Dates
Wednesday, December 04, 2024 - 09:00am to Wednesday, December 04, 2024 - 10:00am
Location
Old Computer Science, Room 2114 & Zoom (see description for Zoom details)
Event Description

All are welcome!

Who: Hoyoung Kim

When: Wednesday, 12/4, 9 AM

Where: Old Computer Science, Room 2114

Zoom: https://stonybrook.zoom.us/j/3902477464?pwd=TVp0bmJWN1o5WkdyU3QzTC84dXMvZz09


Abstract:
The Intelligent Transportation System (ITS) is a sophisticated transportation system that leverages advanced technology, traffic information, and communication to enhance the efficiency and safety of transportation. Its primary goal is to automate and manage transportation systems by applying scientific approaches. ITS encompasses various components: traffic management, lane control, real-time signal control, and public transportation information systems. It is poised to become even more significant in the future autonomous driving era. This thesis focuses on developing ITS solutions specifically in GPS-unavailable environments and addressing data management challenges.

Autonomous vehicles heavily rely on GPS/GNSS for accurate location tracking. This is challenging in GPS-unavailable areas such as tunnels or densely built urban environments. To overcome this limitation, we develop a novel approach by providing simulated GNSS signals underground, allowing for localization using smartphone apps. Through successful tests in a 1.5 km tunnel, uninterrupted GNSS connectivity is achieved with an average distance accuracy of approximately 13 meters, even at speeds ranging from 30 to 70 km/h. This innovative solution provides a viable alternative for reliable and continuous localization in GPS-unavailable environments, contributing to the advancement of autonomous vehicle technology.


The implementation of ITS leads to a substantial increase in data volume due to continuous traffic monitoring. To address the challenges associated with data management, we introduce CLOUD-CODEC, a system designed to alleviate the burden on cloud servers. CLOUD-CODEC employs foreground-background distinction to optimize data processing. This approach enables fast encoding and maintains quality with a high VMAF (Video Multimethod Assessment Fusion) score. Furthermore, the system achieves impressive storage space savings. CLOUD-CODEC provides an efficient solution for managing the substantial amount of data generated by ITS, ensuring streamlined data processing and storage.

Event Title
PhD Thesis Defense: Innovative Solutions for GPS-Enabled Tunnel Navigation and Data Management in Intelligent Transportation Systems