Connected Vehicle-based Truck Eco-Driving: A Simulation Study

Abstract

While technologies such as adaptive cruise control and platooning can increase energy efficiency of freight trucks on highways, to date less attention has been given to technologies that can improve energy efficient of these trucks during their first-mile/last-mile operations. In this paper, we present a Connected Vehicle-based truck eco-driving system, called Eco-Drive, which uses Signal Phase and Timing (SPaT) information from the upcoming traffic signal along with the information about the equipped vehicle and preceding traffic to determine the most energy-efficient speed trajectory to pass through the intersection. The system is pre-calibrated using real-world acceleration and deceleration profile data to create a graph-based vehicle trajectory planning algorithm. A machine learning algorithm is then developed and trained to enable a fast and accurate determination of optimal vehicle trajectory. During the operation, the system uses real-time information about the host vehicle, SPaT, and preceding traffic to optimize vehicle trajectory for energy while ensuring a safe passage through the intersection. The system evaluation in traffic microsimulation environment shows that the proposed system provides statistically significant energy savings for the host vehicle, while maintaining similar travel time, as a result of reduced number of stops and milder acceleration/deceleration. The results presented are specific to the simulation settings used in this paper. More research is needed to better understand the levels of energy savings that the proposed system could provide under a wide variety of settings.