Saturation flow under Adverse Weather Conditions

Abstract

In the modeling of signalized intersections, one parameter is of crucial importance: the saturation flow rate. This value defines the number of vehicles that pass an intersection within 1 h of effective green time per lane. In this study, changes in the saturation flow were investigated under adverse weather conditions, such as precipitation or snow that covered the road surface. Data were obtained from video recordings, and a timestamp was recorded for each vehicle as its rear axle crossed the intersection. Subsequently, all observations were aggregated to a longer time interval. These measurements were then used to train a model by minimizing the squared error between model output and observation. The advantages of the model were the incorporation of various vehicle classes and the consideration of driving behavior at the beginning and the end of the green phase (start and end lags). These parameters were investigated under various weather conditions and showed that the saturation flow rate was significantly influenced by snow on the road surface. To improve traffic models, it is thus important to consider the dependence of the saturation flow rate on the weather. To adjust the saturation flow, adjustments in certain other parameters influenced by prevailing weather conditions were investigated in a microscopic traffic simulation.

Calibrating car‐following parameters for snowy road conditions in the microscopic traffic simulator VISSIM

“…In the pre-study ( [9]), effect of snowfall intensity has also been investigated, but it turned out that there was no influence on saturation flow or start-up delay. This is reasonable, since snowfall reduces visibility, which plays a minor role when speeds are as low as in urban areas.…”

“…In this work the target values are saturation flow rate on the through lane (vehicles per hour green per lane -vphgpl) and start up delay, observed during snowy road conditions as reported in [9] (see left column of Table 1 and 2). In that study all three investigated intersections were located on major arterial roads in Vienna with two or three lanes, dedicated only to through traffic (no mixed lanes).…”

“…Since the slope of a road has an influence on saturation flow and on start-up delay ( [10]), it has been ensured to select only intersections without slope. Data of that study ( [9]) originate from manually annotated video recordings, where snowy road condition have been observed on three different days. The green phase has been divided into equal measuring intervals, and the number of vehicles passing the stop line within one measuring interval has been determined.…”

“…Although only one simulation run has been performed for a given parameter setting, the measurements originate from different independent signal cycles. The saturation flow rate of the traffic model has been calculated in the same was as for the target values in [9], by applying a linear regression model on simulated vehicle data [12].…”

“…The main focus lies on adapting the traffic model to snowy road conditions (that can regularly be encountered in Austria and other mountainous countries during wintertime) without taking into account intensity of snowfall. Saturation flow rate and start up delay have been measured on the through lanes of signalised urban intersections in a previous investigation described in [9]. On right and left turn lanes, saturation flow rates are lower compared to through lanes ( [10], [11]).…”