Visible Light and Device-to-Device Communications: System Analysis and Implementation

Abstract

Radio-frequency based wireless communications have revolutionized our society. Thanks to important wireless communication technologies such as Wi-Fi, LTE, and so on, people can enjoy high data rate and pervasive connections while surfing the Internet. However, new challenges are rising in today’s wireless networks. Increasing capacity demands are requiring more bandwidth and various wireless radio technologies are exacerbating the spectrum problem. New technologies and paradigms are needed to meet these demands. In this thesis, I investigate two technologies towards this direction: Visible Light Communication (VLC) and Device-to-Device (D2D) communication systems. In VLC, transmitters modulate the light intensity of Light Emitting Diodes (LEDs) to send information. Receivers detect these light changes and decode them into data. Although more and more researchers are becoming interested in VLC, the lacking of an open-source platform for VLC research is preventing the fast evolution of VLC. To solve this problem, I design, implement, and evaluate the first open-source platform OpenVLC for embedded VLC research. OpenVLC employs cost-efficient and off-the-shelf optical components and electronics to provide a research platform. The software solutions are developed as a Linux driver and can be easily connected to the TCP/IP layers. This allows for the adoption of various Linux diagnostic tools to evaluate the VLC’s properties and performance. Based on OpenVLC, I propose a new MAC protocol that enables the intra-frame bidirectional transmissions in networks of visible LEDs. The method adopts only a single LED at each node for both transmission and reception. Through this proposed method, the system’s throughput can be greatly improved and the hidden-node problem can be strongly alleviated. Motivated by the envision of the Internet of lights, I study how to provide stable visible light links in VLC. I identify the limitations and trade-off of two different types of optical receivers (photodiode and LED), and design and implement a new optical data link layer that is resilient to dynamic environments. On the other hand, to meet the increasing demands, small cells are proposed and deployed in latest cellular networks. As a result, the number of users served by each cell is decreasing. As the opportunistic gain increases as a concave function of active users, in small cells and when dynamic traffic load are considered, the opportunistic gain will lost. To recoup the opportunistic gain, I propose a base-station transparent method based on D2D communication to dispatch traffic among devices. Dynamic programming is used to find the optimal dispatching policy. The results show this method can improve the average packet transfer delay greatly. To increase the opportunistic gain by a further step, I propose a base-station initiated policy to solve the same problem. An algorithm is therefore designed and implemented, and its performance shows that it can reduce the frame loss ratio significantly.