Using Embedded Agents to Create Ambient Intelligence
Thursday 17 November 2005, 1300-1400
By Dr Hani Hagras, University of Essex
Embedded agents are embedded computational artefacts integrated with intelligent reasoning and learning mechanisms. The embedded agents are able to personalise themselves to the user's needs and preferences by learning from their behaviour and thus configuring and controlling the user's environments on their behalf. Thus these agents can reduce the cognitive load associated with configuring and programming ubiquitous computing environments to create Ambient Intelligence. Due to the limited processing power and memory capacities of embedded computers, the intelligent mechanisms that are used in embedded-agents have to be computationally undemanding. Moreover, the intelligent approaches applied should have transparent internal representations such that their learnt behaviours and decisions can be represented in the form of human readable rules which are accessible to the end user.
This talk will introduce different novel AI techniques that can be used in the agents like type-2 fuzzy logic. The talk will also present the application of these agents to create ambient intelligence within smart environments and the talk will present results obtained from the real test bed of the Essex iDorm.