To do participatory sensing means to be part of a group of people who keep track of and make public where, when, and why they are in a particular spot, usually via a mobile global positioning system (GPS). Until recently, most of the buzz around sensing one’s location based on cell phones concentrated on advertising--if you’re within a few feet of a Saks Fifth Avenue, your phone tells you about the store’s sales. In the personal environmental impact report (PEIR) system, however, the information goes in the opposite direction--it determines what your carbon impact is, depending on where you are and what you’re doing.

Your phone keeps track of four dimensions: your carbon impact (are you walking or driving?), whether you’re emitting airborne particulates near hospitals and schools, your exposure to smog (for example, on the freeway), and how much time you’re spending in or near fast-food restaurants (according to the authors, proximity to fast food has been found to correlate with increased obesity).

By solving these four tracking problems, Mun et al. also solve problems for other projects and developers. For example, how do you know if a PEIR member is on the freeway or walking down the block? You can’t look at feet per hour to differentiate between walking and driving--if you’re stuck in traffic, you may be crawling along at a walking pace. Well, how about identifying the street? Unfortunately, there are enough GPS errors that finding the closest road usually isn’t enough. Also, even though a surface road might appear to be your location, you may actually be on the highway that passes over or next to the surface road. So, the PEIR developers created an “intersection based map-matching.” Instead of accepting the first street identified, the algorithm watches for intersections. If, for example, the user first passes Sawtelle Blvd. and National Blvd., and then passes Sepulveda Blvd. and National Blvd., then he or she is probably on National, not Sawtelle or Sepulveda.

It is clear to me that the ten authors really thought this system through. They developed a clever algorithm for hiding users’ routes. No doubt, partway into the study, some of their users pointed out that they might not want some of their routes to be public (there is a link to Facebook). So, the authors developed a way to substitute innocuous routes for the ones the users want to keep private, without losing the metrics associated with the real route. Very clever!

Although the system was designed to track carbon footprints, the ideas that Mun et al. propose, tested in the field and revised, can be generalized for other uses. The paper is fun to read also, because you, as a reader, can trace their routes through the problems they find and solve, in the same way they trace people through the physical world.