Supervision : Guy PUJOLLE

Multipath Routing Protocol for Ad Hoc Networks

Mobile ad hoc networks (MANETs) consist of a collection of wireless mobile nodes that move freely and self-configure without a preexisting communication infrastructure.

In MANETs, under the arbitrary movement of nodes and highly dynamic topology conditions, the routing protocols for fixed networks do not perform well. To adapt to the rapid change of topology, many routing protocols have been proposed. However, most of them chose the minimum hop-count routes. This choice can lead to significant reduction of the network performance because communication quality via the chosen links is not good. Thus, the evaluation of link quality is indispensable and interference is a key factor to be considered. In the literature, most of the existing interference-aware protocols are single path and the consideration for interference range is limited. Several proposed interference-aware multi-path protocols for MANETs are either not highly efficient or have the computational complexity of NP-hard.

To tackle the above problems, we first determine an interference region of each link including all nodes that can interfere with this link. Then, we propose a formula to evaluate the interference level of each link. Based on the formula of the link interference, we develop a series of interference-aware multi-path routing protocols for mobile ad hoc networks. These interference-aware multi-path routing protocols minimize the impact of interference on the paths from the source to the destination to increase the network performance. The computational complexity of the protocols is polynomial and each of them is most suitable for a type of network size.

To demonstrate the efficiency of the interference-aware multi-path routing protocols, we compare them to prominent protocols in different network sizes and a high mobility environment when the RTS/CTS (Request-To-Send/Clear-To-Send) mechanism is alternatively turned on and turned off. The simulation results show that our protocols significantly improve packet delivery fraction, and reduce end-to-end delay, routing overhead, and normalized routing load.

The RTS/CTS mechanism is to restrict the collision caused by the hidden terminals. However, we show that the RTS/CTS mechanism is not suitable for mobile ad hoc network based on the simulation results.

Phd defence : 09/28/2012

Jury members :

André-Luc BEYLOT, Professeur, IRIT/ENSEEIHT, Toulouse [Rapporteur]
Steven MARTIN, HDR à l'Université Paris-Sud [Rapporteur]
Dominique GAÏTI, Professeur, Université de Technologie, Troyes
Thi Mai Trang NGUYEN, Maître de Conférences, UPMC, Paris
Guy PUJOLLE, Professeur, UPMC Sorbonne Universités

2012 Publications

• 2012
  • Ph. Le : “Multipath Routing Protocol for Ad Hoc Networks”, thesis, phd defence 09/28/2012, supervision Pujolle, Guy (2012)