Intelligent Association Exploration and Exploitation of Fuzzy Agents in Ambient Intelligent Environments

This paper presents a novel fuzzy-based intelligent architecture that aims to find relevant and important associations between embedded-agent based services that form Ambient Intelligent Environments (AIEs). The embedded agents are used in two ways; first they monitor the inhabitants of the AIE, learning their behaviours in an online, non-intrusive and life-long fashion with the aim of pre-emptively setting the environment to the users preferred state. Secondly, they evaluate the relevance and significance of the associations to various services with the aim of eliminating redundant associations in order to minimize the agent computational latency within the AIE. The embedded agents employ fuzzy-logic due to its robustness to the uncertainties, noise and imprecision encountered in AIEs. We describe unique real world experiments that were conducted in the Essex intelligent Dormitory (iDorm) to evaluate and validate the significance of the proposed architecture and methods.