The Environmental Impact of Connected and Automated Vehicles’ Car-following Behavior

Abstract

This thesis focuses on the environmental impact of car-following (CF) driving behavior in mixed autonomy traffic. Given the shortage of real-world mixed autonomy trajectory data, this area of research is not yet well-understood and is often explored using classical simulation models with assumed and unvalidated parameters with little or no consideration of autonomous eco-driving behavior. The thesis introduces a new calibration framework applicable to various CF models. The framework uses the concept of ‘adaptability’ which improves the simulation of AVs’ driving behavior by enabling real-time changes of the CF model parameters based on prevailing traffic conditions. The thesis also provides an environmental assessment of mixed autonomy traffic considering different vehicle arrangements in a platoon using real data and simulation outcomes. The thesis provides several novel findings regarding the impact of the driving behavior of AVs on-road emissions. For instance, the analysis of the trade-off between traffic stability and mobility in a mixed autonomy environment reveals that an automated eco-driving strategy that minimizes the reciprocal of the platoon average velocity produces fewer emissions compared to strategies that aim to minimize traffic instability measured by the standard deviation of velocity or average acceleration. This is because the latter can considerably decrease the platoon average velocity to increase stability. In addition, vehicle arrangements, specifically whether the AV is leading or following, affect the AV’s driving behavior and thus, the emissions, especially in congested conditions. Furthermore, the correlation between mobility and emissions shows that traffic emissions are highly influenced by acceleration at a high-velocity level and by time headway at a low-velocity level. Overall, the thesis explores the environmental implications of AV driving behavior considering different aspects such as traffic network, market penetration rate, and vehicle arrangements. Therefore, the findings of this thesis can be insightful at this early stage of AV deployment to ensure that the future mixed autonomy environment does not only improve traffic and safety but also contribute to a more environmentally friendly transport system.