An Adaptive Architecture for Ambient Intelligence Based on Meta-modeling, Smart Agents, and Wireless Sensor Networks

Abstract

In this paper, an Ambient Intelligence (AmI) architecture for devising and implementing a smart laboratory in a university is presented. The architecture relies on a novel meta-model that abstracts the laboratory ambient using a context model, an adaptation model, an end-user model, and a domain model. Thus, the architecture offers, to the end-users, a set of adaptive AmI services that are supplied by smart agents and wireless sensor networks. The architecture was coded using SQLite, Jade, Protege, and CLIPs and simulated under Oracle's Solarium environment. The latency and scalability of the proposed architecture, as well as its ability to supply effectively the AmI services, were tested using both qualitative and quantitative standard performance metrics.