A Study on Congestion Estimation Method Using Wi-Fi Packet Sensors
Gakuto Sato, Koichi Gyoda · 2025
A set of methods has been proposed to estimate the level of congestion in multiple places by counting the MAC addresses assigned to the Wi-Fi interfaces of mobile devices. Traditionally, since MAC addresses were unique to each interface, the number of MAC addresses is equal to the number of devices, enabling an estimation of the number of people carrying such devices. However, from a privacy perspective, a function that periodically randomizes MAC addresses has been introduced, making estimation more challenging. To solve this problem, a method is proposed to improve estimation accuracy by identifying the same device using the signal strength of probe request frames, transmission characteristics, and the increment of sequence numbers. Experimental evaluations of the proposed method's performance are also presented.