Sight Distance Standards Based On Observational Data Risk Evaluation Of Passing

Abstract

The paper presents an application of a reliability analysis for evaluating the risk associated with passing sight distance (PSD) standards in terms of the expected probability of noncompliance. Calculation of PSD is required to determine where drivers can safely execute passing maneuvers. Traditional PSD standards are based on deterministic, theoretical models, which are calibrated with conservative percentile values for uncertain design inputs to account for uncertainty. The PSD standards do not provide information about the risk of deviation from them. Reliability analysis is a technique that is based on limit state design and that accounts for the propagation of variability from input random parameters to the design outputs. A total of 1,098 passing maneuvers were observed on several two-lane highways in Spain; two data collection methodologies were used: external observations and an instrumented vehicle. The most significant factors affecting PSD were impeding-vehicle speed, passing-vehicle acceleration, and head-ways between impeding and passing vehicles. A uniform acceleration model described the passing-vehicle trajectory. The characterized input parameters and the passing model were used to perform a reliability analysis. The results showed the probability of noncompliance in different scenarios, defined as the proportion of cases in which the required PSD would exceed the available sight distance. American and Spanish PSD standards were evaluated. Geometric design standards presented a probability of noncompliance of about 0.15, whereas some marking standards had a probability of noncompliance exceeding 0.85. These standards may be associated with higher risk levels if they are followed by drivers. As well, PSD risk levels were not consistent for different design speeds, since they underestimated operating speed at some locations