• Department Home
• Undergraduate Students
• Graduate Students
• Faculty & Staff
• Research

Computer Science Colloquia

Embedded networked sensors, those that coordinate amongst themselves to achieve a sensing task, promise to revolutionize the way we live, work and interact with our physical environment. To build large scale, ad hoc deployable, long-lived, densely-deployed sensor networks that operate unattended in unpredictable environments from limited energy sensor nodes, we must design them to self-configure, or autonomously adapt to environmental and application dynamics, or the availability of other nodes.

I will explore this notion in the context of two building blocks fundamental to sensor coordination: (i) node localization, or establishing spatial relationships among sensor nodes, and (ii) query support, or extracting global information from a dynamic sensor network via queries.

First, I will describe the design and implementation of a scalable, decentralized node localization scheme based on radio-proximity to known beacons. I will consider the effect of beacon density on localization quality, and show how the localization system can self-configure by rotating functionality amongst redundant beacons to improve system lifetime without degrading localization quality. I will include performance results from simulations, as well as experimental results with the Berkeley motes (nodes with sharply limited resources).

Next, I will consider the effect of the sensor query type on the amount of communication needed to process it. I will make the case for and describe an information-driven communication architecture that combines information push, or low-rate anisotropic data percolation with information-driven dynamic sensor collaboration and predictive query routing using an application-driven dynamics model, to support complex queries efficiently and with low latency.

I will conclude the talk with some general lessons and future directions.