Machine Learning-based Urban Mobility Monitoring System

• Published in: 2022 IEEE 19th Annual Consumer Communications & Networking Conference (CCNC) pp. 747 - 748
• Main Authors: Bertolusso, Marco, Spanu, Michele, Popescu, Vlad, Fadda, Mauro, Giusto, Daniele
• Format: Conference Proceeding
• Language: English
• Published: IEEE 08.01.2022
• Subjects: Cameras; Convolutional neural networks; Economics; flow monitoring; IoT; Localization; Machine learning; Motorcycles; neural network; Recurrent neural networks; Smart cities; Sniffing; Social impact
• DOI: 10.1109/CCNC49033.2022.9700694

This article analyzes how streams of people move around a city. Wi-Fi sniffing techniques and camera systems are used for classifying vehicles, people, bicycles and scooters. The system is able to detect the presence of people by sniffing the mac address of the smartphone's Wi-Fi radio interface and with the use of cameras to monitor the presence of people, vehicles, and other means of transport, classifying them through the contribution of neural networks. The proposed solution has a high level of reliability overcoming the limits of a single technological system that offers imprecise and inaccurate monitoring.