11 minute read
Urban Innovation
INCREASED URBANIZATION CALLS FOR INCREASED LIVABILITY
With more than half of the world’s population now residing in urban areas and that figure expected to rise to more than 70% by the middle of the century, NYU Tandon is helping envision the cities of the future.
Navigating the world
In partnership with Toyota subsidiary Woven Planet Holdings, Tandon researchers have compiled a dataset of more than 200,000 outdoor images over the course of a year aiming to help visually impaired pedestrians and autonomous vehicles alike better navigate complex urban settings.
The dataset — developed by a team led by Assistant Professor Chen Feng (CUE, MAE, CSE) at his AI4CE Lab and sponsored by C2SMART (Tandon’s U.S.DOT Tier 1 University Transportation Center) and VIDA — is being used to test a range of visual place recognition (VPR) technologies that can improve the accuracy of personal and automotive navigation applications and promote independence for a variety of users. It uses side-view images of sidewalks and storefronts in addition to forwardfacing imagery, allowing researchers to test more applications than traditional mono-perspective sources. For example, side views support navigation for people with impaired vision who navigate in 360 degrees across busy city sidewalks, and the data could also help improve delivery robots, which must move forward and back as well as side to side to reach homes and businesses.
This marks the first work to systematically analyze some of the biggest challenges of visual place recognition, and in addition to providing data from multiple viewpoints, the dataset uniquely
• Captures long-term changes of the same urban area for a year, so researchers can improve VPR under varied conditions like snow and heavy foliage.
• Anonymizes images to protect the privacy of pedestrians and cars. The anonymized images also provide VPR algorithms static and environmentonly information
The NYU-VPR dataset was published at IEEE IROS in 2021, and is publicly available for research and education purposes at https://ai4ce.github.io/NYUVPR/#download
(TRAFFIC) LIGHTS, CAMERA . . . ACTION
In the midst of the COVID-19 pandemic, many research groups sought an effective method to determine mobility patterns and crowd densities on the streets of major cities like New York to provide insight into the effectiveness of stay-at-home and social distancing strategies. But sending teams of researchers out into the streets to observe and collect this data would have involved putting them at risk of exposure to the very infection the strategies were meant to curb.
At C2SMART, which is funded by the U.S. Department of Transportation (DOT), a team co-led by Professor Kaan Ozbay (CUE and Director of C2SMART) and Senior Research Associate Dr. Jingqin Gao (CUE, C2SMART) developed a solution that not only eliminated the risk of infection to researchers and that could easily be integrated with already existing public traffic camera infrastructure, but also provided the most comprehensive data on crowd and traffic densities that had ever been compiled. In order to cover hundreds of locations with 24-hour coverage over many months, the job had to be performed by an AI algorithm instead of human or conventional traffic counters. There are complications that the AI has to overcome from each video feed: the locations are different, the camera angles and height are different, and they are subject to different lighting and positional factors. In other words, it’s difficult for the traditional AI models to learn just one intersection and automatically apply it to another one.
To enable their AI solution, the C2SMART researchers started with an object detection model, namely You Only Look Once (YOLO), which is pre-trained using Microsoft’s COCO dataset. Then they retrained and localized the object detection model with additional images and various customized post-processing filters to compensate for the lowresolution images produced by New York City DOT video feeds that are publicly available. Although the off-the-shelf object detection model could work in this instance with some customization, when it came to measuring the distances between the objects, the team had to develop a novel algorithm, which they refer to as a reference-free distance approximation algorithm.
While the project was inspired by the COVID-19 pandemic, it demonstrated to several New York City agencies that they were sitting on very valuable actionable data that could be used for many different purposes. Ozbay, Gao, and Ph.D. student Fan Zuo (CUE, C2MART), along with a number of graduate and undergraduate students from the Tandon School of Engineering and NYU Abu Dhabi, are now working with New York’s Department of Design and Construction (DDC) and the NYC DOT to use the same kind of approach to analyze traffic around work zones and other key facilities such as intersections and onstreet parking without them needing to actually go out to those locations. They are also incorporating such approaches in predicting pedestrian intention to other emerging fields, such as connected and autonomous vehicles to improve road user safety. They hope that the success of the technology will ultimately lead other DOTs across the country to adopt it themselves.
Additionally, they report, the project has been an excellent platform for engaging graduate, undergraduate and high school students from NYU’s ARISE program in the use of the most advanced AI/ML technologies to solve some of our cities perennial problems.
NYU Tandon professors team up to explore complex urban systems
Each issue of the European Physical Journal Special Topics is focused on a specific subject, with particular emphasis placed upon interdisciplinary topics in physics and related fields. With experts predicting that more than two-thirds of the world’s population will be living in cities by 2050, complex urban systems constitute an area ripe for exploration.
In 2022, Tandon’s Vice Dean for Research, Innovation and Entrepreneurship, Professor Kurt Becker (MAE, AP); Professor Juan Pablo Bello (CSE, NYU Steinhardt, and former Director of CUSP); and Institute Professor Maurizio Porfiri (MAE, BME, CUE, and Director of CUSP) approached the journal with the idea of devoting an issue to that ever-evolving field. Rapid urbanization, they pointed out, poses numerous challenges, including transporting people and goods efficiently, ensuring reliable electrical power and water, considering the health and wellness needs of populations that include older adults and people with disabilities, preparing for the waves of migration expected in the wake of climate change, and building an accessible and inclusive environment. None of those challenges can be addressed in isolation, they explained, since urban environments represent complex, intertwined socioeconomictechnical-environmental systems that require cross-disciplinary scientific and technological approaches — just the type of approaches they and their colleagues were taking on at Tandon and elsewhere.
Becker, Bello, and Porfiri curated a collection of 12 peer-reviewed articles, half of which were authored or coauthored by Tandon faculty or doctoral students, including pieces on what impact a reduced police force may have on response time — a salient concern given recent calls by some activists to defund the police; ways to improve the generally poor performance of most roof-integrated wind turbines; the design of efficient public transit dispatch systems; and more.
Science and technology are key in meeting the complex urban challenges of the present and future, Becker, Bello, and Porfiri say, and the research presented in the European Physical Journal Special Topics issue on complex urban systems proves just how true that is.
The new Director of Tandon’s Center for Urban Science + Progress is bringing his knowledge of complex dynamical systems, including how a deep sea sponge stands up to extreme underwater pressure, to improve urban environments.
COMPLEX ENVIRONMENTS LIKE CITIES POSE COMPLEX CHALLENGES, AND THE NEW DIRECTOR OF CUSP THRIVES ON COMPLEXITY
How do animals behave? How do people reach a consensus? How does a pandemic spread? What drives gun purchases? What happens when particles collide? Can a deep sea sponge hold the answer to stronger skyscraper design?
On the surface, those questions appear to have little in common, but Maurizio Porfiri sees a clear connective thread — dynamic, complex systems whose key to understanding lies in leveraging a distinctive combination of data science, physics principles, and other aspects of engineering. Porfiri’s research space at Tandon is called the Dynamical Systems Laboratory precisely because of his fascination with the unseen relationships in data sets both big and small that reveal how complex systems work. And as an Institute Professor with appointments in the Departments of Mechanical & Aerospace Engineering, Biomedical Engineering, and Civil & Urban Engineering, he’s able to draw on an uncommonly diverse set of expertise to shine a light on previously undetected patterns that can help model outcomes for extremely difficult scenarios.
Now, he is taking that expertise and a unique vision for the potential of urban science to lead CUSP as its new Director.
CUSP had its genesis in 2012, when New York City issued a call for a new kind of academic center that would ensure that the city remains a world capital of applied sciences and technology, dramatically grow its economy, and function in collaboration with nonprofits, the public, and the city itself.
Since its inception, CUSP has grown to become a thriving research hub that uses New York City and other urban areas as living labs to develop novel data- and technology-driven ways to improve city services; optimize decision-making by local governments; create smart urban infrastructures; and address challenging issues such as crime, environmental pollution, and public health. The Center has been celebrated
for major projects that make urban communities safer, healthier, and more livable, including SONYC, an initiative to measure and mitigate noise pollution — a top-ranked quality-of-life concern for all New Yorkers; FloodNet, which in collaboration with the U.S. DOTfunded C2SMART seeks to monitor and predict local flooding; and WE-SAFE, an interdisciplinary project to understand and engineer the firearm ecosystem in the U.S.
And — with the help of a large group of faculty members whose expertise encompasses AI, assistive technologies, data science, environmental science, human-technology interaction, public health, robotics, and numerous other fields — Porfiri foresees CUSP moving in an even more expansive direction. Porfiri’s unique expertise in complex systems makes him well-suited to lead a research center at the forefront of defining what a post-pandemic, modern city can — and should — look like. It’s an especially pressing issue since for the first time in history more than half of the world’s population lives in urban areas; in just a few more decades, that figure will increase to 70 percent. Enabling those cities to deliver services effectively, efficiently, and sustainably while keeping their citizens safe, healthy, prosperous, and well — informed will be among the most important undertakings of this century.
Building upon the work of previous directors, Porfiri plans to hone in on three key application areas: urban environment, urban health, and urban infrastructure — each a cornerstone of equitable, liveable, and sustainable cities. Solving real-world problems in any of these application areas demands fundamental research in a trio of core methodical foundations, which define the full research cycle in urban science, from data collection (sensing) to data analysis (informatics) to the modeling, design, and diagnostics of complex urban systems (complexity).
Working toward the vision of a research center that builds bridges between departments in the School of Engineering, various schools of NYU, and the university and the world’s metropolitan areas, he plans to create
• A research partnership program for public and private organizations that will be invited to engage in informing
CUSP’s research agenda, participate in the life cycle of collaborative grants and projects, and access a pool of exceptionally qualified potential employees
• An Urban Science Colloquium to engage the overall NYU community in urban science and learn from community stakeholders about their needs, values, and desires as NYC residents
• A series of CUSP-led discussion groups on critical urban issues for the general public in partnership with student associations at the School of Engineering: the first planned series, Engineering Safer Cities, will focus on deconstructing violence, criminality, and neighborhood security.
• An interdisciplinary track in urban science for Ph.D. students at the
School of Engineering — paving the way for the first cohort of doctorallevel urban scientists in the history of the school.
That last initiative will allow Ph.D. students to leverage their specific expertise to address urban processes. A biomedical engineer, for example, might choose to explore how more accessible infrastructure can improve the lives of those with mobility challenges, or an electrical engineer could aim to bridge the digital divide by focusing on expanding internet access to underserved communities.
“Cities have social, technical, and environmental components,” Porfiri asserts, “and engineers like those we’re educating at CUSP address the important intersections between them. We want them to be at the leading edge of contributing to fundamental urban science while providing timely, practical, and actionable insights.”
Maurizio Porfiri