Probabilistic Models and Algorithms for Energy-Efficient Large and Dense Wireless Sensor Networks

Research output: ThesisDoctoral ThesisMonograph

Abstract

The concept of sparsity is fundamental to understanding data collection in foreseeable large scenarios. This intuition is based on the fact that future analytics will only perform efficiently over compact information domains. This scenario is even more critical in Wireless Sensor Networks (WSN), since small sensors suffer from major limitations. Hence, Compressive Data Aggregation (CDA) is required to alleviate the work load of the sensors. This research has proposed a compressed sensing-based protocol which follows a random sensing principle. This principle allows for the characterization of sensor interactions using methods from stochastic geometry. This work will deliver: a) a methodology to jointly analyze the compression and communication aspects of CDA in WSNs and b) a novel protocol that effectively implements compressed sensing with collision avoidance. It is called Stochastic Compressive Data Aggregation (S-CDA). Although there have been great advances, there is a lack of methodologies to jointly analyze the compression and communication aspects of CDA in WSNs. Consequently, this dissertation seeks to bridge this gap.

Translated title of the contributionTodennäköisyyspohjaisia malleja ja algoritmeja energiatehokkaille ja tiheille langattomille anturiverkoille
Original languageEnglish
QualificationDoctor's degree
Awarding Institution
  • Aalto University
Supervisors/Advisors
  • Jäntti, Riku, Supervising Professor
  • Caamaño Fernández, Antonio J., Supervising Professor
Publisher
Print ISBNs978-952-60-6681-3
Electronic ISBNs978-952-60-6682-0
Publication statusPublished - 2016
MoE publication typeG4 Doctoral dissertation (monograph)

Keywords

  • compressed sensing
  • data aggregation
  • stochastic geometry

Fingerprint

Dive into the research topics of 'Probabilistic Models and Algorithms for Energy-Efficient Large and Dense Wireless Sensor Networks'. Together they form a unique fingerprint.

Cite this