Estimation of saturation degree for lane groups at signalized intersections using video imaging detection data

the purpose of this paper is to present a new estimation method for the saturation degree that is suitable for both unsaturated and oversaturated conditions, using the data collected by video imaging detectors. We start by analyzing the spatial and temporal variation process for vehicles, from the time they enter to the time they leave a given link, and introduce the notion of the virtual cycle to represent the varying pattern of vehicle travel times or delays. Then, we identify the critical time points in delays, and propose relationships between the parameters of the signal cycle and the corresponding virtual cycle. Next, the process of saturation degree estimation is divided into four parts according to the traffic states of the previous and current traffic signals, and four models are presented based on the number of vehicles entering the link during one virtual cycle being just equal to the number departing from the downstream stop line within the corresponding signal cycle. Finally, the precision of the new models are evaluated using the traffic data collected by the VISSIM simulation, and the results show that the average absolute and relative errors during this simulation period are 0.2083 and 26.0422% respectively.