Patterns, transitions and the role of leaders in the collective dynamics of a simple robotic flock

Abstract

We have developed an experimental setup of very simple self-propelled robots to observe collective motion emerging as a result of inelastic collisions only. A circular pool and commercial RC boats were the basis of our first setup, where we demonstrated that jamming, clustering, disordered and ordered motion are all present in such a simple experiment and showed that the noise level has a fundamental role in the generation of collective dynamics. Critical noise ranges and the transition characteristics between the different collective patterns were also examined. In our second experiment we used a real-time tracking system and a few steerable model boats to introduce intelligent leaders into the flock. We demonstrated that even a very small portion of guiding members can determine group direction and enhance ordering through inelastic collisions. We also showed that noise can facilitate and speed up ordering with leaders. Our work was extended with an agent-based simulation model, too, and high similarity between real and simulation results were observed. The simulation results show clear statistical evidence of three states and negative correlation between density and ordered motion due to the onset of jamming. Our experiments confirm the different theoretical studies and simulation results in the literature about collision-based, noise-dependent and leader-driven self-propelled particle systems. PACS: 05.70.Fh, 05.65.+b, 89.75.Fb Simulation resultsSimilarly to the experiments, the simulations reveal ordered motion in one (with leaders) or in both directions (without leaders) for low noise levels (figure 14).

Linear Analysis of the Vectorial Network Model

“…We further note that the formulation allows for considering additional measures of group coordination. For example, the velocity correlation introduced in [25] and adapted to the VNM reads as…”

Keeping speed and distance for aligned motion

“…O ne o f the m ost frequently observed and best know n group behaviors is collective m otion in biology w hen, for exam ple, cells, insects, birds, or fish move in stable spatiotem poral patterns [11][12][13][14][15][16]. In addition to biological cases at the cellular or anim al level, collective m otion phenom ena are also com m on in physics, inform ation technology, robotics, and the social sciences [17][18][19][20][21].…”

Collective Dynamics in the Vicsek and Vectorial Network Models Beyond Uniform Additive Noise

“…The resulting degree of coordination in the VM and the VNM has been thoroughly analyzed through numerical simulations and analytical treatments based on a number of ad hoc global observables, such as polarization (Aldana et al 2007;Czirók and Vicsek 2000;Pimentel et al 2008;Vicsek et al 1995) and velocity correlation (Tarcai et al 2011), and through the automated classification of spatiotemporal patterns via machine learning (Abaid et al 2012a;DeLellis et al 2013). For example, a continuous phase transition from order to complete disorder is observed in the VM when the intensity of noise is progressively increased (Czirók and Vicsek 2000;Vicsek et al 1995).…”