Impacts of Automated Vehicles in a Mixed Environment on Intersection Sight Distance at Uncontrolled Intersections

Abstract

Mixed vehicle traffic of driver-operated vehicles (DVs) and autonomous vehicles (AVs) on roadways in the future will result in four interaction types between these two vehicle types, impacting current intersection sight distance (ISD) designs. This investigation assessed the ISD for uncontrolled intersections, accounting for each interaction type. ISD design values for DVs at uncontrolled intersections are based on a model adopted in current design guidelines that accounts for human driver behaviors. For AVs, a stopping sight distance model is proposed for ISD. These models and other parameters formed the system demand and supply models relative to an object location, with ISD noncompliance occurring when the demand is greater than the supply. The probability of unresolved conflict (PUC) measure was developed by combining the results of a Poisson-based conflict estimation model and the probability of noncompliance (PNC) of the vehicle interactions obtained by Monte Carlo simulation (MCS). In mixed traffic conditions, the PUC values for a specific object location depend on the traffic volume and speed combinations, although some combinations produce similar PUC values, AV penetration rates, and traffic volume combinations. Target PUC values that can be used to establish design guidelines under mixed traffic conditions were deduced from the results involving a DV-only traffic environment. An example application is presented to demonstrate the application of the developed methodology and models.