 2025

Read Spontaneous gait synchronisation in the wild: exploring the effect of distance and level of interaction by Adrien Gregorj, Zeynep Yücel, Francesco Zanlungo, Takayuki Kanda (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-08967.txt.

Introduction

Adrien Gregorj and co-authors measured in 2025 how walking partners fall into step without being told to. The data came from an underground pedestrian network, where uninstructed walkers were recorded going about their day. Groups were annotated for relationships, interaction level, and physical contact. The team first devised a way to pull the body's gait wobble out of plain trajectory data.

Frequency tools measured how well stride rhythms aligned between companions. Nonlinear tools then judged the stability of the paired motion. Closer distance and richer social interaction strengthened synchronization and steadied stride timing. Three-person formations and each walker's position inside the group also mattered. The authors conclude social bonds literally shape how crowds move. Urban designers and robot builders both rely on such findings.

Catalog record

CREATORAdrien Gregorj, Zeynep Yücel, Francesco Zanlungo, Takayuki Kanda
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE112 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-08967

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Spontaneous gait synchronisation in the wild: exploring the effect of distance and level of interaction Adrien Gregorj, Zeynep Yücel, Francesco Zanlungo, Takayuki Kanda Abstract Gait synchronisation of pedestrians is influenced by a range of factors, including biomechanical properties like leg length, environmental elements such as presence of obstacles and cognitive aspects like sensory feedback. Studying gait data collected in ecological contexts offers unique insights into these numerous factors affecting synchronisation which controlled experimental settings may miss. This study addresses the challenges in assessment of gait coordination in complex real-world interactions by leveraging a dataset of uninstructed pedestrian trajectories. The dataset is recorded in an underground pedestrian street network and annotated for group relation, interaction levels, and instances of physical contact. The main goals of our study is to devise a method to identify gait synchronisation from trajectory data and to provide an in-depth analysis of social factors affecting gait synchronisation in pedestrian groups. To that end, we first propose a method to extract gait residuals from pedestrian trajectories, which capture motion of the body caused by gait-induced oscillations. We thereafter apply a suite of analytical techniques spanning both frequency and nonlinear domains. Frequency-based methods, including the Gait Synchronisation Index and Cross Wavelet Coherence, quantify the alignment of oscillatory patterns in gait. Complementary nonlinear measures, such as Lyapunov exponents, determinism, and recurrence quantification metrics, offer deeper insights into the dynamical stability and predictability of coupled gaits. We demonstrate that higher levels of social interaction are associated with increased gait synchronisation, evidenced by smaller variations in stride frequency, relative phase, and higher synchronisation metrics (GSI and CWC). Distances between pedestrians also influence gait synchronisation, with closer distances leading to stronger synchronisation. Nonlinear analyses indicate that dyads with higher levels of social engagement demonstrate more structured and stable gait dynamics. Additionally, triad formation and relative positioning are shown to influence synchronisation, with certain formations (e.g., ∨ \vee and ⟷ \longleftrightarrow ) showing more stable gait patterns than others (e.g., ∧ \wedge ). Overall, our findings suggest that social interactions shape pedestrian gait coordination, with interaction level and distance being key factors. 1 Introduction Human walking is a complex and dynamic process that requires the coordination of multiple biomechanical and neuromuscular systems. Central to this process is the gait cycle, which represents the sequence of movements from one foot contact to the next. This cycle encapsulates the intricate interplay between these systems and is fundamental to understanding locomotion. Gait synchronisation is influenced by a range of factors, including biomechanical properties such as leg length [ 1 ] , physical coupling [ 2 ] , sensory feedback [ 3 ] , and cognitive demands [ 4 ] . Understanding these factors has implications for a wide range of applications, from urban planning to wearable technology and therapeutic interventions. Studying gait synchronisation in ecological contexts offers unique insights that controlled experimental settings may miss, as real-world interactions often involve a level of complexity absent in laboratory environments. Besides, analysing pedestrian trajectories in natural settings poses challenges, such as noise, variability, and the lack of controlled conditions. This study addresses these challenges by leveraging a dataset of uninstructed pedestrian trajectories, recorded in an underground commercial district. The dataset is annotated with dyadic and triadic relationships, interaction levels (ranging from 0 for no interaction to 3 for strong interaction), and instances of physical contact. The primary aim of this study is to explore whether, and to what extent, social interaction intensity and physical contact influence gait synchronisation within social groups of pedestrians. To achieve this, we apply a suite of analytical techniques spanning both frequency and nonlinear domains. Frequency-based methods, including the Gait Synchronisation Index [ 5 ] and Cross Wavelet Coherence [ 6 ] , quantify the alignment of oscillatory patterns in gait. Complementary nonlinear measures, such as Lyapunov exponents, determinism, and recurrence quantification metrics, offer deeper insights into the dynamical stability and predictability of coupled gaits. 2 Background and literature survey 2.1 Gait cycle and its phases Key phases and parameters of the gait cycle are often described using the frameworks established by Perry and Whittle [ 7 , 8 ] . A complete gait cycle, or stride , encompasses the time and distance between two successive placements of the same foot, with stride length measuring the distance covered during this cycle. In contrast, a step refers to the time and movement between successive placements of opposite feet, with each stride consisting of two steps. Another critical parameter is the step frequency (or cadence), which denotes the number of steps taken per unit time. Since each gait cycle includes two steps, cadence effectively reflects the frequency of half-cycles in walking, equivalent to twice the stride frequency. When individuals walk together, their gaits may exhibit a tendency to synchronise, a phenomenon known as gait synchronisation . It involves some extent of phase locking [ 9 ] , where the timing difference between steps remains constant. Notably, two prominent forms of synchronisation are in-phase , where the same foot contacts align, and anti-phase , where opposite foot contacts occur simultaneously. 2.2 Metrics for detecting gait synchronisation in paired walking Understanding gait synchronisation between individuals requires metrics that effectively capture interpersonal coordination. A comprehensive review of these metrics is provided by Felsberg et al. (2021) [ 10 ] , offering a solid foundation for exploring methods used to detect gait synchronisation in paired walking. One of the earliest approaches to measuring interpersonal synchrony was introduced by Miles et al. (2010) [ 11 ] , who analysed the distribution of relative phase values between individuals. By comparing the proportion of in-phase values (i.e., relative phase close to 0 ) to chance levels, they quantified the degree of synchronisation, laying the groundwork for more sophisticated methods. Building on traditional techniques, Nessler et al. expanded the analysis of gait synchronisation by incorporating nonlinear methods rooted in chaos theory. Their work utilised tools such as recurrence plot analysis and Lyapunov exponents to capture the complex dynamics underlying interpersonal coordination [ 1 , 12 ] . These approaches allowed for a deeper understanding of the stability and variability in synchronised walking. Based on general measure of synchronicity between two oscillators introduced by Tass in [ 13 ] , the Gait Synchronisation Index (GSI) has proven to be a reliable measure of interpersonal gait coordination. Zivotofsky et al. demonstrated its utility across diverse contexts [ 5 , 4 ] , reporting GSI scores as high as 0.4 when multiple feedback modalities (visual, auditory, tactile, and instructions) were integrated. Further extending its application, Soczawa et al. employed GSI in virtual reality settings to explore gait synchronisation with avatars [ 14 , 15 ] . More recently, Liu et al. have studied step synchronisation of pedestrians in the wild using visual inspection to identify synchronisation events [ 16 ] and found that social groups synchronised more often than non-related individuals. Inside social groups, dyads were found to have larger proportions of synchronisation than triads. They also conducted controlled experiments and measured the synchronisation of pedestrians walking side by side by measuring the time difference between instants of the feet touching the ground or being lifted. They showed that this could serve as a reliable indicator of belonging

Read the complete text