糖心TV

Speaker: Mate Nagy (Eötvös Loránd University)

Title: Collective motion and group behaviour – Observations, models, quantitative analysis and robotic applications

Abstract:Many animal groups display impressive levels of coordination in their activities, during which input from one or multiple members is propagated quickly and efficiently to guide the group's behaviour. The role of leaders in achieving such collective behaviour is a highly interdisciplinary question that bridges the fields of animal behaviour, information theory, network science, statistical physics and robotics. The central aim of the talk is to show the background and our recent research results achieved by the group of Prof. Tamás Vicsek at MTA-ELTE Statistical and Biological Physics Research Group, Budapest, Hungary. Some models of collective motion are very general trying to catch universal behaviour, some of the models concentrate on more specific aspects. In parallel with the modelling we studied behaviour of animal groups. To uncover general principles of leadership and dominance we analysed two natural model systems: flocks of pigeons travelling together and packs of dogs in a free running/walking situation. To examine and compare leadership and dominance hierarchies, the animals’ movement paths were recorded by high-resolution GPS loggers, while we also measured social interactions among the same individuals. Collectively feeding homing pigeons were filmed and their social ranks determined using novel computer vision based analyses, where the quantitative analysis of automatically extracted trajectories was complemented by the detection of dominance-induced events. The results revealed some important underlying rules of collective motion and our aim is to utilise those to develop artificial systems. As a realisation of these efforts very recently our group succeeded to build the first outdoor autonomous robotic flock, where each unit performs all necessary computation on board.