The particularly hot summer this year is putting a lot of emphasis on air quality and pollution. It happens to coincide with a planned pilot project that began this summer in Chicago to install advanced air quality sensors on 100 of the city’s bus shelters.

French bus stop advertising and street furniture company JCDecaux and Microsoft Research’s Urban Innovation Group are partnering for this project with the team from Chicago’s Array of Things — a five-year intelligent urban sensor project whose initial funding ended in fall 2020.

“The plan is for the city to include this data with a number of other air quality sensor sources that they use to monitor air quality across the city,” said Charlie Catlett, senior research scientist at the University of Illinois Discovery Partners Institute and leader of the Array of Things team. “I don’t know when they will publish the API to pull the data, but my team has access to it. They intend to have it pushed into the City of Chicago data portal.”

AoT, take two

The new bus stop sensor project is one of two follow-on projects that build on AoT’s past successes. These devices are more advanced, efficient, and potentially longer lasting compared to those installed over the past five years.

AoT is a public-private partnership that formed to ingest and analyze environmental data collected through programmable “nodes” with sensing and computing capabilities. It is composed of scientists from a variety of institutions, including Argonne National Laboratory, the University of Chicago, Northwestern University, the University of Illinois, and the School of the Art Institute of Chicago. Its first urban sensor installment took place in 2016. 

The nodes collect slews of data, including levels of ozone, pollutants, temperature, humidity, noise intensity, and light intensity. They provide researchers and the public with real-time, location-based data about the urban environment, infrastructure, and activity. The intention is for city leaders, industry, and communities to collaboratively use the data to make cities healthier, more efficient, and more liveable.

The AoT team already was working with JCDecaux for a couple years to identify bus shelter sensor opportunities when Microsoft Research approached them recently about testing a new sensor design. The three entities decided to work together for the new bus stop sensor project, which was supposed to launch last summer but was delayed by the pandemic.

Iterative innovation

The previous AoT nodes contained numerous components and sensors, including electrochemical air quality sensors. The electrochemical sensors have a lifespan of about two years because the chemicals dry out, Catlett said. The AoT nodes also contained powerful “edge” computers and used wires to tap into a power source, so they were typically located on street signals or lights. Those nodes were located nearly 30 feet off the ground, which raised questions about whether their air quality measurements accurately represented the air that people breathe.

The new sensors are less expensive, easier to install, and located closer to street level. They do not have wires and instead run on solar power. They “talk cellular” so they’re easy to install anywhere. Part of the research involves assessing how long they last, and Catlett anticipates they will far outlast the previous versions. Plus, when the new devices do die, they are easier to swap out with replacements.

The selected bus shelters have a QR code that users can scan to see the last 24 hours of environmental data collected at that location. It gives information about the localized air quality index and PM2.5, tiny particulates in the air that affect human health. Users also can access a broader map that displays this same data for all the 100 node locations.

The AoT team has been conscious of community desires and reactions to new technologies suddenly appearing in their neighborhoods.

“We never go into a neighborhood where this technology is unwanted. We only go in where residents want to get data about their environment,” Catlett said.

Catlett said the bus stop project already has shown value in the few weeks since the devices were installed. Researchers had to be strategic about where in the city they placed the sensors, for example, and the incoming data is helping to determine if the product locations and project scope are adequate given the devices’ range and granularity. 

“We’re starting to get a good picture of whether 100 nodes is enough, and do we even need 1,000? Maybe we’ll only need 40 if they’re placed cleverly,” Catlett said.

Next steps

Catlett said there are plenty of areas to expand the research and develop real-world applications for the data. For instance, the Chicago Department of Health is aggregating air quality and heat data from multiple sources to influence their policies on air quality and to build a detailed dataset or map that shows neighborhoods most vulnerable to air pollution and excessive heat. 

“A project like this only increases the ability that DPH has to know what is happening across the city,” Catlett said.

These Microsoft sensors are not yet a marketable product, but rather a prototype that’s undergoing testing. A small number of devices have been tested in other cities, including Seattle and Miami. Chicago is the first city to test 100 or more. AoT views this as a good start but would like to further scale up to 1,000 devices in a city the size of Chicago.

“We can readily find funding and we’re ready to do 1,000 units. It’s just a matter of finding a vendor willing to produce them,” Catlett said. “Microsoft has been rightly cautious before moving forward with these. … My hope is Microsoft will have someone build these for them, or will build them [themselves], so we can buy another 900 and put them in Chicago.”

In addition, the Microsoft devices complement the other follow-on AoT project, a sensor network from the National Science Foundation’s SAGE project. Those sensors are higher capacity, better at artificial intelligence, and incorporate cameras. They can calculate traffic safety and risky environments in connection with the bus stop nodes’ data about air quality, sound, and heat.