Collaboration of intelligent, autonomous systems: Situation aware behavior change

Abstract

Random encounters of autonomous devices capable of inter-device communication open up the possibility for collaboration. Often times, however, a change of each device's behavior is needed in order to leverage synergies arising from co-location. We present an approach based on a knowledge base that is able to control a device in an isolated situation as well as during co-location with other devices. In both instances, control is done in a fault tolerant way. The algorithm is capable of smooth behavior transitions as well as behavior shifts where the behavior of the device changes abruptly, which can be seen as a change in strategy. The underlying knowledge base stores all possible behaviors of the device. We present an experiment that illustrates shifts in behavior as well as behavior changes in the context of internal faults. © 2008 Springer-Verlag Berlin Heidelberg.