|
Identifier |
000423793 |
Title |
Decentralized online network performance monitoring for wireless sensor/actuator networks |
Alternative Title |
Αποκεντρωμένη επιγραμμική παρακολούθηση επιδόσεων δικτύου για ασύρματα δίκτυα αισθητήρων / επενεργητών |
Author
|
Τζουγκαράκης, Αντώνιος Κ.
|
Thesis advisor
|
Τσακαλίδης, Παναγιώτης
|
Reviewer
|
Δημητρόπουλος, Ξενοφώντας
Τζαγκαράκης, Γεώργιος
|
Abstract |
Wireless Sensor Networks (WSNs) have been introduced to our daily lives in several
application domains such as Smart Water, Smart Grids, Smart Homes and Health
Monitoring. All these applications require the production of robust and reliable
applications which are based on the WSN's optimal performance. Various factors can
affect the operation of a WSN ranging from the operational space, the adopted
communication protocol, the intra-network dynamics and the status of each individual
node.
The main goal is the unattended and the continuous operation of the network in real-life
deployments. As such, characterization of the network's high-level performance when it
is based exclusively on link-quality estimation can yield episodic snapshots of the
performance specific point-to-point links.
The objective of this thesis is to characterize network performance beyond the
constraints of 1st hop neighbors and across different layers of a fully functional protocol
stack, ranging from the Physical to the Transport and Application layers. Heterogeneous
metrics are fused and machine learning methods provide the required means to discover
patterns in the data and provide the features that are the most dominant ones by
employing feature selection techniques in an unsupervised fashion. Thus, systematic
study of end-to-end links' performance could provide the means for understanding the
multi-dimensional behavior of an entire network.
|
Language |
English |
Subject |
Network measurement and analysis |
|
Testbeds and experimental evaluation |
|
Unsupervised learning |
|
Wireless sensor networks |
|
Δοκιμαστικές μονάδες και πειραματική αξιολόγηση |
|
Μέτρηση και ανάλυση δικτύων |
|
Μη εποπτευόμενη μάθηση |
Issue date |
2019-07-26 |
Collection
|
School/Department--School of Sciences and Engineering--Department of Computer Science--Post-graduate theses
|
|
Type of Work--Post-graduate theses
|
Views |
537 |