Roadside Sensor Based Vehicle Counting Incomplex Traffic Environment

Abstract

The 5G networks are expected to support autonomous driving to enhance driving experience and travel efficiency. Toward this goal, the valuable data generated by the complex and dynamic transportation system need to be collected. In this paper, we propose a roadside sensor-based vehicle counting scheme for collecting traffic flow information in complex traffic environment. In the scheme, the roadside sensor can sense the magnetic data, where the magnetic flux magnitude will be changed if a vehicle passes though the sense coverage of the sensor. Based on this, we first analyze the change of the magnetic signals in the complex traffic environment and process the magnetic signals collected by the roadside sensor. Then, an integrated algorithm is designed to detect and count the traffic flow by considering the features of the collected signals. After this, we carry out experiments to evaluate the performance of the proposed vehicle counting scheme and analyze the vehicle counting error. According to the features of the error, we further design the error compensation strategy to correct the experiment results. Experimental verification results show that the vehicle counting accuracy before and after the error compensation in the complex traffic environment are 97.07% and 98.5%, respectively.