Territory Formation in Mobile Robots

摘要

Animals navigate through their environment using diverse strategies. These navigation strategies might be highly dependent on the animal's ecological niche. Though a lot of work is focused on biological navigation mechanisms, the cognitive ecology of spatial memory is poorly understood. The performances of different navigation mechanisms in the context of a biologically relevant behavior can be assessed by using mobile robots. For this task, territorial behavior was built on Khepera miniature robots. Territoriality is strongly dependent on spatial learning and thus provides a powerful context for testing of sensory and computational complexity needed by animals solving special tasks. Simultaneously, the principles underlying the establishment of territories in a previously uninhabited area and how it is dependent on environmental parameters can be investigated. In the present paper, territoriality in robots is introduced. The robots navigate by relying on poor sensor input and a representation of the environment of a low complexity. Qualitative results show that territories form. Subsequently, the consequences of the introduction of a newcomer in already established territories are presented.