The small project establishes a conceptual framework for deploying less-precise mobile sensors en masse, with the ultimate goal of allowing scientists to analyze the measurements, detect hot spots and understand trends over time.
As cities, nonprofits and private companies begin setting up networks of sensors in an effort to create vast data sets, two Bay Area companies have opened up an experiment of one relatively untested approach: smaller, less-precise mobile sensors deployed en masse rather than a few high-quality stationary sensors.
Google announced last week that it has started putting air quality and weather sensors on a few of its Street View cars — the colorful cruisers that take ground-level photos for Google Maps — in an attempt to get information at the street level instead of at the city or regional levels. Air-quality districts, following federal standards, already have sensors set up around metropolitan areas across the country. But they’re stationary, and thus limited in what they’re able to show.
“Our network being fixed provides really good information about trends, things of that nature,” said Eric Stevenson, director of technical services for the Bay Area Air Quality Management District (BAAQMD). “What it doesn’t really provide is a spatial representation of the Bay Area at large.”
The hope is that the Google project will essentially add better understanding of the data BAAQMD is already gathering from 31 stations situated around the Bay Area.
“This will help us, the folks at Google, the folks at Aclima — it will help take the information that they’re gathering and apply modeling and functions to it to see what the meteorology of it is,” he said.
Pete Beckman, a scientist with the Argonne National Laboratory who’s working to set up the “Array of Things” in Chicago — a network of sensors that could be used to map air quality across the city, among other things — said the Google-Aclima project could help to shed light on an as-yet unexplored approach to building an Internet of Things. The concept is that rather than setting up a few high-quality sensors in fixed locations, smaller sensors could be deployed en masse throughout a city and travel about, capturing data as they go.
“This is an exploding technology area that is moving very rapidly, this notion that lightweight sensors can be deployed," he said. "We’ve tinkered with putting them on quad copters. So they’re going to go everywhere, and then the question … on the data science side is what can we learn and what kind of decisions can we make?”
Of course, every approach has its limitations, Beckman said. Though the Google cars will add mobility to the data the Bay Area Air Quality Management District receives, it might also be fleeting data — that is, a car might pass through a certain city block, and then not revisit that block for some time. After all, how often does Google need to take pictures of the same spots on the map?
“This isn’t to say that none of these challenges prevent us from using these, but they change how we do data collection,” Beckman said. “They affect what we can deduce.”
Another idea — one the Argonne National Laboratory is considering with its own technology — is to put sensors on city buses.
“Buses are much better … in that there is a very fixed schedule, and from a data science perspective, I am collecting data in a pattern and I am collecting it over and over and over again in that pattern,” he said.
The drawback to that approach is that buses tend to drive on busier roads, taking the sensors along routes that don’t necessarily have a lot of houses on them.
“(I) want data on where I spend 10 hours of the day, which is in my house,” he said.
But as the technology is deployed, Beckman said it’s possible that all those sensors — be they in fixed locations, attached to Google Maps cars or affixed to city buses — could all be pooled together to provide a bigger picture than each would provide on its own.
That’s part of the hope with the Google project, which will open its data up to the public, according to Google Earth Outreach Program Manager Karin Tuxen-Bettman.
“Our goal is to allow any scientist to analyze the measurements to create maps of air pollution, detect hotspots and understand trends over time,” Tuxen-Bettman wrote in an email to Government Technology. “We also share the goal to make those maps available to the public, for citizens, governments and organizations to access and take action on.”
That could allow for a wide range of possible applications, she wrote in a blog post last week.
“With street-level air pollution data, a parent of an asthmatic child could reduce exposure to air pollution that causes asthma attacks when they go to the park to play,” she wrote in the post. “Bike commuters and outdoor enthusiasts could find the healthiest route for their trips. Or a city planner could pinpoint areas of low air quality in her city and devise specific solutions to improve it.”
The project is starting out small with six sensors, and measuring air quality around the Bay Area, Los Angeles and the Central Valley. But Davida Herzl, chief executive officer of Aclima, sees it as a launch pad to expand the technology to other cities.
“We envision a day when this data is available as a layer on Google Maps,” she said.
But California is a good place to start. As the country’s most populous state, and one home to large urban populations in valleys where inversion factors can lead to smoggy air, Herzl said the data from the project will be available.
“California suffers from some of the worst pollution in the nation," she said. "In order to drive solutions, we need a much deeper and more granular understanding of air quality."
Eventually, she said she wants to see the deployment of sensors en masse to deliver granular, street-level information for large swaths of people.
“We can get very large spatial distribution and really map an entire city,” she said.