The challenge of intelligence in pervasive computing scenarios is a difficult and fascinating one. While MASs (Multi-Agent Systems) provide for a principled approach to pervasive intelligent systems, large scale and physical distribution call for expressive coordination models and technologies promoting distribution of intelligence. Also, the issue of situatedness of intelligence mandates effective agent models and technologies enabling the engineering of agent-based systems immersed within dynamic and articulated pervasive computing environments. Moving MAS coordination toward situatedness -- from both a theoretical and a technological viewpoint -- is then a fundamental step toward intelligent MAS in pervasive computing scenarios. In this paper we show how a language for MAS coordination --^sReSpecT -- can be extended according to a meta-model which provides a simple and solid grounding for MAS situatedness --^sA&A (Agents & Artifacts) -- and present a simple case study where the extended A&A ReSpecT is used to coordinate an agent-oriented pervasive system for sensing and controlling environmental properties.
展开▼
