AI Advances and Applications Virtual Seminar Series
Artificial Intelligence for Smart City Transportation by Dr. Sanjay Ranka
Professor in Department of Computer Information Science and Engineering
September 16, 2020
We live in an era of real-time gathering of information and its dissemination. The size of the digital universe is estimated to be around 50 Zettabytes today and expected to double every two years. The growth of exploitable data has the potential to trigger disruptive changes in the transportation sector and is having a profound impact in urban planning, intelligent mobility, and safety.
Vehicle loop detectors that have traditionally been deployed at intersections to detect the passage of vehicles have high deployment and maintenance costs; and are not always useful for observing the movements of pedestrians and scooters. The use of other modalities, such as Video, Radar and LIDAR has great potential to improve accuracy and timeliness in the detection of vehicles, pedestrians, bicyclists, etc. As part of the I-Street Trapezium project we are collecting this information at 27 intersections in Gainesville using Video Camera, Radar and LIDAR. Additionally, data is being collected from interactions of vehicles with onboard units (that mimic the behavior of future connected vehicles) with Road Side Units. This information is then synthesized to create a real-time city-wide traffic palette that is used to improve traffic safety and operations.
In this talk I will present our work on use of artificial intelligence and machine learning techniques for real-time incident detection, vehicle classification, space-time trajectories, near-misses and travel-time distributions of vehicles and pedestrians while maintaining individual privacy. I will also present our work on developing signalized intersection control strategies and sensor fusion algorithms for jointly optimizing vehicle trajectories and signal control for a mixture of autonomous vehicles and traditional vehicles at every intersection. All these applications directly impact traffic safety and operations.
This work is in collaboration with Anand Rangarajan and Lily Elefteriadou. It is supported by NSF Foundation (CPS and Smart Cities program), and Florida Department of Transportation. We will also like to thank the City of Gainesville and FDOT District 5 for their support and participation.