- Courses
- Sistemi Multi-Agente LM 2010/2011
- Projects
- JaSA: semantically aware agents to improve adaptivity and context-awareness of smart environments
JaSA: semantically aware agents to improve adaptivity and context-awareness of smart environments
JaSA: semantically aware agents to improve adaptivity and context-awareness of smart environments
Authors
Abstract
The aim of this project is to accomplish the need for adaptivity and context-awareness in smart environment applications, combining agent-oriented programming with ontological reasoning. This concretises in the realization of Jason Semantically Aware (JaSA) agents, that will show their potentiality using JaCa-smart, an extension of the Jason-CArtAgO framework, to act in the Smart-M3 computational environment.
Outline
- Introduction
- Towards semantically aware agents
- Interaction model for Smart Environment Applications
- Agent-Oriented Programming is the way
- Use of BDI agents in Smart Environment Systems
- Integrating BDI agents in Smart-M3
- Smart-M3: an open source platform within SOFIA project
- JaCa-smart
- Goal of this work
- JaSA agents
- Working of a Jason agent
- Needed Modifications
- Modifications to the belief base
- Modifications to the Plan Library
- Modifications to the "overall agent architecture"
- Further Modifications
- JaSA project
- From JASDL to JaSA
- From CArtAgO to CArtAgO-DL
- Artifact manuals as OWL ontologies
- A new artifact to interact with Smart-M3 SIB(s)
- New artifact operations interfaces
- SIB events handling
- How to build a JaSA MAS - the thermostat example
- Problem: controlling the temperature of a room
- Setting up the MAS configuration file: thermo.mas2j
- MAS thermo artifacts
- MAS thermo agents
- Conclusions & future work
- Acknowledgements
- References
Paper and bibliography
The paper and bibliography related to JaSA project can be found in the attached
.Code repository
The sources of JaSA technology, including the examples, can be downloaded here: http://cartago.svn.sourceforge.net/viewvc/cartago/jasa/trunk/?view=tar