Student Presenter(s): John-Paul Cantalino
Faculty Mentor: Michael Colef
Department: Electrical and Computer Engineering
School/College: School of Engineering and Computing Sciences, Long Island

This project was borne out of a terrible 45-minute commute from Bay Shore to Glen Cove at a job I used to have. During three years of sitting in traffic, I started to realize that certain intersections I was stopping at had traffic signals which were not intelligently programmed. Queues of vehicles were sitting at a red light, while no one was passing in the opposing direction. Some junctions had signals that were not synchronized, allowing long queues of vehicles to move from one junction to the next only to have to stop again barely 200 yards later. This led me to investigate several routes to fix these problems. The first of which is using data science and machine learning to optimize current junction timings. This is the cheapest but least adaptable option. The second was performing the same optimization but on a continuous basis using computer vision and cameras. Finally, the most expensive but likely most reliable option is using LIDAR technology to monitor queues as well as track vehicle approach speeds and use this data to intelligently change traffic light phases. These ideas create long term benefits with the most obvious being time savings for drivers. Indirectly, the economy is boosted by less man-hours being wasted behind the wheel, as well as a reduction in common accident types. This project is in its infancy still, but I do have some slightly working computer vision examples to display.