Evolving Type-2 Fuzzy Agents for Ambient Intelligent Environments

Abstract

This paper presents an overview of our work to produce type-2 fuzzy agents that can realize an intelligent ambience in everyday environments to form Ambient Intelligent Environments (AIEs). The agents are embedded in the user environment where they learn the user behavior in a non intrusive mode and control the environment on the user behalf to realize the intelligent ambience. Type-2 fuzzy systems are able to handle the different sources of uncertainty and imprecision encountered in AIEs to give a very good response. However, there is a need to evolve the type-2 agents by evolving the type-2 Membership Functions (MFs) and rules in a life long learning mode to handle and accommodate for the uncertainties associated with the long term operations and the changing environmental conditions and user preferences. This paper will present an overview of the evolving type-2 agents which will be evaluated in real world test beds for intelligent environments.