Intelligent Machines

Robot Lights Up at Pollution

A hacked Roomba detects the specific spots where indoor air is bad.

Jul 18, 2011
Robots may someday be able to tell you if the air in your house or office is polluted. Researchers at the Public Laboratory for Open Technology and Science (PLOTS), in conjunction with the Rhode Island School of Design, attached a battery pack (in black, above), circuit boards, and a chemical sensor (next to the light) to a floor-cleaning Roomba robot. The robot’s light would change slowly from yellow to blue if it detected an increasing amount of airborne alcohol. The group plans to add sensors for other volatile organic chemicals (VOCs), such as formaldehyde, which is present in a lot of wood furniture and can cause dizziness and nausea and may be linked to asthma. “There are not any good tools for the average consumer to use to monitor indoor air pollutants,” says Gennet Paauwe, spokeswoman for the California Air Resources Board.
The team decided to create a “do-it-yourself” hack that would let schools, community centers, and motivated consumers test and track their environments’ air quality with cheap, off-the-shelf parts, pictured above and captioned by the researchers. Team members Jae-Ok Lee and Byeongwon Ha attached the VOC sensor MQ 135 and an LED that can glow red, green, and blue to an Arduino programming board, and rigged it to a $50 Roomba they bought secondhand. “Indoor air pollution is a huge problem,” says Sara Wylie, an MIT researcher and the director of toxins and health research at PLOTS. “We wanted to figure out ways people could start to recognize what kinds of pollutants are in the house.”
The researchers sprinkled rum around a room, mounted a camera overhead, turned out the lights, and let the robot loose. The long-exposure photo above shows the prototype moving as its floor-cleaning algorithm dictates. The light glows yellow when it doesn’t detect the evaporating alcohol, and blue when it does detect it. The image drives home one of the advantages of using a robot to sense air quality: it moves around an entire room, unlike stationary sensors. The PLOTS team next plans to develop a handheld sensor wand as a cheaper means for pollution detection. “The Roomba was our first attempt to think through how we could generate a compelling, single picture of a whole room rather than individual data points,” Wylie says.
The team superimposed the raw light-data image onto an illuminated photo of the room to more easily locate the spots containing alcohol. They plan to encourage users who create the DIY robot sensor to send their photos to an online gallery, and the team will help them compare their images to other images of air quality. Wylie would love to see many people sending in photos so her team can figure out what kinds of furniture or other indoor fixtures give off toxins. Last month, the lab was awarded a Knight News Challenge grant to support its plans to add more sensitive sensors and to disseminate the technology in schools and other community settings.