An Energy-Efficient Multi-Source Fusion Mechanism for Vehicle Detection Scenarios Based on Seismic and Magnetic Signatures

Currently, the magnetometer-based Wireless Parking Detector (WPD) powered by a battery has the advantages of low cost and convenient installation, which leads to its widespread use in collecting parking status information for Intelligent Parking System (IPS). However, existing magnetometer-based parking detection (MPD) schemes waste a lot of energy to process the redundant magnetic data when the parking space is free or occupied. Although energy consumed by WPD devices can be optimized by reducing the magnetic sampling rates, it may reduce the detection accuracy. For this reason, a geophone-magnetometer based parking detection method (GMPD) combing the signatures extracted from a self-powered geophone and magnetic sensor, is proposed. In the work, MPD algorithm can only be awakened when a vehicle-induced seismic feature is detected. Therefore, the magnetometer and microcontroller can sleep as long as possible when the parking space is free or occupied. Finally, the experimental results show that the overall accuracy of GMPD is 99.33%, and the energy consumption is only 18.07% of the existing algorithms, which provides a valuable reference for the WPD industry and reduces the pollution of waste batteries to the environment.