Ordonnancement monoprocesseur pour les applications temps réel récupérant l énergie ambiante
Energy management is a central problem in real-time systems design, in particular for embedded wireless devices. Whereas a great number of work was carried out in order to minimize the energy consumption and maximize the lifetime achieved as in classical battery operated devices, few stud- ies have...
Enregistré dans:
Auteurs principaux : | , , , , , |
---|---|
Collectivités auteurs : | , , |
Format : | Thèse ou mémoire |
Langue : | anglais |
Titre complet : | Ordonnancement monoprocesseur pour les applications temps réel récupérant l énergie ambiante / Hussein El Ghor; sous la direction de Maryline Chetto |
Publié : |
Nantes :
Université de Nantes
, 2012 |
Description matérielle : | 1 vol. (166 p.) |
Note de thèse : | Thèse de doctorat : Automatique et informatique appliquée, systèmes temps réel : Nantes : 2012 |
Disponibilité : | Publication autorisée par le jury |
Sujets : |
Résumé : | Energy management is a central problem in real-time systems design, in particular for embedded wireless devices. Whereas a great number of work was carried out in order to minimize the energy consumption and maximize the lifetime achieved as in classical battery operated devices, few stud- ies have been devoted to emerging harvesting systems which need to operate perennially thanks to the environmental energy. In such systems, the consumed energy comes at the same time from the computing nodes and the supports of transmission/reception. In this thesis, we focus on systems that receive their energy from an environmental source (for example, solar energy). Under these assumptions, the objective is to ensure a perennial functioning of the real-time system that must be able to operate indefinitely in the respect of its timing require- ments. So the problem we have to deal with is: How can we schedule tasks so as to guarantee their timing constraints perpetually by suitably exploiting both the processor and the available ambient energy. These problems can be solved by different types of scheduling approaches: Clairvoyant scheduling strategies which take into account the characteristics of the energy source, the capacity of the energy storage as well as the energy consumed by the tasks, and time. Non clairvoyant scheduling strategies which take decisions at run-time without any prior knowledge about the future energy production and task characteristics. In this thesis, we intoduce new scheduling strategies dedicated to real-time energy harvesting applications. Our objective is to prove both the superiority of the proposed schedulers and their integrability thanks to comparative performance experiments |
---|---|
Variantes de titre : | Uniprocessor scheduling for real time energy harvesting applications |
Notes : | Titre provenant de la page de titre du document électronique L'impression du document génère 182 p. Ecole(s) Doctorale(s) : École doctorale sciences et technologies de l'information et de mathématiques (STIM) (Nantes) Autre(s) contribution(s) : Frank Singhoff (Président du jury) ; Rafic Hage Chehade, Damien Masson (Membre du jury) ; Pascal Richard (Rapporteurs) |
Configuration requise : | Un logiciel capable de lire un fichier au format pdf |
Bibliographie : | Bibliogr. p.159-166, 116 réf. |