Date of Award

December 2016

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Electrical Engineering and Computer Science

Advisor(s)

Qinru Qiu

Subject Categories

Engineering

Abstract

In the past decade, mobile embedded systems, such as cell phones and tablets have infiltrated and dramatically transformed our life. The computation power, storage capacity and data communication speed of mobile devices have increases tremendously, and they have been used for more critical applications with intensive computation/communication. As a result, the battery lifetime becomes increasingly important and tends to be one of the key considerations for the consumers. Researches have been carried out to improve the efficiency of the lithium ion battery, which is a specific member in the more general Electrical Energy Storage (EES) family and is widely used in mobile systems, as well as the efficiency of other electrical energy storage systems such as supercapacitor, lead acid battery, and nickel–hydrogen battery etc. Previous studies show that hybrid electrical energy storage (HEES), which is a mixture of different EES technologies, gives the best performance. On the other hand, the Energy Harvesting (EH) technique has the potential to solve the problem once and for all by providing green and semi-permanent supply of energy to the embedded systems. However, the harvesting power must submit to the uncertainty of the environment and the variation of the weather. A stable and consistent power supply cannot always be guaranteed. The limited lifetime of the EES system and the unstableness of the EH system can be overcome by combining these two together to an energy harvesting embedded system and making them work cooperatively.

In an energy harvesting embedded systems, if the harvested power is sufficient for the workload, extra power can be stored in the EES element; if the harvested power is short, the energy stored in the EES bank can be used to support the load demand. How much energy can be stored in the charging phase and how long the EES bank lifetime will be are affected by many factors including the efficiency of the energy harvesting module, the input/output voltage of the DC-DC converters, the status of the EES elements, and the characteristics of the workload.

In this thesis, when the harvesting energy is abundant, our goal is to store as much surplus energy as possible in the EES bank under the variation of the harvesting power and the workload power. We investigate the impact of workload scheduling and Dynamic Voltage and Frequency Scaling (DVFS) of the embedded system on the energy efficiency of the EES bank in the charging phase. We propose a fast heuristic algorithm to minimize the energy overhead on the DC-DC converter while satisfying the timing constraints of the embedded workload and maximizing the energy stored in the HEES system. The proposed algorithm improves the efficiency of charging and discharging in an energy harvesting embedded system.

On the other hand, when the harvesting rate is low, workload power consumption is supplied by the EES bank. In this case, we try to minimize the energy consumption on the embedded system to extend its EES bank life. In this thesis, we consider the scenario when workload has uncertainties and is running on a heterogeneous multi-core system. The workload variation is represented by the selection of conditional branches which activate or deactivate a set of instructions belonging to a task. We employ both task scheduling and DVFS techniques for energy optimization. Our scheduling algorithm considers the statistical information of the workload to minimize the mean power consumption of the application while satisfying a hard deadline constraint. The proposed DVFS algorithm has pseudo linear complexity and achieves comparable energy reduction as the solutions found by mathematical programming. Due to its capability of slack reclaiming, our DVFS technique is less sensitive to small change in hardware or workload and works more robustly than other techniques without slack reclaiming.

Access

Open Access

Download

Included in

Engineering Commons

Share

COinS