Friday, April 24, 2009

The Best of CS263

Well, we're done discussing papers in CS263 this term. We read and talked about 41 papers over the course of the semester, and by and large they were really great.

We decided to run the course in a mock "program committee" style, in which our goal was to pick the top 10 or so papers over the course of the semester. This is great practice for students but also provided a focal point for the discussion: Does a given paper belong in our "program"?

After much deliberation, we have selected the following eight papers as The Best Papers from CS263. This could be seen as a kind of "best sensor network systems papers" award. (Of course, there are many papers we did not read at all, so caveat emptor!)

System Architecture Directions for Networked Sensors, J. Hill et al., ASPLOS'00

This is the classic paper on TinyOS, which lays out many of the challenges in the sensor network domain and focuses on the new OS and communication model requirements.
Radio Interferometric Localization, M. Maroti et al., Sensys'05

There's a lot to like about this paper, which applies RF interferometry to low-power wireless sensor nodes; a lot of technical depth.

The Design of an Acquisitional Query Processor for Sensor Networks, S. Madden et al., SIGMOD'03

TinyDB has been hugely influential in terms of new approaches for querying and aggregating data in sensor nets.

Sensor Network-Based Countersniper System, G. Simon et al., Sensys'04 (BEST PAPER AWARD!)

This system uses acoustic sensors to localize the source of gunshots in an urban setting. It brings together event detection, time synchronization, localization, and a nice sensor fusion algorithm to develop a complete system, which is deployed and evaluated in a real field setting. This paper was universally voted as the students' favorite so we've given it our "best paper award."

Distributed Image Search in Sensor Networks, T. Yan et al., Sensys'08

This paper proposed to use a network of wireless cameras to store and index images which can later be queried by similarity. The authors do a very good job at motivating their design and diving into the details of image processing, flash storage, and so forth.

Dozer: Ultra-Low Power Data Gathering in Sensor Networks, N. Burri et al., IPSN 2007

This paper does a great job at a complete cross-layer design spanning the MAC, link, routing, and application layers to achieve very low power and high reliability for data collection. In some sense this is the first paper I'd give someone working on communication in sensor nets, since it nails down how to do it right.

An Analysis of a Large Scale Habitat Monitoring Application, R. Szewczyk et al., Sensys'04

This is another "classic" paper on the sensor net deployment at Great Duck Island. It's a great case study of the challenges faced by a sensor network in the real world.

Synopsis Diffusion for Robust Aggregation in Sensor Networks, S. Nath et. al, Sensys'04

This was one of the only "algorithms" papers we read, and it does a great job at developing an aggregation technique that is robust to duplicate and lost packets. A superb example of plucking a somewhat obscure concept from the literature and demonstrating its value in a real system.

It's been a great term and I gained a lot by having these heated discussions with the students, who were approaching many of these papers for the first time.

Next week we're doing in-class presentations of the final student projects -- I'll be blogging summaries of those as well!