Increasingly, sensing, computation, and communication are an integral part of the physical world. As we interact with our surroundings, we encounter many small computers, often invisible, networked, and embedded into everyday objects. Networked mobile devices have become commonplace enough that it is hard to imagine a world without them. This new generation of mobile devices with access to wireless networking and inexpensive sensors has enabled a variety of context-aware applications that provide services based on the personal context data of individual users. We are now moving beyond the initial challenges in ubiquitous computing of creating a mobile infrastructure, leading to a more urgent need to consider higher level challenges, including managing energy constraints. The increasing amounts of data that can be captured by sensors, analyzed and used by applications, stored for various reasons, and distributed throughout networks highlight both expanding possibilities for new services and the difficulty of providing long lasting battery life for mobile devices.
In this talk I outline the emerging field of ubiquitous computing, highlighting rich possibilities for new systems and applications, and the major barriers to their realization. I describe my research on data fidelity in context-aware systems, leading to a framework for using fidelity characteristics to deliberately design for energy management in such systems.