Vehicle/Pedestrian Localization System Using Multiple Radio Beacons and Machine Learning for Smart Parking

Abstract

In recent years, the number of casualties and injuries at intersections and roads has been decreasing due to wide spread of safe driving support systems, but the number of casualties and injuries due to low-speed traffic accidents in parking lots has not decreased. In the parking lot, it is necessary to drive while looking for an empty slot, which may result in contact accidents with pedestrians. Therefore, in this research, we propose a new smart parking system that prevents low-speed contact accidents by estimating availability of slots in the parking lot and the position of pedestrians. The proposed system attempts to estimate positions of user’s smartphones by deploying a small number of beacon devices on the parking lot, and by analyzing the radio wave intensity measured by the smartphones. In addition, estimation accuracy of the position of the pedestrian / driver is evaluated by experimental evaluation in a parking lot. Through the performance evaluation, estimation accuracy of the vehicle’s position to higher than 98%, and estimation accuracy of the pedestrian’s position is about 70%.