Skip to Main Content
In the pervasive computing paradigm, there is a strong demand towards user-centric services based on the context and service management. For implementing the user-centric services recent researches on the service-oriented system is focused on the context recognition, analysis and reasoning. Furthermore, the service-oriented systems also need to manage the structural and hierarchical context model based on ontology including the situational and environmental contexts. These systems enhance the system scalability and increase the service diversity according to the service reasoning and the context analyzing. However, the conventional systems spend too much time and effort on the context analysis and the pattern learning. Moreover, the similarity of users, spaces, and service characteristics has not been adequately considered in the conventional systems. These facts reduce the effectiveness of the pattern learning and the service prediction. Therefore, we propose a Service-oriented Multi Agent Middleware which provides the information fusion and the group-based situation management for effectively managing the service prediction in pervasive environments. Our system dynamically analyzes the group-based situation and creates the novel services with patterns and policies. Our system reduces the service prediction delay up to 12% at the real home testbed and emulation system.