Date of Award
Santa Clara : Santa Clara University, 2021.
Computer Science and Engineering
Traffic management is an important public infrastructure component that can greatly benefit from an intelligent reactive system. Vehicle counting is a vital aspect of an intelligent traffic system that could mitigate traffic congestion and increase the efficiency of traffic lights. While previous research has made progress in developing machine learning algorithms for vehicle counting and tracking, they did not have any hardware constraints and were not designed to run on IoT devices in real time. Our team proposes a decentralized method of vehicle counting and tracking through a system of IoT devices, namely Jetson Xavier NX’s, that communicate vehicle identification information with other identical IoT devices at a neighboring intersections through a custom protocol. By decentralizing the process and pushing the workload out to a system of IoT devices, our team achieves real time efficient and accurate counting and tracking of multi-class vehicles in a multi-intersection environment.
Ladhad, Jay; Rioux, Colin; Dong, Maggie; Tian, Chris; and Allen, Donovan, "Multi-Class Multi-Intersection IoT-Based Vehicle Counting" (2021). Computer Science and Engineering Senior Theses. 208.