Publication: Arquitectura de pizarras distribuidas para sistemas de Inteligencia Ambiental
Loading...
Files
Official URL
Full text at PDC
Publication Date
2015-02-06
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad Complutense de Madrid
Citations
Exportar
Abstract
La información en un sistema de IAm (Inteligencia Ambiental) es de naturaleza diversa. Los datos sin procesar de los sensores se agregan y ltran para crear información abstracta, que puede ser procesada por componentes que observan sus cambios para decidir qué acciones tomar. Este proceso involucra las siguientes tareas: encontrar las fuentes de información disponibles y sus tipos, reunir los datos proporcionados por estas fuentes, facilitar la fusi ón (agregación y derivación) de los fragmentos de información, y actualizar la representación de este contexto para que sea usada por aplicaciones. El proceso inverso también aparece en los sistemas IAm, cuando cambios en la representación del contexto disparan acciones en dispositivos actuadores. Además, los dispositivos pueden ser agregados, cambiados o fallar, modi- cando la topología del sistema. También, pueden cambiar otras circunstancias afectando a las actividades de los usuarios, p.ej., el tiempo, la localizaci ón, o la presencia de otros usuarios. Estos cambios modi can la información que el sistema tiene disponible para satisfacer las necesidades de los usuarios (es decir, el contexto). Los sistemas IAm, deben adaptarse a estas condiciones cambiantes para ser capaz de proporcionar sus servicios, siendo tan poco intrusivos como sea posible. También hay requisitos de rendimiento que el sistema debe cumplir para proporcionar respuestas en tiempo real a los estímulos del entorno. Los mecanismos de control oportunista abordan estos aspectos monitorizando el contexto, y suspendiendo o resolviendo objetivos cuando las condiciones apropiadas se cumplen. Este trabajo presenta FAERIE (Framework for AmI: Extensible Resources for Intelligent Environments), una infraestructura que aborda los requisitos anteriores facilitando la gestión y fusión de información del contexto a distintos niveles. Está implementada siguiendo un modelo de pizarras distribuidas. Cada nodo del sistema tiene una pizarra privada para gestionar fragmentos de información a los que pueden acceder componentes observadores, ya sea local o remotamente (desde otras pizarras) de una forma transparente. También proporciona facilidades para implementar patrones de adaptación tales como a cambios en la topología o a ciertas actividades en un ujo de trabajo. La infraestructura se ofrece con un conjunto de elementos para facilitar su uso en el desarrollo de aplicaciones IAm. Concretamente, se proporciona una guía de desarrollo, que explica el proceso desde la extracción de requisitos a la implementación y despliegue siguiendo la arquitectura. Además, la infraestructura software viene con un conjunto de herramientas orientadas a hacer un desarrollo con integración continua de los módulos software. Este concepto persigue controlar en todo momento el estadio de desarrollo y la salud del software, entre otros aspectos, mediante la ejecución y validación automática de pruebas, bajo distintos despliegues y considerando diferentes casos de uso. Para automatizar en la mayor medida posible todo el proceso de validación de una aplicación, se ha desarrollado un adaptador de la infraestructura para hacer uso de una plataforma de simulación de entornos 3D. De esta manera, las aplicaciones pueden desplegarse en el entorno simulado como si fuese el real, abaratando los costes de detección de errores en las primeras fases de pruebas. Frente a otras arquitecturas, FAERIE ofrece un marco de desarrollo más robusto y sencillo, que otorga su ciente control para manejar la complejidad de este tipo de aplicaciones, y reduce el esfuerzo del seguimiento y validación de su desarrollo. Con respecto a su propia arquitectura, el modelo de pizarras distribuidas permite que la obtención del contexto sea transparente para su uso, manteniendo la escalabilidad y evitando el problema de tener un único punto de fallo. Aparte de esto, el modelo de procesamiento del contexto utilizado es menos restrictivo que el de otras arquitecturas, permitiendo a cada componente del sistema acceder a información a cualquier nivel de abstracción.
Information in an AmI (Ambient Intelligence) system has diverse natures. Raw data coming from sensors are aggregated and ltered to create more abstract information, which can be processed by application components that observe it to decide what actions should be performed. This process involves several activities: nding the available sources of information and their types, gathering the data from these sources, facilitating the fusion (aggregation and derivation) of the di erent pieces of data, and updating the representation of this context to be used by applications. The reverse process also appears in AmI systems, when changes in the context representation trigger actions in actuator devices. In addition, these devices can be added, change or fail, modifying the system topology. Also, other circumstances may change, a ecting users' activities, e.g., time, location, or the presence of other users. These changes modify the information the system has available to satisfy users' needs (i.e., the context). AmI systems need to adapt to these evolving conditions in order to be able to provide their services, but being as unobtrusive as possible for their users. There are also performance requirements that a system must ful ll to provide responses to environment stimuli in real time. Opportunistic control mechanisms address these issues by monitoring the context, and suspending or resolving goals when the appropriate conditions are met. This work presents FAERIE (Framework for AmI: Extensible Resources for Intelligent Environments), a framework that addresses the previous requirements by providing means for the management and fusion of context information at di erent levels. It is implemented as a distributed blackboard model. Each node of the system has a private blackboard to manage pieces of information that can be accessed by observer components, either locally or remotely (from other blackboards) in a transparent way. It also provides facilities to implement adaptation patterns, such as adaptation to topology changes or to certain activities in a workflow This framework is o ered with a set of elements to facilitate its use for developing AmI applications. Concretely, a development guide is provided, which explains the process from the extraction of requisites to the implementation and deployment using the architecture. In addition, the framework software comes with a set of tools oriented to make a development with continuous integration of the software modules. This concept pursues controlling at every moment the development state and the health of the software, among other aspects, via automatic execution and validation of tests, under di erent deployment and use cases. In order to automatize most of the validation process of an application, an adapter to execute 3D simulations has been developed for the framework. This way, applications can be deployed in a simulated environment as if it was the real one, reducing the costs of error detection in the early stages of testing. Compared to other architectures, FAERIE o ers a more robust and simpler development framework, which gives enough control to handle the complexity of this kind of applications, and reduces the e ort of their tracking and validation. Regarding its own architecture, the distributed blackboard model makes context acquisition transparent for its use, maintaining the scalability and avoiding the problem of having a single point of failure. In addition, the context processing model used is less restrictive than in other architectures, allowing each component to access information at any abstraction level.
Information in an AmI (Ambient Intelligence) system has diverse natures. Raw data coming from sensors are aggregated and ltered to create more abstract information, which can be processed by application components that observe it to decide what actions should be performed. This process involves several activities: nding the available sources of information and their types, gathering the data from these sources, facilitating the fusion (aggregation and derivation) of the di erent pieces of data, and updating the representation of this context to be used by applications. The reverse process also appears in AmI systems, when changes in the context representation trigger actions in actuator devices. In addition, these devices can be added, change or fail, modifying the system topology. Also, other circumstances may change, a ecting users' activities, e.g., time, location, or the presence of other users. These changes modify the information the system has available to satisfy users' needs (i.e., the context). AmI systems need to adapt to these evolving conditions in order to be able to provide their services, but being as unobtrusive as possible for their users. There are also performance requirements that a system must ful ll to provide responses to environment stimuli in real time. Opportunistic control mechanisms address these issues by monitoring the context, and suspending or resolving goals when the appropriate conditions are met. This work presents FAERIE (Framework for AmI: Extensible Resources for Intelligent Environments), a framework that addresses the previous requirements by providing means for the management and fusion of context information at di erent levels. It is implemented as a distributed blackboard model. Each node of the system has a private blackboard to manage pieces of information that can be accessed by observer components, either locally or remotely (from other blackboards) in a transparent way. It also provides facilities to implement adaptation patterns, such as adaptation to topology changes or to certain activities in a workflow This framework is o ered with a set of elements to facilitate its use for developing AmI applications. Concretely, a development guide is provided, which explains the process from the extraction of requisites to the implementation and deployment using the architecture. In addition, the framework software comes with a set of tools oriented to make a development with continuous integration of the software modules. This concept pursues controlling at every moment the development state and the health of the software, among other aspects, via automatic execution and validation of tests, under di erent deployment and use cases. In order to automatize most of the validation process of an application, an adapter to execute 3D simulations has been developed for the framework. This way, applications can be deployed in a simulated environment as if it was the real one, reducing the costs of error detection in the early stages of testing. Compared to other architectures, FAERIE o ers a more robust and simpler development framework, which gives enough control to handle the complexity of this kind of applications, and reduces the e ort of their tracking and validation. Regarding its own architecture, the distributed blackboard model makes context acquisition transparent for its use, maintaining the scalability and avoiding the problem of having a single point of failure. In addition, the context processing model used is less restrictive than in other architectures, allowing each component to access information at any abstraction level.
Description
Tesis inédita de la Universidad Complutense de Madrid, Facultad de Informática, Departamento de Ingeniería del Software e Inteligencia Artificial, leída el 16-07-2014