Faculty Tech Talk: Building Smarter Cities

May 20 2019 | By Jesse Adams | Culligan Photo Credit: Jeffrey Schifman | Smyth Photo Credit: Eileen Barroso | Jiang Photo Credit: Timothy Lee Photographers

So-called “smart cities” promise a future of better living through data; urban environments where new technologies optimize services and streamline waste out of the system. But along with this bold vision, comes big questions about who can access that data stockpile and how these new technologies impact our quality of life.

As cities get smarter, engineers will increasingly be the ones to shape our experience of built environments, noted three Columbia Engineering faculty at the forefront of the field, in a public conversation with Dean Mary C. Boyce on April 11. By working closely with communities and prioritizing resilience and sustainability over monetizing data collection, researchers can ensure they’re creating biophilic, human-centered environments that safeguard the planet while best serving people.

Sitting down in Carleton Commons, Professors Patricia Culligan, Andrew Smyth, and Xiaofan (Fred) Jiang delved into recent insights gleaned from their labs, broader ideas on coming trends in urban technology and key obstacles delaying implementation.

Culligan, the Robert A.W. and Christine S. Carleton Professor of Civil Engineering, explores novel solutions to urbanization challenges. Applications for her work include green infrastructure, social networking, and advanced sensing technologies for water and energy management.

Smyth, a professor of civil engineering and engineering mechanics, faculty director of the Carleton Strength of Materials Laboratory, and chair of the Smart Cities Initiative at Columbia’s Data Science Institute, specializes in structural health monitoring and combining sensor information with artificial intelligence to determine the condition of critical infrastructure.

Jiang, an assistant professor of electrical engineering and director of the Columbia Intelligent and Connected Systems Lab, concentrates on areas such as cyber-physical systems, smart and sustainable buildings, and data analytics.

We’ve excerpted a few edited highlights of the conversation below.

Patricia Culligan

Andrew Smyth

Xiaofan (Fred) Jiang

Q: Could you give us a quick overview of how each of your work is contributing to smarter cities?

Fred Jiang: In my lab, we’re focusing right now on air quality, energy, safety, and health. A few projects we’ve been working on are pedestrian safety, recognizing that people are paying more attention to their phones than to crossing intersections. We’re adding sensors and using machine learning algorithms to detect and localize vehicles potentially in a collision trajectory with you and sending you a warning in advance. Another one is called energy footprinting; we’re calculating the individual energy responsibility associated with every one of us 24/7. Our system tells you—in real time—how much energy the building is spending on your behalf to make you comfortable, and we use algorithms to discover what concrete actions you could take to save energy. Also, we’re working on using computer vision to quickly identify mosquito breeding grounds and monitoring perspiration levels with wearable sensors using conductive fabrics.

Patricia Culligan: Right now, infrastructure systems essential for the sustainability of our cities are under significant stress. A lot of the infrastructure that we have in cities like New York is aging past its design life and was never even designed for the functions its being asked to serve as the city population has expanded. In other cities, where population expansion is occurring more rapidly than planned, people find themselves living in informal settlements without any, or with very few, infrastructure services. We also have situations, like Detroit, where population densities have significantly decreased over time, leading to severe infrastructure maintenance issues because there isn’t the tax base to provide maintenance revenue. Plus, there are the impacts of climate change to think about, and integrating rapid innovations into our infrastructure systems. A group of us here at Columbia Engineering are interested in whether there’s a new way we can deliver infrastructure services using localized or distributed infrastructure solutions, which are in contrast to the centralized infrastructure that’s been the pattern of development in the last century. Examples include getting energy from solar panels as opposed to big coal-burning power stations, and dealing with wastewater within a building rather than sending it to a centralized treatment plant. Localized or distributed infrastructure can evolve to meet whatever needs come up in the future, financing can occur over time and there isn’t a need to lock into one specific technology. We’ve been asking what the best mix of local vs. non-local infrastructure might be, depending on the city typology. We’re trying to understand how a distributed infrastructure system, which can be composed of tens of thousands of different components, behaves as a system, so we have to use quite advanced sensing and statistical analysis.

Andrew Smyth: Primarily the way I approach the challenges of smart cities is looking at the opportunities that sensing and data can afford us to understand the city really as an organism. There’s an incredible opportunity at the convergence of sensing and data science to learn how a system actually works. We did a project with the New York City Department of Transportation learning from their data from a fleet of 27,000 vehicles. We suggested to them that we could take this mobile network of sensors to learn about the road network and accumulating the data over time we’ve learned a tremendous amount. We were able to create a profile of every single road segment all over the city, during different times of day, days of the week, and times of year to sort priorities from the point of view of safety, so as to inform a planner perhaps where to put in a speed bump or bike lane. City planners have had intuition and a lot of experience but not the raw data to make connections, so this kind of study really unlocks that for them across lots of different applications. In a completely different context, data applies to structural health monitoring; you have to manage and keep infrastructure alive while you’re using it. Typically, I’ll instrument a bridge with a variety of inertial sensors and develop a math, physics, and data driven model that can be updated over time, and that can inform decisions. What we’re looking to do now in the Smart Cities Center is to model how humans actually interact with buildings and provide improved design and management over time, and the long-term vision is heading toward the city scale.

A group of us here at Columbia Engineering are interested in whether there’s a new way we can deliver infrastructure services using localized or distributed infrastructure solutions...Examples include getting energy from solar panels as opposed to big coal-burning power stations.

Patricia Culligan
Robert A.W. and Christine S. Carleton Professor of Civil Engineering

Q: Urban centers are comprised of many elements—people, buildings, transportation, and the environment, as well as the matter of how these elements work together. With all the new technologies on the horizon, what strategies might we be able to employ at different inflection points across the system?

Culligan: That’s a really complicated question, and in order to answer it we need to understand a lot more about how people use infrastructure and what they value. For instance, with green infrastructure, there’s a move in the city to plant these systems with native plants, of which there are many grasses. It turns out many people dislike grasses because they’re untidy, they’re unkempt, they’d much rather have decorative flowers like roses. To design ideal buildings, we need to mesh human behavior with infrastructural needs and environmental quality to find the optimal nexus, and I think our buildings are going to have to be adaptive.

Jiang: In the past we’ve seen buildings as these sorts of cold objects that we simply spend some time in, but moving forward we should really think of humans and buildings as coexisting. As buildings become more aware of their occupants and adapt to the people inside, in some way in the not too distant future humans can also be aware of the needs of the buildings and adapt to those. We can both reduce energy consumption overall and also improve the comfort level of those in it. Hopefully we’ll see this happening very soon.

Smyth: A focus on adaptation is really key, and that’s a really difficult thing to do. How do you make the building change? One pesky thing about hardware of infrastructure is that it’s very expensive, very bulky, and usually quite big. I think one thing that’s going to come into the fold is the area of robotic construction—so 3D printing is a good strategy for making small things, but in the case of 3D printing buildings, the building becomes bigger than the robot. Assembling things in a modular way is easier for a robot to do, and makes it easier to change and adapt the structure.

Q: When it comes to urban design, there are already a number of noteworthy technologies available that haven’t been widely implemented. Can you talk a bit about what you think is impeding wider adoption?

Culligan: The answer to that is something we all need to own. There’s a significant gap between the knowledge that we have and action, and I think we need to be much more deliberate about ensuring that gap closes. There’s new technology available, we know that it’s good. We have to be a lot more deliberate about getting it out there.

Jiang: I also think that sometimes economics and the cost of things go against adoption of innovation. One thing I’ve personally run into is that the cost of electricity is too low: when we try to propose some new project that will save this many kilowatt hours per month, people run a calculation of how long it takes to recoup the investment in hardware. Right now energy is very cheap and it’s hard to convince others. We need to be more farsighted.

Smyth: I don’t think there’s a really good solution other than better foresight, better planning, and thinking about future loads and having future redundancy built in—it’s that notion of adaptation again. But it’s sometimes difficult to forecast the future…Technology can help us to an extent, but there has to be a willingness to communicate needs and look for help. I see these kinds of very basic human types of elements as being essential to overcoming that gap between knowledge and need.

9 Ways to Make Your City Smarter