Work

User Experience Aware System Optimizations for Mobile Systems

Public

Downloadable Content

Download PDF

Over the last few years, understanding user experience within mobile systems has become a popular phenomenon as a means to manage hardware resources. Across the many issues being studied in this area, I focus on how to utilize user satisfaction in this dissertation. Notably, I examine user experience by incorporating real end user satisfaction in mobile system designs and optimizations by conducting a variety of user studies. While the importance of end user is not debated, the individual end user preferences are still widely ignored as the modern system designs hinge on the average user by developing a “one-size fits-all” approach. In my work, I begin with showing that there exists a significant variation in user satisfactions across different users - even under the same workloads and/or conditions. Thus, it is certainly vital to not only focus on the raw performance of the systems, but also to concentrate on the needs and desires of the end user. I then proceed with demonstrating the tools and methods needed in order to build highly accurate prediction models for user satisfaction. For the prediction models, I use system metrics and/or built-in sensor data that is available in modern mobile devices. I also validate that the proposed models can easily be adopted in any modern mobile systems at system-level with negligible power consumption. Lastly, I end with conducting user studies in order to demonstrate how the prediction of user satisfaction can be utilized in hardware component managements for the mobile devices. Within these user experiments, I study specifically the two most power hungry components in order to maximize system-level energy savings: CPU and screen brightness. All in all, this dissertation points toward the importance of incorporating the end user’s satisfaction with the hardware component management to the mobile systems research. Various user studies that have been conducted as a part of this work illustrates that if we can succeed in placing the end user into the design and optimization process, we can achieve significant improvements with regards to the energy efficiency of current mobile systems while maintaining - or even improving - individual levels of user satisfaction concurrently.

Creator
DOI
Subject
Language
Alternate Identifier
Date created
Resource type
Rights statement

Relationships

Items