Experimental and numerical detection of dynamic emergence in a human crowd

Dynamic emergence defined as a time-dependent entanglement (in the sense of coexistence and mutual influence of phases) of aligned, levorotatory (counterclockwise) and dextrorotatory (clockwise) phases has been recently put forward as a means to characterise collective behaviour in active matter [1, 2]. Up to now, dynamic emergence has only been detected in numerical simulations (interacting boids), hence this work is aimed at experimentally detecting it by drone recording different human crowds, each consisting of 30 members, moving within the area of a basketball court. The crowd was instructed to follow only two simple rules, namely, 1) To jog within a basketball court, and 2) To try to stay together at all times even if the crowd is disturbed by a simulated attack. The recorded emergent collective behaviour was characterised by extracting individual paths and velocity vectors in time, which were used to build local order parameters that revealed the existence of phases entanglement, thus confirming the presence of dynamic emergence. This result highlights the importance of using local order parameters to characterise collective behaviour. Additionally, an IABP (inertial active Brownian particle) model with three different interaction rules is proposed and compared with the available experimental data. This comparison shows that an IABP with visual weighted topological interactions reproduces the dynamics of a human crowd. Furthermore, a new parameter called rotational dispersion is introduced in order to identify dynamic emergence in a phase diagram.