“Smart Cities” Use Analytics to Improve Safety and Well Being

October 2018
Topics: Data Analytics, Enterprise Architectures, Data Management, Emergency Management, Data Encryption, Data (General), Innovation
Using new technology and data analytics, cities are harnessing a wealth of data flowing from devices, sensors, and social media. Our researchers are helping cities manage their resources, respond to emergencies such as hurricanes, and keep citizens safe.
Smart city

Cities around the world are striving to become "smart" by taking advantage of new technologies, such as digitization and next-generation electronic sensors, as well as the massive amounts of data they produce. Developing new analytics based on these technologies can help cities better manage the major challenges of urbanization by, making cities safer, more efficient in delivering services and resources, and more responsive to citizens' needs.

For example, Los Angeles is piloting a program to improve parking availability, while Liverpool is at work reducing its carbon footprint. Brisbane is driving a new wave of economic development, and Dubai wants to make traffic congestion a thing of the past.

MITRE has been researching the potential of smart city technology for several years. Robert Case, chief scientist for analytics, who oversees MITRE's Intelligent Cities (ICITY) Innovation Area within our independent research program, says the potential applications are staggering. He can imagine a day when cities will use the approach to guide economic development, reduce crime, lessen the impact of natural disasters, manage infrastructure upgrades, control the airspace, and reduce the carbon footprint.

"We definitely have a grand vision for this," he says. "And one that fits right into MITRE's mission, which is solving problems for a safer world."

To achieve this, however, most cities must overcome formidable challenges. For instance, how can a city combine the overwhelming amount of data from dramatically diverse sources into a coherent stream that will help identify meaningful information for cities? How does a city combine the content of a tweet with the output of a traffic congestion sensor? A 911 call with satellite imagery?

"Cities increasingly have more data on human activities, mobility, and the environment. This coincides with the emergence of the field of data analytics, which can be applied to applications (apps) to help citizens in their daily life and help city leaders plan and manage operations. Much of the innovation will come from the citizens of smart cities," says Jennifer Mathieu, principal investigator of the ICITY umbrella research project.

"With increased automation and ubiquitous city sensing, the challenge is how to foster data-driven processes that integrate numerous, diverse data sources at a variety of temporal and spatial resolutions, and that produce dynamic, actionable information for citizens as well as city planners and managers."

Jennifer Mathieu, principal investigator for MITRE's "smart cities" research project, envisions an ever-increasing role for data analytics in helping citizens and civic leaders improve daily life in urban areas.

Smart Cities Need Smarter Engagement, Tools, and Decision Making

In other words, smart cities need ways to embrace smarter engagement with stakeholders and tools to support making decisions. MITRE is developing an engagement framework and a set of analytics that could help fill this void. Our goal is to produce new technology that both combines disparate data sources and moves the needle toward prediction, rather than simple reaction. These approaches could have important applications in everything from security and safety to increasing quality of life—for example, by improving environmental protection, traffic flow, and access to city services.

MITRE launched its ICITY work in 2014. When Mathieu joined the project a year later, she brought with her expertise in social media analytics, which opened new areas of exploration. While access to some data can be an issue, social media data—text, images, videos, audio recordings, comments, reviews—is both abundantly available and often of high quality. Social data integrated with city data can provide analytics to support decision-making.

When a City Floods

Mathieu and her team were looking for a real-life example they could use to show the value of smart city initiatives to local and state governments. The team joined forces with the College of William and Mary in southeast Virginia to create and test a proof of concept that would help the city deal with flooding. The National Oceanic and Atmospheric Administration has identified the Hampton Roads region as the second largest population center most at risk for rising sea levels in the country.

The team first identified and assembled appropriate data sources and then developed new analytics from the available information to extract patterns, forecast events, and provide options to respond to events more efficiently.

The team then applied the proof of concept to flooding events (stranded motorists, unpassable roadways, broken levees) in the aftermath of Hurricane Matthew in late September and October 2016. The team combined data from a range of sources, including Twitter, 911 calls, webcam feeds, and ship locations, to create predictive analytics and a dashboard to show the information to decision-makers.

The use case demonstrated that the MITRE methodology could successfully integrate information from disparate sources to spot transportation disruptions in time to reroute emergency response vehicles and ensure timely responses. Under normal conditions, for instance, emergency response vehicles would have been deployed from a rescue station seven minutes away from the event. But using this methodology, help could be dispatched from a different station and with an arrival time within four minutes.

"Three minutes may not seem like much on its own, but it provides cities a return-on-investment example of why they should install more sensors, share more data, and invest in infrastructure to do analytics," Mathieu says.

Situational Awareness on Land and Sea

While Mathieu's focus is on land applications, such as improving traffic flow and identifying event hotspots, her counterpart—researchers Ryan Hollins and Jeff Gold—are applying the same "smart" strategies to a city's marine realm. A city port is its own unique microcosm with established patterns of activity, including container vessels coming in to load and unload, ferries carrying commuters, fishing trawlers looking for the daily catch, and recreational watercraft, such as kayaks and sailboats.

This research project is working to increase situational awareness at ports by creating a baseline of normal activity for each unique setting. Such a baseline will provide the foundation for anomaly detection—allowing stakeholders to recognize unusual activity, including the navigation of fishing vessels or the emissions from vessel smokestacks.

Ports that use MITRE's approach to automate the collection and analysis of data would not only be able to react to these outlying events, but also to predict them. "Right now, ports have a good grasp on near-real time activity. For example, they can tell if a ship goes into an environmentally sensitive area," Hollins says. "Yet they often lack analytics to detect patterns and deviations from normal activities, which could vary from port to port.

"In addition, such analytics could help ports move from being reactive to more proactive. For instance, instead of alerting when a vessel enters a sensitive area, can we alert when we have a high confidence that the vessel is likely to enter the sensitive area in the next hour based on historical navigation and trajectories? This is how a port can move beyond situational awareness and plan a response to likely events before they happen."

Hollins' approaches are currently being tested at the Ports of San Diego, Long Beach, and Baltimore.

Prediction Is Key

Many of these capabilities, which move us closer to true predictive analysis, are the latest milestone in MITRE's work in social analytics. Six years ago, the research program created the Social Radar technology suite, which uses machine learning and other approaches to integrate and analyze social media to track attitudes and behaviors around the world. (We've licensed many of these tools to government, industry, and non-profit organizations.)

Combining online sources with physical data sources, however, is a significant step forward. "In adding physical sensors to our understanding of human behavior, we can better understand not only what people are saying and doing, but what, in the relevant physical world, strongly correlates with such behaviors," says Barry Costa, director of MITRE's Technology Transfer Office, and a developer of the Social Radar tools.

"Imagine an accident, an earthquake, a flood, or a pandemic disease . . . our goal is to create capabilities that will predict the human behaviors related to the event and put the necessary remediation in place to help."

Costa says while many companies are working in the Smart Cities space, few organizations are addressing the specific problem of combining disparate, multi-type city data sources, which is an area in which MITRE has tremendous experience. And people are listening: This past June, Mathieu and her colleagues presented some of their ideas to an audience of traffic-monitoring professionals at the National Travel Monitoring Exposition and Conference.

MITRE believes many cities could benefit from applying our methodologies to increase their safety and security, for example by improving traffic flow and transportation systems, identifying crime spots, and managing emergency responses.

—by Twig Mowatt

Explore more at MITRE Focal Point: Innovation.

Learn more about how MITRE works as a mission partner with cities and states or contact us: citiesandstates@mitre.org.


Publication Search