During the Summer of 2024, CITIES will offer several Summer Research Program positions that provide students with hands-on research experience under faculty supervision.
Spatial Distribution of Probe Vehicles in Urban Network
Student: Komal Neupane
Supervisor: Yinjie Luo
The role in the project of a student research assistant is Modeling, Programming, Data Processing, and Analysis. The probe vehicle is an ordinary vehicle equipped with GPS, which can measure and report the real-time traffic state. The research focuses on data-driven modeling and usually does not involve working with hardware devices. The core problem discussed in the research will be around “How to establish a generalized deployment strategy for probe vehicles in urban networks, considering the structure of the network and time-varying traffic demand.”The role of a student research assistant encompasses harnessing available datasets to derive meaningful insights into urban traffic dynamics. Utilizing data processing and analysis techniques, the student will work closely with the team to preprocess, clean, and analyze the probe vehicle data, enabling us to understand the spatiotemporal distribution of traffic conditions within the urban network. The responsibilities will extend to understanding statistical methodologies to determine the optimal penetration rate of probe vehicles required to capture traffic patterns effectively. Additionally, the student will also learn about and work on macroscopic traffic simulation, optimization, and signal processing to integrate data-driven insights into our models and strategies.
Comparative analysis of natural and anthropogenic influences on air quality dynamics in Abu Dhabi: The role of mangroves as a primary natural sink in the mitigation of airborne pollutants
Student: Sipan Hovsepian
Supervisor: Jose Balsa Barreiro
The methodology for this research will be based on empirical data collection from analytically selected field sites and an extensive literature review of existing pollution data archives. The selection of data collection sites will align with the specific nature of the pollutants, categorizing them into those of natural and human origin. The analytical focus will be to uncover spatial patterns of the prevailing pollutants and assess their potential interception and absorption by the city & mangroves. This will involve generating maps for visualization to aid in understanding these complex dynamics. The ultimate goal of this research is to evaluate the effectiveness of current environmental policies in Abu Dhabi and gauge the city’s inherent capacity for pollutant capture. By doing so, this study aims to provide valuable insights into sustainable urban planning and environmental conservation strategies tailored to the unique challenges faced by Abu Dhabi.
Time-resolved synchrotron micro-CT data analysis with machine learning-aided approaches for multiscale material characterization
Student: Will Mekvabidze
Supervisor: Kemal Celik
This project will leverage machine learning and computer vision techniques, specifically using PyTorch, to develop an algorithm for detecting cracks in images of processed samples. Additionally, the team will be responsible for documenting research findings and writing reports on the analyzed data as part of the project’s objectives.
Benchmarking study and implementation of autonomous navigation
Student: Uday Menon
Supervisor: Borja Garcia de Soto
A multi-agent robotic system, designed to autonomously navigate a construction site and capture relevant information through point cloud scans, will be developed over the course of this project. To this end, the student will be involved in the implementation of the navigation and exploration component of the project. To ensure that the best exploration and Simultaneous Localization And Mapping (SLAM) algorithms are utilized, it will be necessary to test the relevant algorithms prior to implementation. The algorithms will be tested in a simulation environment using the robot operating system (ROS). Additionally, the project will also encompass the real-world implementation of the robot, which will be thoroughly documented in order to produce a formal publication of the study.
Machine Learning cyber security
Student: Prakrati Mamtani
Supervisor: Muhammad Shafique
The role of this summer research will focus on Adversarial training in machine learning. Since the use of machine learning in our everyday life is increasing exponentially, especially after the release of chat gpt it is essential to consider the cyber attacks and risks associated with it. This project will focus on security against image data inputs and will analyze how a computer processor reads image data. The human brain contains an inbuilt mechanism to classify objects in micro-microseconds; however, with computers, it’s not that easy. There are multiple examples in which, although the image contains something, it is classified as something else. These types of images are called adversarial images. Hence, the primary objective of this research will be to construct various types of adversarial attacks, including the Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD), onto neural networks and develop a defense against these attacks.
SONIX (representing Sonicscape/ audiotape of NYUAD campus)
Student: Marwan AbdElhameed
Supervisor: Matteo Marciano
The SONIX project at NYU Abu Dhabi is a cutting-edge initiative to redefine the campus’s soundscape, combining detailed sound sampling and advanced analytics to uncover and mitigate noise pollution. By charting the ebb and flow of sounds—from construction din to the hum of daily traffic—using deep learning and traditional methods, SONIX aims to enhance urban acoustic environments. The project’s core lies in developing a machine-learning system for precise sound categorization, supporting an innovative platform for visualizing the audioscape. In doing so, SONIX not only seeks to illuminate noise sources and their effects but also to explore sound solutions that boost community well-being, setting a new standard for sound-conscious urban development.