Pollution is invisible and knowing how much is around you is not always easy. But a new system called Citisense, which consists of a mobile air quality sensor and smartphone app, could one day give people real-time information about the air around them.
"Asthmatics, who number in the millions, would find this valuable to their immediate health," said William Griswold, a computer science professor at UC San Diego, who lead the group that developed the system. "What we found is that people are very interested in their personal exposure, even if they are not asthmatic."
The system, which is still in the research stages, has a mobile sensor that a person wears while walking or biking around a city. The sensor detects the levels of pollutants in the air and sends the information to a server that uses machine learning to analyze the information for the app. Users with the app can see maps that display levels of pollutants, estimates of a user's exposure to those pollutants as well as a color-coded scale for air quality that uses EPA standards, i.e. green for good and purple for bad.
The sensors were tested for four weeks by 30 people all over San Diego, most of them faculty at the university. According to the press release, one tester found that she was exposed the most to pollutants while she rode her bike to work.
Griswold said in the release that, “The people who are doing the most to reduce emissions, by biking or taking the bus, were the people who experienced the highest levels of exposure to pollutants.” The field tests also found that pollution levels varied throughout the day, depending on variables like traffic.
For the most part, the sensors are mobile and proximity to them is necessary for the app to receive data. However, Griswold said in an email to Discovery News that if enough sensors were put out into an area, personal sensors wouldn't be necessary to receive feedback on the pollutants nearby. "With the machine-learning component in the backend," he said, "it will be possible to get an estimate of your exposure from the machine learning estimates, even if you don't have a sensor."
Toward the end of the testing phase, a few fixed sensors were tested, but Griswold said that they didn't affect the user experience enough to continue.
One of the hurdles facing the project now is battery life. The data exchanges between the sensors and mobile devices takes up a lot of power. When testing, users had to carry around two chargers, one for the sensor and one for the smartphone. Currently, the team is experimenting with replacing constant updates by spacing out times when data is transferred to every 15 minutes to save battery life, or making it a transfer that occurs on demand.
Griswold said in an email that sensors like this will be start appearing on mobile phones in about a decade or so.
Credit: Jacobs School of Engineering