Capacity and delay analysis of arterials with mixed autonomous and human-driven vehicles

Abstract

This paper investigates the traffic flow characteristics of mixed stream of autonomous and human-driven vehicles. The proposed model aims at understanding the fundamental properties of mixed flow of human-driven (N) and autonomous vehicles (AV) such as headway, capacity, and delay at signalized intersections. This is challenging because of intrinsic differences between longitudinal driving characteristics of these two types of vehicles and the convoluted dynamics of car following situation within various combinations of AV and human-driven vehicles. The expected headway of the mixed flow is determined based on the penetration rate of AV and the headways between two successive AV-AV, AV-N, N-AV, and N-N. Furthermore, the upper and lower bounds of mixed flow headway is presented. The theoretical headway is validated by microsimulation data. The estimated headways are then incorporated to derive the delay of a mixed flow at a signalized 2-lane link. Four combination of (i) mixed lanes, (ii) dedicated lanes for AV and human-driven vehicles, (iii) one mixed lane and one AV dedicated lane, and (iv) one mixed lane and one human-driven vehicle dedicated lane are considered. The results demonstrate the performance of the four lane configurations for various stages of AV deployment penetration rate.