phd thesis



In this work, we study the relationship between topological metrics of Wireless

Ad Hoc Networks and the network performance. We are interested in applying

different concepts and metrics related to the network topology to three different

network models, namely (i) wireless sensor networks (WSNs), (ii) mobile ad

hoc networks (MANETs), and (iii) vehicular ad hoc networks (VANETs). These

three models cover a wide variety of network topologies, ranging from typically

static or nearly static topologies (WSNs) to highly dynamic topologies such as the

ones present in VANETs. The main contributions of this work are: firstly, we propose

an expressive topology model able to describe a wide variety of deployment

strategies for WSNs. We present a topology-related feature estimator derived

from the betweenness metric, suitable for representing the energy depletion related

to the sensor relay task in WSNs. We developed a distributed algorithm to

compute this metric, which was used to design a routing algorithm that aims to

make a fair balance of the relay task of nodes in a WSN. For MANETs, we developed

a new localization system for Internet capable devices, based on A-GPS

technology, which offloads the GPS raw signal data to the cloud. We show that

this technique is able to reduce the energy consumption up to 80% when compared

to traditional A-GPS. To tackle with the highly dynamic topologies present

in VANETs, we proposed the use of a cooperative target tracking solution to track

the quick changes of the topologies due to the high velocity of vehicles and used

this solution to improve the performance of a video distribution mechanism over


You can download the full document by using the link below.