Online learning and adaptation for intelligent embedded agents operating in domestic environments

  • Authors:

  • Hani Hagras;Victor Callaghan;Martin Colley;Graham Clarke;Hakan Duman

  • Affiliations:

  • Department of Computer Science, University of Essex, Wivenhoe Park, Colchester, CO4 3SQ, England;Department of Computer Science, University of Essex, Wivenhoe Park, Colchester, CO4 3SQ, England;Department of Computer Science, University of Essex, Wivenhoe Park, Colchester, CO4 3SQ, England;Department of Computer Science, University of Essex, Wivenhoe Park, Colchester, CO4 3SQ, England;Department of Computer Science, University of Essex, Wivenhoe Park, Colchester, CO4 3SQ, England

  • Venue:
  • Autonomous robotic systems
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this chapter we show how intelligent embedded agents situated in an intelligent domestic environment can perform learning and adaptation. A typical domestic environment provides an environment where there is wide scope for utilising computer-based products to enhance living conditions. Intelligent embedded agents can be part of the building infrastructure and static in nature (e,g. lighting, HVAC etc.), some will be carried on the person as wearables, others will be highly mobile, as with robots. Both non-intrusive and interactive learning modes (including a mix of both) are used, depending on situation of the agent. For instance mobile robotic agents use an interactive learning whilst most building based agents use non-intrusive background learning modes. In this chapter we will introduce the learning and adaptation mechanisms needed by the Building and Robotic embedded agents to fulfil their missions in intelligent domestic environments. We also present a high-level multi embedded-agent model, explaining how it facilitates inter-agent communication and cooperation between heterogeneous sets of embedded agents within a domestic environment.