指導教授：周俊廷臺灣大學：電信工程學研究所林硯澤Lin, Ian-TseIan-TseLin2014-11-302018-07-052014-11-302018-07-052014http://ntur.lib.ntu.edu.tw//handle/246246/264358With advances in numerous technologies, our world is moving towards an ``always connected" paradigm. Internet of Things (IoT) is a tangible realization in this new paradigm and enables connectivity from “anytime, anywhere for anyone” toward “anytime, anywhere for anything”. With IoT, objects (e.g., toothbrush, door, window, etc.) in our daily lives are able to be connected with each other. These connected objects can sense the environment, communicate with each other and even transport various information to cloud servers, allowing service providers to make better decisions and take more appropriate actions. Home automation is a promising application in IoT that provide users a more convenient, comfortable and secure living environment through connected objects. There are generally two steps involved to set up a home automation system. The first step is to establish physical connections between objects in the house and a gateway that is connected to the Internet. The second step is to set up logical connections between objects so that the system can provide service accordingly. Setting up physical connections in current home automation systems usually requires professional installation, which is expensive and difficult to modify once they are set up. In this thesis, we propose an automatic solution for both steps. In our solution, clustering algorithms are used, base on the received signal strength measurements, to group objects that are in the same control zone. In home automation, the topology of control zones follows certain patterns, which are different from the random topology in other applications such as the intelligent transportation system (ITS). Based on this observation, hierarchical clustering algorithm is adopted. Simulation and experiment results show that more than 90 percent of objects can be automatically set up.ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi CHAPTER 1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . 1 1.1 Evaluation Towards the Internet of Things . . . . . . . . . . . . . . 1 1.2 Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.2.1 Home Automation . . . . . . . . . . . . . . . . . . . . . . . . 4 1.2.2 Home Automation System Architecture . . . . . . . . . . . . 5 1.2.3 Setting Up the Home Automation System . . . . . . . . . . . 6 1.3 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.4 Thesis Overview and Organization . . . . . . . . . . . . . . . . . . . 11 CHAPTER 2 RELATED WORK . . . . . . . . . . . . . . . . . . . . . 13 2.1 Indoor Localization in Wireless Sensor Network . . . . . . . . . . . . 13 2.1.1 Anchor-based Techniques . . . . . . . . . . . . . . . . . . . . 14 2.1.2 Anchor-free Techniques . . . . . . . . . . . . . . . . . . . . . 22 2.2 Summary of Indoor Localization Techniques . . . . . . . . . . . . . . 23 CHAPTER 3 GROUPING SCHEME . . . . . . . . . . . . . . . . . . 25 3.1 Home Automation Properties . . . . . . . . . . . . . . . . . . . . . . 25 3.2 Grouping Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.3 Clustering Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.3.1 Partitioning-based Clustering . . . . . . . . . . . . . . . . . . 28 3.3.2 Hierarchical-based Clustering . . . . . . . . . . . . . . . . . . 30 3.3.3 Density-based Clustering . . . . . . . . . . . . . . . . . . . . 32 3.3.4 Model-based Clustering . . . . . . . . . . . . . . . . . . . . . 32 3.3.5 Grid-based Clustering . . . . . . . . . . . . . . . . . . . . . . 34 3.3.6 Summary of Clustering Algorithms . . . . . . . . . . . . . . 34 3.4 Classi cation Algorithms . . . . . . . . . . . . . . . . . . . . . . . . 35 3.4.1 Neural Networks . . . . . . . . . . . . . . . . . . . . . . . . . 35 3.4.2 Support Vector Machines . . . . . . . . . . . . . . . . . . . . 36 CHAPTER 4 SIMULATION AND EXPERIMENT RESULTS . . 38 4.1 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 4.1.1 Performance Metric . . . . . . . . . . . . . . . . . . . . . . . 39 4.1.2 Multi-room Scenarios . . . . . . . . . . . . . . . . . . . . . . 39 4.1.3 Single Room Scenarios . . . . . . . . . . . . . . . . . . . . . 41 4.2 Experiment Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 4.2.1 Experiment Setting . . . . . . . . . . . . . . . . . . . . . . . 44 4.2.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 CHAPTER 5 DISCUSSIONS . . . . . . . . . . . . . . . . . . . . . . . 45 5.1 Correction through User Feedbacks . . . . . . . . . . . . . . . . . . 45 5.2 E ectiveness of Clustering Algorithms under Di erent Topologies . . 47 5.2.1 Spherical Topologies and Line-shaped Topologies . . . . . . . 47 CHAPTER 6 CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . 507353392 bytesapplication/pdf論文公開時間：2014/09/03論文使用權限：同意有償授權(權利金給回饋學校)自動配置家庭自動化分群智慧聯網thesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/264358/1/ntu-103-R01942112-1.pdf