An IoT-Driven Vehicle Detection Method Based on Multisource Data Fusion Technology for Smart Parking Management System

Recently, with the increasing difficulty of parking in cities, the wireless vehicle detectors (WVDs) based on the Internet-of-Things (IoT) technology and magnetic sensor have been widely used to obtain parking space status data, which lay the foundation for smart parking management system (SPMS). However, there is a weak magnetic region under every vehicle where the magnetic variation is even smaller than that distorted by adjacent vehicles, which leads to the accuracy decrease of the magnetometer-based vehicle detection algorithm (MB-VDA). To improve the performance of the WVDs, an IoT-Driven vehicle detection method that combines the heterogeneous data feature of the UWB channel with that of the magnetic signal is proposed in this article. In the proposed method, the length of the propagation path and signature of the channel impulse response (CIR) with respect to the vehicles, which can be obtained from UWB modules, are introduced to achieve vehicle detection. Moreover, considering that the energy consumption of UWB modules is much higher than that of magnetic sensors, the UWB-based vehicle detection algorithm is only activated when the signature of magnetic signals cannot enable MB-VDA to output an accurate decision. Finally, the proposed method was evaluated in a commercial parking lot under different conditions. The experiments show that it has a detection accuracy of 98.81% when the sampling rate of the magnetic sensor is 1 Hz.