Abstract
With the growth of the Internet of Things, the sheer amount of wireless traffic doosnot stop to increase and will cause network congestion in a near future. This requires
a rethinking of the way we process information, especially in high-density networks.
This subject has attracted a lot of interest for the average consensus problem.
In this work, we first provide an overview of actual technology for processing information
in sensor network and review the basis of low-complexity sensor and
high-density network.
Then, we present the Peper’s Average Consensus algorithm in asymmetric broadcasting
wireless sensor networks in asynchronous case. We provide a new mathematical
model for this algorithm and expose numerical results about the impact of
network topology.
After that, we develop a new Average Consensus algorithm with multiple updates
as an improvement of the Peper’s algorithm and prove its better performance through
computer simulations. We also highlight links between some algorithm parameters
and give a mathematical model of this algorithm.
Finally, we show a Clustering Consensus algorithm to solve the problem of computing
average inside clusters instead of the whole network. We expose our motivations
and describe different policies before presenting a numerical evaluation.
Date of Award | 20 Jun 2017 |
---|---|
Original language | English |
Awarding Institution |
|
Supervisor | Anthony Cleve (President) & Marie-Ange Remiche (Supervisor) |