Energy and Environmental Implications of Automated Vehicles Under Mixed Autonomy Traffic Environment

With the evolution of technologies related to automated vehicles (AVs), vehicles with automation are increasingly approaching large-scale deployment nowadays. This research concentrates on mixed traffic scenarios in the near future, which capitalize on the implications of AVs on the transportation system and environment. How to co-exist with AVs in a complex environment is quite an unfamiliar challenge for human drivers of conventional vehicles. To conquer the difficulty, we investigate the behavioral response of human-driven vehicles (HVs) to AVs in multiple mixed traffic situations based on the real-world trajectory dataset and propose the driving characteristic indicator based on machine learning to conduct the quantitative comparison of different carfollowing (CF) behaviors. Additionally, to precisely replicate the driving features of HVs and longitudinal control of AVs in mixed traffic, a microscopic behavior modeling framework consisting of model-based and data-driven CF strategies is designed, which accurately reproduces the sophisticated microscopic behavior and maintains the interpretability of the intrinsic controller mechanism. Furthermore, simulation experiments on a real-world urban-scale network in Suzhou, China are performed to thoroughly evaluate the implications of widespread implementation of AVs under different market penetration rates (MPRs). The result indicates that the low penetration rate may exert adverse effects on the mixed traffic environment.