23.08.2015 Views

Here - Agents Lab - University of Nottingham

Here - Agents Lab - University of Nottingham

Here - Agents Lab - University of Nottingham

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Although some success [13] [14] have been made in the literature for solvingthe joint action and configuration planning problems at a global level for heterogeneousrobotic ecologies or to enable their purely reactive configuration [15], wepropose to distribute certain aspects <strong>of</strong> these tasks over a number <strong>of</strong> agents withvarying responsibilities and functionalities in order to lead to better scalability,robustness and fault tolerance.Agent and Multi Agent Systems (MASs) are regarded as a general-purposeparadigm to facilitate the co-ordination <strong>of</strong> complex systems built in terms <strong>of</strong>loosely-coupled, situated, autonomous and social components (the agents). Inparticular, the Belief Desire Intention (BDI) agent model, used in the Self -OSGiframework discussed in the following section, provides a simple but extensiblemodel <strong>of</strong> agency that explicitly addresses the production <strong>of</strong> rational and autonomousbehaviour by agents with limited computational resources.3 Self -OSGiSelf -OSGi [3] [4] is a BDI agent framework built on OSGi Java technology andpurposefully designed to support the type <strong>of</strong> collaboration envisaged within theubiquitous and embedded systems targeted by RUBICON.The component & service orientation used in the design <strong>of</strong> Self -OSGi is anhighly popular, mainstream approach used to build modular s<strong>of</strong>tware systems.Component & service frameworks operate by posing clear boundaries (in terms<strong>of</strong> provided and required service interfaces) between s<strong>of</strong>tware components andby guiding the developers in re-using and assembling these components intoapplications. Self -OSGi addresses the lack <strong>of</strong> common adaptation mechanismsin these frameworks by leveraging their previously unexploited similarities withthe BDI agent model.3.1 OSGiOSGi specification [5] is currently the most widely adopted technology for buildingmodular control systems for networked home applications, with many implementationstargeting computationally constrained platforms. Within the AALdomain, OSGi-based middleware have long been used to provide the technicalbasis for integrating network devices and services, e.g. in EU projects such asAmigo, OASIS, SOPRANO, and their recent consolidation in the UniversAALplatform.OSGi defines a standardised component model and a lightweight containerframework, built above the JVM, which is used as a shared platform for networkprovisionedservices and components specified through Java interfaces and Javaclasses. Each OSGi platform facilitates the dynamic installation and management<strong>of</strong> units <strong>of</strong> deployment, called bundles, by acting as a host environmentwhereby various applications can be executed and managed in a secured andmodularised environment. An OSGi bundle is packaged in a Jar file and organisesthe frameworks internal state and manages its core functionalities. These73

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!