Project topics and materials
Get Complete Project Material File(s) Now! »
Constitutive Hypothesis
In the research context outlined above, we pointed out that emerging network technologies contribute in the heterogeneity of building systems by only barely focusing on interoperability at the application layer. Furthermore, only few existing approaches allow to interconnect applications, especially when it comes to address the requirements of sensor networks composed of constrained devices. Based on the preceding observations, this work is grounded on the following hypotheses:
• The Internet of Things and its related technologies are emerging.
• A standardisation on top of the Internet of Things is inevitable to ensure interoperability.
• Internet of Things devices will penetrate the building automation market with novel sensors actuators.
• The paradigm of the Web of Things and Web technologies will increasingly be adopted in sensor networks.
These hypotheses will be further detailed and confronted with a literature survey presented in Chapter 2 and Chapter 3. As it will be demonstrated, we have good reasons to believe that theses assumptions will be true in a near future. Considering these, our constitutive hypothesis and main research question in this thesis is the following:
Can building automation systems composed of novel IoT nodes and legacy devices be homogenised around emerging Web technologies and especially the paradigm of the Web of Things?
More research questions bound to this main question are detailed in the next section.
Research Questions and Methodologies
In addition to the aforementioned assumptions and the constitutive hypothesis, this thesis is driven by the following research questions:
• Are applicative Web protocols compatible with building automation system interaction models?
• Are RESTful APIs descriptive enough to allow automatic service consumption in the context of buildings?
• Can applications be built as in-network clouds and distributed on smart things?
• Are data-driven building management algorithms compatible with a Web approach?
• Do applicative Web protocols used by notification mechanisms differentiate between themselves in terms of energy efficiency and traffic load?
• Can legacy building automation systems such as KNX be seamlessly interfaced with Webbased sensor networks?
These research questions are addressed throughout this thesis by surveying related works and discussing the current shortcomings and limitations of existing solutions. We then propose to overcome these problems not only from a conceptual point of view, but through the development of proof of concepts to validate our proposals. We then test our experimental implementations as far as possible in scenarios reflecting real use cases. Finally, and whenever possible, we confront our results to existing state-of-the-art solutions in order to position our contributions.
Thesis Outline
The remainder of this thesis is structured as follows: Chapter 2 presents a high-level survey of related works with emphasis on the emergence of new sensor network technologies and building automation systems.
In Chapter 3 we provide insights on current Web-related paradigms and technologies aiming to push the Web down to daily objects and sensor networks. We also discuss their potential use in smart building applications.
In Chapter 4 we present our Web of Buildings architecture. We expose the required building blocks to successfully turn building automation systems into an homogeneous Web ecosystem. Our proposal includes four blocks addressing different levels of functionalities required by building automation systems. The Field Level provides communication capabilities with sensor and actuator devices. The Automation Level challenges the creating and deployment of automation rules. The Intelligent Adaptation Level outlines how the machine learning training phase can be integrated into the Web. Finally the Management Level contributes to the overall working with a set of general supporting services for a building network.
In Chapter 5 we challenge our proposition of the Web of Buildings architecture from two angles. First, we apply our Web of Buildings layered architecture to integrate the KNX and EnOcean legacy networks into the Web. For this, we present two gateway implementations that were successfully deployed in real environments. Then, we address the energy consumption of smart things and propose a set of mechanisms to limit their footprint. We first apply an intelligent module on data producers that automatically switches to the most economic application protocol. We then suggest to adapt CoAP in order to make multicast notifications possible. Finally, in Chapter 6 we provide a summary of our contributions. By taking a step back, we also discuss some open challenges and interesting future directions the Web of Buildings will have to address.
Table of contents :
1 Introduction
1.1 Problem Statement
1.2 Research Context
1.3 Constitutive Hypothesis
1.4 Research Questions and Methodologies
1.5 Contributions
1.6 Thesis Outline
2 Background and Motivations
2.1 Emergence of New Technologies in Building Automation Systems
2.1.1 Emerging Wireless Sensor Networks
2.1.2 Heterogeneity in Smart Buildings
2.1.3 Data-Driven Building Management
2.2 Building Automation Systems (BASs)
2.2.1 KNX
2.2.2 EnOcean
2.2.3 BACnet
2.2.4 Internet of Things
2.2.5 Discussion
2.3 Distributed Systems Architectures
2.3.1 Object-Oriented Architectures (OOA)
2.3.2 Service-Oriented Architectures (SOA)
2.3.3 Resource-Oriented Architectures (ROA)
2.3.4 Discussion
2.4 Middleware for Building Automation Systems
2.5 Requirements for Open and Scalable Smart Buildings
2.6 Discussion and Summary
3 Towards Things-Oriented Building Networks
3.1 Relying on the Web
3.2 The Web of Things
3.2.1 Representational State Transfer (REST)
3.2.2 Addressing Resources
3.2.3 Exchanging Information
3.2.4 Interacting with Resources
3.3 Pushing Data from Things
3.3.1 HTTP Long Polling
3.3.2 HTTP Streaming
3.3.3 HTTP Callbacks
3.3.4 WebSockets
3.3.5 MQTT
3.3.6 Discussion
3.4 The Constrained Application Protocol (CoAP)
3.4.1 Observer Pattern
3.4.2 Group Communication
3.5 Discussion and Summary
4 The Web of Buildings
4.1 Management Level
4.1.1 Hybrid REST Library
4.1.2 A Building-Oriented Naming System
4.1.3 Service Descriptions
4.1.4 A Semantic-Oriented Query Engine
4.1.5 Related Work
4.2 Field Level
4.2.1 Web APIs for Sensing and Actuation
4.2.2 Towards Homogeneous Building Automation Systems
4.2.3 A Cloud-Like Storage for Historical Measurements
4.2.4 Related Work
4.3 Automation Level
4.3.1 A Machine Learning Approach
4.3.2 MaLeX: A Web Interchange Format
4.3.3 Distributing the Machine Learning Runtime
4.3.4 A Mashup-Based Room Controller
4.3.5 Related Work
4.4 Intelligent Adaptation Level
4.4.1 Training
4.4.2 Adaptation
4.4.3 Related Work
4.5 Summary
4.5.1 Related Work
4.5.2 Discussion
4.5.3 Contributions
5 Challenging the Web of Buildings Architecture
5.1 Integrating Legacy Building Automation Systems in the Web of Buildings
5.1.1 KNX
5.1.2 EnOcean
5.1.3 Related Work
5.1.4 Discussion
5.2 Making the Web of Buildings More Energy Efficient
5.2.1 Selecting the Appropriate Protocol
5.2.2 Towards Multicast Notifications
5.2.3 Related Work
5.2.4 Discussion
5.3 Summary
6 Conclusions and Outlook
6.1 Contributions
6.2 Discussion and Future Challenges
A Energy Consumption Measures
A.1 Raw Energy Consumption of the Protocols
A.2 Energy Consumption Measures with the Selection Intelligence Module
References
