University of Patras
Wireless sensor and actor networks (WSANs): αλγόριθμοι αναγνώρισης γεγονότων βασισμένων σε τεχνικές ασαφούς λογικής με πρόβλεψη για απαίτηση απόκρισης πραγματικού χρόνου
Abstract
dc:descriptionEvent detection in real environments is a research area that has attracted a lot ofinterest. The discipline of artificial intelligence seems to have gained significant ground.The main objective of these areas is the creation of appropriate algorithms and buildingmodels on real data sets, so that the evaluation metrics are able to confirm their abilityto detect whether or not the existence of real events. Assuming a reliable system(evaluated by the relevant metrics (Accuracy, Sensitivity, Specificity)) the basic researchproblem is to investigate the performance of the above system when the application ishosted in a WSN topology, and more particularly when the system is executed in realWSN hardware platforms which are characterized by limited-resources and under theerror-prone nature of wireless communications. Hence, the problem is examined fromtwo perspectives: i.) What are the limitations during the development and theimplementation of such systems in WSN infrastructures and ii.) how the performance ofsuch systems is affected while the Network QoS varies within the network topology.For these reasons we designed and developed two fuzzy logic systems, whichevaluated in the Matlab environment. Then the above systems implemented in C andexecuted by the embedded WSN platforms TelosB hardware platforms. Theexperimental evaluation confirmed our initial beliefs about the impact of networkconditions on the performance of the algorithm and the monolithic approach of localprocessing demonstrated to affect vital to the application itself.Driven by these observations, we propose novel techniques in order to implement fuzzylogic systems in WSN infrastructures targeting optimal resource allocation within anetwork topology. Towards this objective, the technique denoted as Directed AcyclicGraph to Network Graph (DAG2NG) is presented leading to the distributed implementation of the Fuzzy Logic. Moreover, the event triggered execution of theFuzzy Inference Systems is implemented in order to achieve optimal CPU utilization.The task scheduling problem, is extended beyond the framework of the fuzzy logic. Sowe design, study and evaluate a series of scheduling models that allow the distributionof tasks within the wireless nodes. The application of these models in the field of WSNsallows the study of the energy efficiency, response time and justice between nodes in aseries of algorithmic problems (parallel, nonparallel, hybrid), which can exist in variousWSN applications.The semantic relationship between the wireless nodes, is another object of this thesis.Both the event triggered implementation of fuzzy logic systems and the management ofthe nodes participating in a task assignment model, involve the discovery of semanticneighbors. So we propose two new semantic discovery algorithms neighbors, whom weevaluate.The thesis concludes with the proposal of a new algorithm that seeks to give a newdimension and answer to the problem of node-to-actor coordination.The assignment of the sensors at specific actors, taking into account the actors’capabilities as well as the heterogeneity of the network, lead to the proposal of a newsensor-to-actor coordination algorithm based on the Voronoi tesselation.
Degree
thesis:*- Grantor dc:publisher
- University of Patras
- Year dc:date
- 2016
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Dima, Sofia-Maria
- Δήμα, Σοφία-Μαρία
Subjects
dc:subject × 20- Aναγώριση Γεγονότων
- Ασύρματα Δίκτυα Αισθητήρων Ενεργοποιητών
- Καταμερισμός εργασιών
- Σημασιολογικοί Γείτονες
- Συντονισμός κόμβων ενεργοποιητών
- Ασαφής λογική
- Event detection
- Wireless Sensor and Actor Networks
- Task Scheduling
- Semantic Neighbors
- Node-to-Actor Coordination
- Fuzzy logic
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Natural Sciences
- Computer and Information Sciences
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/38442