Erweiterte Suche

Integration of sensor data by means of an event abstraction layer

Diese Doktorarbeit stellt eine Methodik vor, die aus Zeitreihen von Sensorbeobachtungen Events (Ereignisse) nahezu in Echtzeit ableiten und darstellen kann. Die Analyse der Sensor-Daten erfolgt unter zur Hilfenahme von Prozessen und Technologien des 'Semantic event processing'. Aus den Daten abgeleitete Ereignisse werden eindeutig und maschinenlesbar als Instanzen bestehender Wissensbasen (Ontologien) dargestellt. Der Einsatz einer erweiterten Form der 'Semantic Sensor Network' Ontologie ermöglicht in diesem Zusammenhang eine Modellierung von spezifischem Fachwissen in einer mehrstufigen Ontologie-Struktur. Infolgedessen können differenzierte Perspektiven verschiedener Fachgemeinschaften auf die gleichen Daten integriert und verglichen werden.

This thesis presents a methodology to infer and represent events from time series of sensor observations in near real-time. Semantic event processing is used to analyse sensor data. Inferred events are modelled using an extension of the Semantic Sensor Network ontology. Domain knowledge is represented in a multilevel ontology structure. The proposed methodology allows information communities to integrate different views on the same data.

Titel: Integration of sensor data by means of an event abstraction layer
Verfasser: Llaves Arellano, Alejandro GND
Gutachter: Kuhn, Werner
Organisation: FB 14: Geowissenschaften
Dokumenttyp: Dissertation/Habilitation
Medientyp: Text
Erscheinungsdatum: 09.01.2014
Publikation in MIAMI: 09.01.2014
Datum der letzten Änderung: 27.07.2015
Schlagwörter: Event Verarbeitung; Raumzeitliche Modellierung; Datenzusammenführung; Semantische Interoperabilität; Sensoren
event processing; spatio-temporal modelling; data integration; semantic interoperability; sensors
Fachgebiete: Informatik, Wissen, Systeme; Geowissenschaften, Geologie
Sprache: Englisch
Format: PDF-Dokument
URN: urn:nbn:de:hbz:6-54359642460
Permalink: https://nbn-resolving.org/urn:nbn:de:hbz:6-54359642460
Onlinezugriff:
Inhalt:
1 introduction 1
1.1 Overview 1
1.2 Motivation 3
1.2.1 The data avalanche in the Digital Earth 3
1.2.2 Modelling spatio-temporal change 4
1.2.3 Semantic interoperability problems 5
1.3 Research questions 6
1.4 Scope and Limitations of the Thesis 7
1.5 Thesis Outline 9
2 background and related work 11
2.1 Sensors and Environmental Monitoring 11
2.2 Semantic Web and Sensor Web 12
2.3 Event Modelling 14
2.3.1 The nature of events 14
2.3.2 Events in DOLCE 16
2.3.3 Event modelling in GIScience 18
2.3.4 Events and observations in the Semantic Sensor
Network ontology 20
2.3.5 Generic event models 21
2.4 Event Processing 24
2.4.1 ESP vs. CEP 25
2.4.2 EDA, SOA and the pub/sub paradigm 26
2.4.3 Semantic event processing for general purpose 27
2.4.4 Inferring events from sensor observations in GIScience
28
2.5 Eventing Standards 30
3 inferring events from in situ sensor observations 33
3.1 My Perspective of Events 33
3.2 The Event Abstraction Layer 34
3.3 Semantic Annotation of Event Patterns 35
3.4 The Event Abstraction Ontology 37
4 implementation of the event abstraction layer 43
4.1 Semantic Event Processing Architecture 43
4.2 Event Processing Service Prototype 45
4.2.1 Scheduling sensor data requests 45
4.2.2 Registering event patterns 46
4.2.3 Publication of event instances 47
5 real-time flood monitoring in the danube river 51
5.1 Overview 51
5.2 Two Views on the River Floods: a Governmental Body
and a Hydroelectric Power Plant 51
5.3 Inferring Events from Flood Monitoring Observations 53
5.3.1 A domain ontology for flood monitoring 54
5.3.2 Application ontologies and event patterns 55
5.4 Conclusion 59
6 evaluation 61
6.1 Testing the Prototype: Evaluation Experiments 61
6.1.1 Analysis of historical data: the Romanian floods
in 2006 61
6.1.2 Simulating floods in real-time 65
6.1.3 Conclusion 67
6.2 Comparing my Approach to Existing Solutions 67
7 conclusions and future work 71
7.1 Answer to Research Questions 71
7.2 Contribution 73
7.3 Future Work 74
8 appendix 77
8.1 EPS Software Engineering 77
8.1.1 EPS class diagram 77
8.2 SOS data request 78
8.3 Event Patterns 79
8.3.1 Event patterns for Romanian Waters 79
8.3.2 Event patterns for Hidroelectrica Romania 82
8.3.3 Event patterns for real-time experiment 86
8.4 Properties of an Event Abstraction Instance 87
8.5 SPARQL Queries 88
8.6 Real-time Data Simulation 92
bibliography 93