A Lightweight Architecture to Support Context-Aware Ubiquitous Agent System

Abstract

Autonomous agents or semi-autonomous agents feature largely in their dynamic adaptation of its behaviors for changing environments to achieve some set of goals, especially in ubiquitous environments. It demands mechanism for coherently satisfying agent goals depending on changing availability of resources one the fly. Leveraging context-aware techniques and agent-oriented approaches, the paper proposes a lightweight architecture to build adaptive but predictable goal-driven ubiquitous agent systems based on OSGi (Open Service Gateway initiative), an open standard service-oriented framework. The proposed platform not only supports context acquisition, discovery and reasoning, but also provides a centric goal resolution mechanism using goal-tree to automatically specify service components for the satisfaction of agent goals. A Context Pair Language and the production rule are used as semantic basis to model contextual information as well as the goals.