Emergent behavior for real world robots

摘要

An autonomous mobile robot operating in everyday life conditions will have to face a huge variety of situations and interact with other agents (living or artificial). Using the behavior-based framework, mobile robots have demonstrated how a functionality can emerge from the interactions between the behaviors and the environment. However, to increase the level of 'intelligence' required for such robots to operate in real life settings, other decision abilities must be integrated. This paper presents a general architectural methodology that allows the combination of various properties required for robots to behave intelligently, by still preserving the emergence of functionality at all decision levels. Validation of these abilities in different experiments are also described.