Project topics and materials
Get Complete Project Material File(s) Now! »
Research Hypotheses
In this section, we provide research hypotheses that we address in this thesis:
• Hypothesis 1: The semantic engine is not too resource consuming. This is a first step to later embed the semantic engine in constrained devices. We are expecting that it takes few minutes to run the semantic engine. This is done by measuring software performances which is explained in Section 3.7.
• Hypothesis 2: The semantic engine is generic enough to support various kind of IoT measurement. We are interested in checking that the semantic engine is adapted to heterogeneous IoT domains such as healthcare, smart home, smart cities, weather forecasting, smart car, etc. and handles the data generated by simple IoT devices.
This is done by running M3, more precisely, S-LOR on datasets with different kind of measurements which is explained in Section 3.7.
• Hypothesis 3: The semantic engine enables building cross-domain IoT applications. This is done by running M3, more precisely, S-LOR on datasets with different kind of measurements and by loading cross-domain templates which is explained in Section 3.7.
• Hypothesis 4: Users are interested to integrate semantic web technologies to Internet of Things. This is done by looking at the visitor map and Google Analytics on our proof of concept available on the Web which is explained in Section 3.7. We keep it as an hypothesis even if people do not consider it as a research hypothesis. For us, it is important to evaluate whether it encourages persevering in this research topic.
• Hypothesis 5: The dataset of M3 rules is reliable to interpret IoT data. This aspect is important to ensure the reliability of the results provided by the reasoning engine. This is done by looking at completeness and correctness of the rules which is explained in Section 4.8.
• Hypothesis 6: A dataset of ontology-based projects relevant for IoT can be exploited outside of the M3 framework. It shows that the catalogue is relevant and exploited by people. This is done through a user form which is explained in Section 4.8.
• Hypothesis 7: The knowledge base built to interpret IoT data encourages the interoperability of data and domains. This can be done by following semantic web best practices which is explained in Section 4.8. It is important to encourage the reuse of our knowledge base.
• Hypothesis 8: The security knowledge base is built using the same methodology that for the M3 interoperable domain knowledge. It shows that M3 is generic enough for other domains such as security which is explained in Section 5.5.
• Hypothesis 9: A security knowledge base help non-experts in security choose security mechanisms fitting their needs to secure IoT applications. This is done through a user form which is explained in Section 5.5.
Understanding SemanticWeb of Things Related Research Fields
In this section, we explain the main challenges of heterogenous research fields having overlapping goals: Ubiquitous Computing (UbiComp), Pervasive Computing, Ambient Intelligence (AmI), Context-Awareness, Ambient Assisted Living (AAL), Smart Homes, Semantic Sensor Networks (SSN), Machine-to-Machine (M2M), Internet of Things (IoT), Web of Things (WoT), Semantic Web of Things (SWoT), Smart Cities and Physical-Cyber-Social Computing (PCS). The evolution of Ubiquitous Computing has given place to new terms like ’Pervasive Computing’, ’Context-aware Computing’, ’Mobile Computing’, ’Wearable computing’ and now ’Internet of Things’.
Ubiquitous Computing (UbiComp)
Ubiquitous computing is a research field aiming at integrating computers into objects. In 1993, Weiser introduces the notion of ubiquitous computing [Weiser, 1993]. In 2003, Chen et al. integrate semantic web technologies to pervasive computing [Chen, 2004]. They design the Standard Ontology for Ubiquitous and Pervasive Ap- plications (SOUPA) ontology to describe user profiles, beliefs, desires, etc. [Chen et al., 2003] [Chen et al., 2004] [Chen et al., 2005b] The SOUPA ontology is integrated in the Context Broker Architecture (COBRA) architecture to build smart meeting rooms [Chen, 2003] [Chen et al., 2005a] [Chen, 2004]. COBRA is a centralized architecture for context-aware systems in smart environment. Then, the authors developed EasyMeeting, an intelligent meeting room based on the COBRA architecture. They define a policy language for users to control the sharing of their information and two ontologies SOUPA and COBRA-ONT.
The ontology COBRA-ONT is for modeling context in an intelligent meeting room: places, agents, agents location and agent’s activity.
The CONON ontology [Wang et al., 2004] [Gu et al., 2004] has been designed and integrated to the Service-Oriented Context-AwareMiddleware (SOCAM) [Zhang et al., 2005] [Wang et al., 2002] architecture. Neither prototype nor ontologies are available online. The SOCAM architecture is an OSGibased architecture that converts various physical spaces where contexts are acquired into a semantic environment where context-aware applications can share and access them easily. The context ontology CONON [Wang et al., 2004] [Gu et al., 2004] defines several concepts like computational entities (services, applications, devices), location, person, activity and indoor space (building, room, corridor and entry). In 2006, Hilera et al. survey the ontologies in ubiquitous computing: SOUPA, CONON, FIPA and GUMO [Hilera and Ruiz, 2006].
Jeong et al. [Jeong et al., 2006] propose in the Ubiquitous Computing Architecture (UTOPIA) project, an ontology to describe: (1) environment (e.g., humidity, temperature), (2) objects with person, computing device (computer, display, printer), environmental devices (e.g., air conditioner, curtain, door, window), (3) space represents building or room, (4) activities, (5) preferences, and (6) people. Within the UTOPIA project, some services such as U-Restaurant Service (some information are provided about religious beliefs, if the meal is vegetarian), U-Museum service, and U-Theme Park service have been developed.
Table of contents :
Abstract
French Abstract
Acknowledgments
I Introduction & State of the Art
Chapter 1: Introduction
1.1 Motivation
1.2 Problem
1.3 Our Approach
1.4 Assumptions
1.5 Research Hypotheses
1.6 Contributions
1.7 Organization of Thesis
Chapter 2: State of the Art: SemanticWeb of Things (SWoT) & Related Research fields
2.1 Understanding Semantic Web of Things Related Research Fields
2.1.1 Ubiquitous Computing (UbiComp)
2.1.2 Pervasive Computing
2.1.3 Ambient Intelligence (AmI)
2.1.4 Context-Awareness
2.1.5 Ambient Assisted Living (AAL)
2.1.6 Smart Homes
2.1.7 Semantic Sensor Networks (SSN)
2.1.8 Machine-to-Machine (M2M)
2.1.9 Internet of Things (IoT)
2.1.10 Web of Things (WoT)
2.1.11 Semantic Web of Things (SWoT)
2.1.12 Smart Cities
2.1.13 Physical-Cyber-Social Computing (PCS)
2.1.14 Discussions
2.2 Identifying Main SWoT Challenges
2.3 Identifying Existing Tools Limitations for Each Challenge
2.3.1 Interoperable IoT Data
2.3.2 Interpreting IoT data
2.3.3 Inter-Domain Interoperability
2.3.4 Designing Interoperable IoT applications
2.3.5 Sensor Plug & Play
2.3.6 Semantics Applied to Constrained Devices
2.3.7 Securing IoT
2.4 Concluding Remarks: Limitations of these Works
2.4.1 Describing interoperable IoT data
2.4.2 Interpreting IoT Data
2.4.3 Inter-domain Interoperability
2.4.4 Securing IoT
2.4.5 Summary
Chapter 3: The Machine-to-Machine Measurement (M3) Framework
3.1 Assisting Developers in Designing SWoT applications
3.2 M3 Architectural Overview
3.3 SWoT generator
3.4 Designing Interoperable Semantic Web of Things Applications with M3 .
3.4.1 Generating M3 templates
3.4.2 Semantically annotate IoT data
3.4.3 Interpreting IoT data
3.4.4 Making use of M3 templates for IoT EU projects
3.5 Integrating M3 in a Semantic-Based M2M Architecture
3.6 Implementation
3.6.1 Scenario 1: Suggesting safety devices according to the weather
3.6.2 Scenario 2: Suggesting activities or clothes according to the weather 84
3.6.3 Scenario 3: Suggesting home remedies according to health measurements
3.7 Evaluation
3.7.1 Evaluating software performances
3.7.2 Evaluating the semantic engine with different IoT datasets
3.7.3 Evaluating with end users
3.7.4 Discussions
3.8 Concluding Remarks
Chapter 4: Sensor-Based Linked Open Rules (S-LOR)
4.1 Assisting IoT developers in Interpreting IoT Data
4.2 M3 Nomenclature & Ontology
4.3 Linked Open Vocabularies for Internet of Things (LOV4IoT)
4.3.1 LOV4IoT, an extension of the LOV catalogue
4.3.2 LOV4IoT table
4.3.3 LOV4IoT RDF dataset
4.3.4 Extracting a dictionary to describe sensor measurements
4.3.5 Extracting rules to interpret sensor measurements
4.3.6 Extracting domains
4.3.7 Lessons learned
4.4 Interoperable M3 Cross-Domain Knowledge
4.4.1 Designing an interoperable M3 domain knowledge
4.4.2 Combining domain knowledge expertise through M3 rules
4.5 The semantic engine S-LOR integrated in the M3 Approach
4.6 S-LOR: A ’Share and Reuse’ Based Reasoning Approach
4.7 Implementation
4.8 Evaluation
4.8.1 Evaluating M3 rules with completeness and correctness
4.8.2 Evaluating LOV4IoT
4.8.3 Evaluating M3 domain knowledge with semantic web methodologies
4.8.4 Discussions
4.9 Concluding Remarks
Chapter 5: Security Toolbox: Attacks & Countermeasures (STAC)
5.1 Assisting Developers in Securing IoT Applications
5.2 STAC generator
5.3 Interoperable STAC Cross-Domain Knowledge
5.3.1 Reusing Security Knowledge with LOV4IoT
5.3.2 STAC ontology
5.3.3 STAC dataset
5.3.4 Updating STAC
5.4 Implementation
5.5 Evaluation
5.5.1 Evaluating STAC domain knowledge with semantic web methodologies
5.5.2 Evaluating STAC with end users
5.5.3 Discussions
5.6 The novelty of the STAC knowledge base
5.7 Concluding Remarks
Chapter 6: M3 Framework at Work
6.1 Using and Contributing to M3
6.2 Developing Mobile SWoT Applications with M3
6.2.1 Application Provisioning Phase
6.2.2 Design Application Phase
6.3 Integrating M3 in Smart Cars
6.4 End-User Centric Approach: M3 Embedded in Smart Fridges
6.5 End-User Centric Approach: M3 Embedded in Smart Luggage
6.6 Designing Secure IoT Applications with STAC
6.7 Concluding Remarks
Chapter 7: Conclusion and Future Directions
7.1 Conclusion
7.2 Short Term Challenges, Future Directions and Discussions Regarding M3 .
7.2.1 Synergizing efforts with standardization
7.2.2 Extracting the domain knowledge
7.2.3 Enhancing Sensor-based Linked Open Rules
7.2.4 Polishing the M3 framework
7.3 Long Term Challenges
7.4 Social impacts
Bibliography
