RESEARCH
RESEARCH
Research interests
Deep Learning, Computer Vision, Transportation and Traffic Safety Research, Autonomous and Connected Vehicles, Digital Twins and Smart Cities and Intelligent Transportation Systems (ITS).
Summary of my projects
While serving as a Postdoctoral Research Associate at the University of British Columbia (UBC), I focused on Autonomous Vehicle (AV) safety through the design of a safety-first trajectory planning algorithm and the development of an AV simulation platform using inverse reinforcement learning to simulate its behavior when interacting with other road users such as human driven vehicles, cyclists and pedestrians. This proactive approach to road safety, focused on AV interactions, resulted in my publication titled: “AV-Pedestrian Interaction Modeling Platform for Urban Environments: A Case Study in Four Major Cities”, which provides insights into the degree of competitive behavior between AVs and pedestrians and can be used for the development of more efficient and safe autonomous mobility systems.
My Ph.D. work has centered around the development of a speech-powered traffic prediction platform that harnesses a novel deep learning algorithm using Connected Vehicles (CV) data. Specifically, I designed a Graph Convolution - Gated Recurrent Unit (GC-GRU) architecture and demonstrated faster traffic forecasting with improved performance in comparison to state-of-the-art models. This research underscores the value of a multiscale approach, merging CV data with conventional sources, emphasizing CV data's superiority in detecting short-duration incidents. I further introduced an algorithm to rapidly process substantial CV data, achieving a 70-fold speedup in data processing for Missouri's entire day of unique CV journeys. This processed data aids a specialized UNet model to predict comprehensive traffic features. The interactive web application, powered by speech queries, provides visual insights into historical traffic data and prediction results. Collectively, this work offers a cutting-edge tool for transportation authorities to manage congestion effectively.
During my M.Sc, my research has centered around the influence of Transportation Network Companies (TNCs) on urban mobility, particularly in the Munich region. Despite the growing prominence of TNCs, there's a dearth of in-depth research in this area. I assessed Munich transportation users' willingness to pay for TNC services using a stated preference survey of 500 participants. Findings highlight TNCs' popularity among larger households and those with fewer cars. To understand TNCs' impact, I incorporated an incremental logit approach into the MATSim model. Notably, a larger TNC fleet significantly reduced waiting times without affecting in-vehicle trip durations. However, smaller fleets were more efficient during peak demand. While centered on Munich, this study offers broader insights into the effects of TNCs on urban mobility.
During my B.Sc., I studied the use of recycled concrete aggregate, sourced locally from Bee'ah's facility in Sharjah, UAE, as a sustainable alternative in new concrete mixes. By replacing up to 100% of coarse aggregate with recycled material, we aimed to produce concrete with a compressive strength of 30-35MPa over 28 days. The research was divided into two phases: analyzing the recycled aggregate properties and the design of the concrete mix, followed by shear strength tests on reinforced concrete beams made from the various mixes. Test results revealed that beams with up to 100% recycled aggregate exhibited performance characteristics similar to those made with entirely natural aggregate.
Technical Skills
Languages: [Python, Java, JS, SQL, C, MATLAB, R]
Frameworks: [Pytorch, Tensorflow, Keras, Scikit, Django, OpenCV, ReactJS, Node.js]
Softwares: [VISSIM, Synchro, SimTraffic, AutoCAD, Microstation, Carla]