Analysis of Queue Estimation Methods Using Wireless Magnetic Sensors

Abstract

Four queue estimation methodologies were studied with wireless magnetic sensors installed on a single-lane loop on-ramp. Queue length estimation based on (a) occupancy measurements at the ramp entrance, (b) vehicle counts at the on-ramp entrance and exit, (c) speed measurements at the ramp entrance, and (d) vehicle reidentification were considered. These queue estimation methods were evaluated with available raw and processed sensor data retrieved from the test site through mobile data communication and downloaded from a server. The accuracy and reliability of the queue estimation methods were studied with ground truth data obtained from video. Discrepancies between inconsistent sensor and ground truth data were identified and further analyzed with raw data coming directly from the vehicle detection system at the on-ramp. Each of the methods analyzed had deficiencies that needed to be taken into account and compensated for if they were to be used for applications that involved ramp metering with accurate queue control. The worst performance of the queue estimation methods was observed when the on-ramp was under congestion (queue extends around or beyond the ramp entrance). On the basis of the observations at the on-ramp test site, it was also possible to point out some of the main factors that affect the performance of the queue estimation methods.