What is Array of Things?

The Array of Things (AoT) is an experimental urban measurement project comprising a network of interactive, modular devices, or "nodes," that are installed around Chicago to collect real-time data on the city’s environment, infrastructure, and activity. These measurements are published as open data for research and public use. AoT essentially serves as a “fitness tracker” for the city, measuring factors that impact livability in Chicago such as climate, air quality and noise.

The project was funded by the U.S. National Science Foundation and by late 2019 roughly 130 nodes were operating throughout the city, typically at street intersections, 24-28 feet above the street. The original project plan was to grow to 150 nodes by mid-2020, involving both replacing older nodes (some of which were installed as early as 2016) and adding locations. However, the COVID-19 pandemic has put all installation activity on hold for the time being (Nov. 2020).

The technology platform used to build the nodes is called Waggle, an open source hardware/software system developed at Argonne National Laboratory. AoT nodes are manufactured in Chicagoland. Waggle is being significantly expanded and enhanced through a new project, "SAGE: A Software-Defined Sensor Network," which began in late 2019. New, "SAGE" nodes have nearly 200 times more powerful processors and improved sensors and cameras. We expect to begin testing SAGE nodes in Chicago some time in 2021 (modulo pandemic issues). See "When and where will nodes be installed?" below.

What are the goals of Array of Things?

Array of Things was conceived as an experimental project to explore the concept of an urban-scale “instrument” to enable the City, urban planners, residents, and researchers to monitor and examine Chicago’s environment, infrastructure and activity, including detecting trends and changes over time. Ultimately, the goal is to measure the city in sufficient detail to provide data to help engineers, scientists, policymakers and residents work together to make Chicago and other cities healthier, more livable and more efficient.

The project is also a computer science research effort to understand how to create a resilient platform that can support not only traditional sensing but "software-defined" sensing. That is, nodes have remotely programmable computing systems in addition to sensors, and this allows programs to be run within the nodes to analyze the data, for instance "sensing" the number of vehicles at an intersection by using computer vision software to analyze images. This concept, often called "edge computing," also supports very strong privacy protections such as are covered in the project's privacy policies. modified and improved over time without

How does AoT benefit Chicago?

AoT provides near-real-time, location-based data about the city’s environment, infrastructure and activity to researchers and the public. Over time, these measurements will allow for studies of longer term changes such as how noise levels are affected by changes in zoning, or how air quality changes following investments in clean energy. In late 2018 a new set of application programming interfasces (or, "APIs") was released to support the development of applications, portals, and other tools to provide value from the data. For instance, several groups are developing smartphone applications that would provide navigation tips based on air quality, noise levels, or excessive heat. Others are developing services that would allow a resident to subscribe to air quality alerts. This initiative has the potential to allow researchers, policymakers, developers and residents to work together and take specific actions that will make Chicago and other cities healthier, more efficient and more livable.

The project also aspires to engage Chicago's youth in science, technology, engineering, arts, and mathematics ("STEAM"). Beginning in 2016, with funding from the Motorola Solutions Foundation, Chicago Public Schools faculty members at Lane Technical High School worked with the AoT team and with the School of the Art Institute of Chicago to develop a curriculum teaching students about wireless sensor devices, how they can support science and policy investigations. Over the course of eight weeks the students learn how to formulate their inquiries, how to design and build a wireless sensor device, and then how to analyze the data once they have installed the device to operate for several weeks. Over 600 students have been trained in the curriculum in the past four years. In July 2018 the faculty leaders provided a one-week training session to teachers from eleven other CPS middle- and high-schools, six of which elected to try the curriculum out in the 2018-2019 school year. Similarly, partners at the University of Chicago and at Northern Illinois University have developed tutorials for high-school and undergradate students to learn how to analyze and graph or otherwise visualize AoT data.

What data is collected by AoT?

Today the nodes measure temperature, barometric pressure, light, vibration, partiulate matter (PM2.5), carbon monoxide, nitrogen dioxide, sulfur dioxide, ozone, and ambient noise levels. A novel aspect of AoT nodes is that they can analyze images and sound without saving or transmitting the data. Today there are three generations of nodes installed, the earliest of which were installed in 2016, and each generation involved improvements in the enclosure and reliability features. A growign number of the newest nodes (primarily those deployed in 2019 and 2020) analyze an image from their cameras several times a minute to count the number of pedestrians and the number of vehicles, reporting these numbers and then deleting the images. Continued research and development in computer vision is being done to enable the nodes to measure additional factors such as flooding, recognizing vehicle type (e.g., bus, bicycle, automobile, or truck) and eventually more detail such as whether a vehicle is electric or gas-powered. Additional computer vision research is underway for analyzing the flow of pedestrians and vehicles.

Who has access to the AoT data?

Data collected by AoT is open, free, and available to the public. The nodes transmit their sensor readings every 30 seconds to a secure central database server at Argonne National Laboratory. Data is then published openly to allow individuals, organizations, researchers, engineers and scientists to study urban environments, develop new data analysis tools and applications, and inform urban planning. New data is published to a download site every 24 hours and to an API service every five minutes. For more information on using the API, see the API documentation site.

What can be done with this data?

Potential applications of data collected by the Array of Things include:

But importantly, the team expects that many unanticipated, innovative applications will be constructed by outside parties using the open Array of Things data, realizing the potential of an open, community-based initiative.

When and where will nodes be installed?

In partnership with the Chicago Department of Transportation (CDOT) Division of Electrical Operations (DEO), 100 nodes were installed in 2018. In early 2019 several dozen of the older nodes were replaced and the total number of sites grew to 130. The nodes are typically mounted on traffic signal light poles, though some of the nodes installed in 2019 were mounted on buildings (such as at the University of Chicago campus) or in other locations (such as at Crown Castle Communications cellular towers). The team continues to seek research and development funding to expand and evolve and improve the system with additional node locations and with new technologies in the coming years. Significantly improved nodes have been developed and should be ready for testing in early 2021 as part of a new project called SAGE: A Software-Defined Sensor Network"-- also funded by the National Science Foundation.

Node locations are chosen with input from researchers, neighborhood groups, city departments, and community members. The goal is to work with these and other groups to place nodes in support of local concerns and interests, such as in air quality, traffic safety, or noise.

The Nodes have cameras. How is privacy protected?

In late 2014 the AoT team drafted a set of privacy policies, governance structures, and decision procedures to ensure that the project would protect personal privacy in a transparent and accountable fashion. These policies include an executive oversight committee comprising Chicago leaders from industry, academia, government, law, and the community, and a process for reviewing privacy policies and their implementation, as well as change that might be proposed over time.

To protect privacy, all AoT images are analyzed within the nodes rather than transmitted and stored for later use. Analysis of images is also limited to pre-approved functions, such as "count the number of pededstrians" or "detect standing water (flooding)." The privacy does allow for a limited volume of images to be collected in order to develop the computer vision software, but these images are protected and only accessible to academic researchers who have signed a data use agreement. The full policies are availabe at AoT Policies, which also includes profiles of all of the members of the AoT executive oversight committee.

Will the nodes collect information about individual people?

No.

Array of Things is supporting analysis of, and monitoring over time, the city’s environment and activity, not individuals. In fact, the technology and policy have been designed to specifically avoid any potential collection of data about individuals, so privacy protection is built into the design of the sensors and into the operating policies.

I heard that sensors will be installed to monitor Bluetooth-enabled devices and WiFi signals. Is this true?

No.

None of the AoT nodes will measure the presence of Bluetooth/WiFi devices or communicate in any way with other devices. The nodes are programmed to communicate using cellular data, and only with the central database servers at Argonne. In its formative stages, the project considered including such sensors in the nodes, but decided against it for two reasons. First, research by other groups indicated that the measurements would be biased toward people who own connected devices and thus would not provide the most accurate indicator of foot traffic. Second, it was clear that most people (ourselves included!) would be apprehensive about a system that could contact their personal devices in any way.

What other cities will be using AoT nodes?

The project received over 100 requests from cities and universities to partner with the Array of Things team. Through an experimental partnership program, nodes were tested and in some cases also installed in Seattle, Portland, Palo Alto, Denver, Syracuse, Chapel Hill, Atlanta, Nashville, and Taichung City (Taiwan). Nearly all of these involved partnerships with local universities. Working with these partners we learned that the partnership in Chicago is unique in that many partners were unable to arrange for installation in their cities. The SAGE project invovles partners in several cities in Australia as well as Bristol (England), Champaign (IL) and Houston (TX). Nearly all of these projects are led by research partners at universities within those cities.

Who is the Array of Things team?

The Array of Things project is led by Charlie Catlett from the University of Illinois Discovery Partners Institute, with partners from the Northwestern-Argonne Institute for Science and Engineering (NAISE) and the Urban Center for Computation and Data of the Mansueto Institute for Urban Innovation at the University of Chicago. The underlying software and hardware platform, known as Waggle, was developed at Argonne National Laboratory by Pete Beckman, Charlie Catlett, Rajesh Sankaran, Nicola Ferrier, and dozens of student interns over multiple summers. The custom enclosure for the sensor nodes was developed by Product Development Technologies, based on early designs from Douglas Pancoast and Satya Mark Basu of the School of the Art Institute of Chicago.

While each AoT node contains commercial components from many sources, the majority of the custom aspects of the nodes are sourced in Chicagoland. The white sensor shields, for instance, are injection molded new O'Hare Airport. All of the custom circuit boards are made by Surya Electronics in Glendale Heights, Illinois. Surya also does all of the assembly of the nodes. The project is executed in partnership with the City of Chicago, with coordination by the City's Chief Information Officer (as of November 2020, Carleton Nolan).

The Array of Things project also involves partnerships between the Argonne and UChicago team and scientists at academic institutions including Northwestern University, Northern Illinois University, University of Illinois at Chicago, University of Illinois at Urbana-Champaign, DePaul University, Illinois Institute of Technology, Purdue University, University of Notre Dame, and DePaul University. The project has benefited from support and technical advice provided by corporate partners including AT&T, Arm, Cisco, Crown Castle Communications, Microsoft, Schneider Electric, Intel, Motorola Solutions, and Zebra Technologies.

Who is funding AoT?

The University of Chicago received a $3.1 million grant from the National Science Foundation for the AoT project. The project also received investments of $150,000 each from the Chicago Innovation Exchange and Argonne National Laboratory to fund the initial pilot project. The underlying platform technology is the result of over $1 million of internal research funding from Argonne National Laboratory. Additionally, the City of Chicago has provided in-kind support through installation and power for the nodes.

How much does an AoT node cost?

A full-equipped node cost roughly $2,500 for parts and assembly.

How much will this project cost the City of Chicago?

The City of Chicago provided node installation and supplies electricity to them from the City’s power supply via the traffic signal poles to which nodes are attached. The amount of electricity drawn by each node is roughly 20 Watts - less than what is required to charge a cell phone). The nodes, mounting supplies, and cellular communications were all funded through the National Science Foundation grant.