Thèse de doctorat en Informatique
Sous la direction de Kamel Barkaoui.
Pas de résumé disponible.
This dissertation deals with a proposition of new efficient energy protocols for wireless sensor networks (WSN) which are characterized by a low traffic, a high latency and a limited and non renewable energy. To do so, several works dealing with communication architecture, routing protocols and medium access control (MAC) protocols for Ad hoc networks and WSN were firstly analyzed. This, allowed us to propose a new MAC (ECo-MAC) and routing protocols that benefit from the low traffic feature of the majority of WSN applications enabling thus the saving of energy and an increase of the network lifetime. These protocols are used to create two cross-layered models for the two nodes sensor and base station (BS). To evaluate our proposition, we have used these models in the Optimized Network Engineering Tool (OPNET) simulator. The obtained results, under different network configurations, show important gains in terms of energy saving and latency, compared to works based on the S-MAC protocol. In a second step, we build a sensor and a BS models according to the timed automata formalism and using UPPAAL model checker tool we check the correctness of the new ECo-MAC protocol behaviour and the time slot duration in the worst case. Finally, from a discrete time Markov chains (DTMCs) modelling, we provide, by using probabilistic symbolic model checker (PRISM) tool, a formal justification of our adoption for some parameters values of the backoff procedure of our ECo-MAC protocol.