A Cross-Correlation Technique for Vehicle Detections in Wireless Magnetic Sensor Network

Abstract

Vehicle detections are an important research field and attract many researchers. Most research efforts have been focused on vehicle parking detection (VPD) in indoor parking lot. For on-street parking, strong noise disturbances affect detection accuracy. To deal with vehicle detections in on-street environment, we propose two vehicle detection algorithms based on a cross-correlation technique in wireless magnetic sensor networks. One is for VPD, and the other one is for vehicle speed estimation (VSE). The proposed VPD algorithm combines the state-machine detection and the cross-correlation detection. In the VSE, speed estimation is based on the calculation of the normalized cross correlation between the signals of two sensors along the road with a certain spacing. Experimental results show that the VPD has an accuracy of 99.65% for arrival and 99.44% for departure, while the VSE has an accuracy of 92%.