Trying to imitate-a step towards releasing robots from social isolation

Abstract

This paper proposes a scenario which allows one to study one-to-one interactions between robots in 'individualized' societies. We are inspired by the 'social intelligence hypothesis', which comes from primatology and states that primate intelligence originally evolved to solve social problems. If this is one important factor in the evolution of intelligence and cognition in primates, we suggest, that it might be a general principle in the evolution of intelligent artefacts, not necessarily restricted to a biological substrate. Therefore we work with robots which should be able to collect experiences from individual interactions and recognize each other in order to establish individual 'personal' relationships. We use a hilly habitat which (1) allows one to concentrate on aspects of social interaction without using obstacles or object manipulation and (2) provides enough complexity and ways for the robots to behave 'meaningfully'. The robots communicate explicitly by synchronization and matching of movement. This behavior is the key to a symbiosis-like relationship, i.e. it is necessary for the survival of the individual.