Skip to Main content Skip to Navigation

Protocoles localisés de diffusion et économie d'énergie dans les réseaux ad hoc et de capteurs

François Ingelrest 1, 2
2 POPS - System and Networking for Portable Objects Proved to be Safe
LIFL - Laboratoire d'Informatique Fondamentale de Lille, Inria Lille - Nord Europe, IRCICA
Abstract : A wireless ad hoc network is composed of a decentralized set of mobile and self-organized hosts. Such a network does not rely on any infrastructure, and is thus completely autonomous and dynamic. Based on these concepts, any set of objects equipped with appropriate communication interfaces may spontaneously form such a network. As wireless communications are used, only closest hosts may directly communicate with each other. Long distance communications must be done through a multi-hopping mechanism: this simply means that intermediate objects have to forward the packets until the destination is reached.

Wireless sensor networks are similar to ad hoc networks, as they are also decentralized and autonomous. A sensor is a small device with monitoring capabilities, allowing it to gather information on its environment. For instance, typical use case of sensor networks may be the monitoring of military areas (movements detection) or forests (heat detection).

Among the common problems studied in these networks is broadcasting. In such communication, a message is sent from a given host to all the other ones in the network. Applications of this process are numerous: route discovery, synchronization\ldots{} As mobile objects rely on a battery, it is mandatory for the broadcast protocol to be energy efficient. This is generally done by reducing the quantity of needed relaying nodes, or by limiting the transmission power at each host. Of course, all these optimizations must preserve the reliability of the protocol.

The goal of my work was to study broadcasting in ad hoc and sensor networks, in order to highlight the characteristics and the flaws of existing mechanisms, and to propose new efficient protocols. In all our proposals, we always tried to provide realistic solutions: a lot of previous studies was based on centralized mechanisms, where a global knowledge was needed to perform the broadcasting process. We focused on reliable and localized solutions, using only close information on the neighborhood of each node. This kind of solution is also more scalable, as the quantity of needed information does not grow with the size of the network. Some of our studies, especially in \cite{2005-medhoc}, show that these solutions may be as efficient as centralized ones.

As power adjustment is a a mechanism of prime importance for energy preservation, we proposed in \cite{2006-tpds, 2004-medhoc, 2004-icc} an original broadcasting method based on the concept of optimal communication range. The latter is theoretically computed thanks to the considered energy model, and is the best compromise between the spent energy at each node and the quantity of relays. We thus proposed two different protocols based on this concept, each of them being specifically targeted either at ad hoc (TR-LBOP) or sensor networks (TR-DS).

To further reduce energy savings, we studied a well-known centralized and very efficient protocol named BIP. Its efficiency is due to its consideration of the coverage obtained thanks to a single omnidirectional transmission, instead of considering all links separately. We proposed in \cite{2007-winet, 2005-iscc} a localized solution based on BIP, in order to incrementally construct a broadcasting structure. We experimentally showed that results are very close to those provided by BIP, especially in highest densities, while only local information is needed at each node.

We finally considered the removal of a strong hypothesis, broadly spread in the ad hoc and sensor networks community, referred to as the unit disk graph. The latter defines the communication area of each device to be a perfect circle. In \cite{2006-msn}, we replaced this hypothesis by a more realistic one in order to study a well-known broadcasting mechanism, the multipoint relay protocol (MPR). We showed that MPR does not provide anymore sufficient performance in such a realistic environnement. We also proposed efficient modifications in order to get back efficient results.
Complete list of metadata
Contributor : François Ingelrest <>
Submitted on : Tuesday, November 14, 2006 - 7:01:35 PM
Last modification on : Wednesday, April 7, 2021 - 3:00:01 PM
Long-term archiving on: : Tuesday, April 6, 2010 - 7:35:10 PM


  • HAL Id : tel-00113869, version 1



François Ingelrest. Protocoles localisés de diffusion et économie d'énergie dans les réseaux ad hoc et de capteurs. Réseaux et télécommunications [cs.NI]. Université des Sciences et Technologie de Lille - Lille I, 2006. Français. ⟨tel-00113869⟩



Record views


Files downloads