Uppsamling och drivning av artificiella och levande flockar
Tidsperiod: 2016-01-01 till 2018-12-31
Projektledare: Daniel Strömbom
Bidragstyp: Bidrag för anställning eller stipendier
Budget: 3 150 000 SEK
Herding of sheep by dogs is a powerful example of one individual causing many unwilling individuals (agents) to move in the same direction. However, the need to herd autonomous, interacting agents is not unique to the sheep and sheep dogs. Similar herding is required to keep animals away from sensitive areas, when cleaning up the environment, during crowd control, and for guiding groups of exploring robots. In collaboration with my proposed host, we recently developed a new herding algorithm based on two very simple rules: collecting agents when they are too dispersed and driving them once they are aggregated. This novel algorithm reproduces key features of data we collected from sheep-dog interactions, and suggests new ways in which robots can be designed to influence movements of living and artificial agents. During my Fellowship, I will implement our new algorithm in robot shepherds, and test their success at herding magnets, other programmed robots, and insects. All experimental work will be performed in Swansea, in collaboration with leading experts on collective behavior, in world-class facilities for acquisition and analysis of animal/robot movement data. I expect the results of this work to be of direct relevance to the real-world shepherding problems described above, and in addition, represent a major advance in behavioural research with significant implications for fields as diverse as biology, mathematics, physics, engineering, and technology development.