Estimating Bus Passenger Counts Using Wi-Fi Packet Sensors: A Background Noise Cancellation Approach

抄録

It is important for bus operators to understand the number of passengers on their buses in order to create operation plans, such as the number of buses and routes. This study uses Wi-Fi packet sensors on small buses to estimate passenger volume by detecting Wi-Fi signals from smartphones. It distinguishes passenger signals from non-passenger signals, treating the latter as background noise. The background noise is pre-estimated through regression analysis, using route characteristics like road width, land use, and the number of buildings as dependent variables. Several estimation methods, linear and non-linear regression and machine learning were tried and compared. The regression equation and machine learning method estimating the amount of background noise did not have a high coefficient of determination. It will be necessary to increase the number of independent variables for estimating background noise to improve the goodness of fit.