Date of Completion


Embargo Period



Smart Grid, Electrical Vehicle Networks, V2G, Internet of Things

Major Advisor

Bing Wang

Co-Major Advisor

Song Han

Associate Advisor

Mohammad Khan

Associate Advisor

Shalabh Gupta

Associate Advisor

Donald Sheehy

Field of Study

Computer Science and Engineering


Doctor of Philosophy

Open Access

Open Access


Energy sustainability is a pressing issue facing the modern society. The twin pillars of sustainable energy are renewable energy and energy efficiency. In this dissertation, we propose novel architectures and approaches to improve energy sustainability in two application domains: transportation and the Internet of Things. Transporta- tion is one of the major sources of energy consumption and environmental pollution. Plug-in hybrid electric vehicles (PHEVs) present many opportunities in improving energy efficiency and reducing greenhouse gas emissions. In addition, with batteries and built-in mobility, PHEVs can form a mobile and distributed energy network, where energy can be conveniently transported from place to place. In the first part of this dissertation, we investigate how to optimally distribute renewable energy in a distributed PHEV energy network under two system architectures. The first archi- tecture assumes that each charge station is equipped with energy storage to serve as an energy exchange point. Some PHEVs can be charged by renewable energy sources and discharge energy at a charge station. Other PHEVs passing by the charge sta- tion can withdraw energy from the charge station, and therefore indirectly use the energy from the renewable energy sources. The second architecture assumes that the charge stations do not have energy storage. Instead, they are connected using under- ground cables to a central energy storage (CES), which has a limited capacity and is charged by renewable energy sources. PHEVs can withdraw/deposit energy from/to the CES (and thus indirectly use renewable energy) through the charge stations. We formulate and solve the optimal renewable energy transfer problem under each of the two system architectures. The two optimization problems share the same objective function to maximize the total amount of renewable energy used by the PHEVs but are subject to different constraints derived from the system architectures. Simulation results using the data set from the Manhattan city bus system demonstrate that our approaches significantly outperform baseline schemes and provide effective ways to share renewable energy in PHEV energy network and thus improve energy sustain- ability. We further study the energy sustainability problem in a broader application do- main - Internet of Things (IoT). IoT is the network of things that enables devices to exchange data with manufacturers, operators or other devices. It is expected that there will be nearly 26 billion devices on the IoT by 2020. Energy efficiency is a critical issue in the IoT. In the second part of the dissertation, we investigate energy- efficient packet transmission in the IoT. Specifically, we consider a mobile network with group-based encountering. The optimization goal is to minimize the delay for transmitting a set of packets from a source to a destination while limiting the en- ergy consumption. The challenge lies in how to schedule packet transmission among a group of nodes that meet each other so that information carried by the different nodes can be exchanged effectively. We first assume that node encountering is known beforehand, and develop a max-flow based algorithm that obtains the optimal solu- tion. While the assumption is clearly unrealistic, the optimal solution is useful to quantify the effectiveness of different heuristic algorithms. Specifically, we propose two network coding based heuristic algorithms. One algorithm uses full signaling where nodes exchange their coefficient matrix with each other while the other incurs much less signaling overhead in that nodes only exchange rank information when meeting each other. Both algorithms use a token-based technique to limit the total number of transmissions, and only incur signaling at the beginning of a group meet- ing. Simulation results demonstrate that both algorithms achieve delays close to the minimum latency for moderate number of tokens. They present different tradeoffs in the number of transmissions and the signaling overhead.