Over the past several decades, the development of technologies and the production of autonomous vehicles have enhanced the need for intelligent intersection management systems. Subsequently, growing interest in studying the traffic management of autonomous vehicles at intersections has been evident, which indicates a critical need to conduct a systematic literature review on this topic. This paper offers a systematic review of the proposed methodologies for intelligent intersection management systems and presents the remaining research gaps and possible future research approaches. We consider both pure autonomous vehicle traffic and mixed traffic at four-way signalized and unsignalized intersection(s). We searched for articles published from 2008 to 2019, and identified 105 primary studies. We applied the thematic analysis method to analyze the extracted data, which led to the identification of four main classes of methodologies, namely rule-based, optimization, hybrid, and machine learning methods. We also compared how well the methods satisfy their goals, namely efficiency, safety, ecology, and passenger comfort. This analysis allowed us to determine the primary challenges of the presented methodologies and propose new approaches in this area.INDEX TERMS Autonomous vehicle, intelligent intersection management system, mixed traffic, vehicleto-infrastructure (V2I) communication, vehicle-to-vehicle (V2V) communication.

To address sustainable development issues of urban traffic, electric buses will join traditional bus system, and the scheduling of bus fleet should be adjusted due to the distinct features of electric buses. To this end, this paper develops a Multi-objective Bi-level programming model to collaboratively optimize the vehicle scheduling and charging scheduling of the mixed bus fleet under the operating conditions of a single depot. The upper level determines the vehicle scheduling to minimize the operating cost and carbon emissions under the constraints of connecting time between trips and the limited driving range of electric buses. The lower level is a charging scheduling problem that considers the charging time and the limited driving distance constraint to minimize the charging cost. The proposed model is solved with an integrated heuristic algorithm. The vehicle scheduling problem is addressed with the iterative neighborhood search algorithm based on simulated annealing, while the charging scheduling problem is solved with a greedy dynamic selection strategy based on the approach of multi-stage decision. Finally, case study is carried out based on a mixed bus fleet in Beijing, and the results validate the availability of the proposed model and solution algorithm. INDEX TERMS Vehicle scheduling, charging scheduling, mixed bus system, multi-objective bi-level programming, collaborative optimization.

This paper presented an energy-efficient adaptive cruise control, called Energy-Efficient Electric Driving Model (E 3 DM), for electric, connected, and autonomous vehicles (e-CAVs) in a mixed traffic stream. E 3 DM is able to maintain high energy efficiency of regenerative braking by adjusting the spacing between the leading and the following vehicles. Moreover, a power-based energy consumption model is proposed to estimate the on-road energy consumption for battery electric vehicles, considering the impact of ambient temperature on auxiliary load. Using the proposed energy consumption model, the impact of E 3 DM on vehicle energy consumption is investigated. In particular, single-lane vehicle dynamics in a traffic stream with a mixed of e-CAVs and human-driven vehicles are simulated. The result shows that E 3 DM outperforms existing adaptive cruise control (i.e. Nissan-ACC) and cooperative adaptive cruise control (i.e. Enhanced-IDM and Van Arem Model) strategies in terms of energy consumption. Moreover, higher market penetration of e-CAVs may not result in better energy efficiency of the entire fleet. The reason is that more e-CAVs in the traffic stream results in faster string stabilization which decreases the regenerative energy. Considering mix traffic streams with battery electric (BEVs) and internal-combustion engine (ICEVs) vehicles, the energy consumption of entire fleet reduces when the market penetration of BEV (contains both e-CAV and human-driven BEV) increases. A higher ratio of e-CAV to human-driven BEV results in higher energy efficiency.

Connected and autonomous vehicle (CAV) technologies are likely to be gradually implemented over time. In this paper, an adaptive cruise control, named Smart Driver Model (SDM), is proposed to describe the autonomous vehicles flow. The stability criteria is proposed for SDM to judge the stability of homogeneous traffic flow. Numerical simulations were conducted to verify the results of the theoretical analysis. Single-lane vehicle dynamics in a traffic stream with connected and autonomous vehicles are simulated by varying model parameters. Simulation results are consistent with the results of linear stability analysis. As a result, a set of parameters is proposed to investigate the stabilization effect of the proposed model on homogeneous traffic flow considering realistic driving cycle and cut-in condition. By simulating a platoon with a lead vehicle which follows the Urban Dynamometer Driving Schedule (UDDS), we find out that the proposed model can stabilize the traffic flow with proposed parameters. The results from simulation and linear stability analysis show that SDM outperforms the IDM-ACC and the ACC proposed by Milanés and Shladover in terms of stabilization effect on homogeneous traffic flow. The simulation result shows that the SDMequipped vehicles are able to stabilize the homogeneous traffic flow under cut-in condition.

Travel time and its reliability are intuitive system performance measures for freeway traffic operations. This paper proposes a method to estimate travel times based on data collected from roadside radar sensors, considering spatially correlated traffic conditions. Link-level and corridor-level travel time distributions are estimated using these travel time estimates and compared with the ones estimated based on probe vehicle data. The maximum likelihood estimation is used to estimate the parameters of Weibull, gamma, normal, and lognormal distributions. According to the log likelihood values, lognormal distribution is the best fit among all the tested distributions. Corridor-level travel time reliability measures are extracted from the travel time distributions. The proposed travel time estimation model can well capture the temporal pattern of travel time and its distribution.