Parking Detection Method Based on Finite-State Machine and Collaborative Decision-Making

Abstract

A high-accurate parking detection method for wireless magnetic sensor networks is proposed in this paper, which is based on the combination of finite-state machine and collaborative decision-making. The magnetic disturbance induced by a vehicle can be sensed and processed to obtain the availability of a parking space by a magnetic sensor. However, the main challenge lies in the difficulty of eliminating the interferences from adjacent vehicles that decrease the accuracy. The vehicles include moving vehicles on adjacent roads and parking vehicles in adjacent parking spaces. For simplicity and low-energy consumption, a multi-interim finite-state machine (MiFSM) is proposed to deal with the interferences from moving vehicles. Our method contains preliminary detection and final detection. Using MiFSM, most of the parking vehicles can be correctly detected in the preliminary detection. However, the adjacent parking vehicles may cause more complicated interferences. It is hard to distinguish these interferences from the disturbances induced by “weak-magnetic” vehicles above on the detecting sensor. The “weak-magnetic” vehicles cause small magnetic disturbance because of their high chassis or short car-body. Therefore, by using the collaborative information of adjacent sensors, a Dempster–Shafer evidence theory-based collaborative decision-making is developed to cope with these complicated interferences in the final detection. The experimental results show that our work has a significant improvement in detection accuracy, as about 99.8% for vehicle arrival and 99.9% for vehicle departure. The proposed method can also be extended for moving vehicle detection, speed estimation, and vehicle classification in the applications of intelligent traffic system.