Energetic Impacts of Autonomous Vehicles in Real-World Traffic Conditions From Nine Open-Source Datasets
Jianchang Huang, Guohua Song, Feng He, Zhe Tan
Abstract
Potential increases or decreases in energy consumption for various traffic operation modes is a research focus in light-duty autonomous vehicle (AV) studies. However, AVs have unclear energy consumption patterns compared with human-driving vehicles (HV), especially under diverse real-world traffic conditions. Considering nine open-source AV activity datasets from six nations, three nontrivial characteristics of light-duty AVs under real traffic conditions were found. 1) The energy consumption of AVs was similar to that of HVs in the low-speed range and was lower than HVs in the high-speed range. 2) The turning point between low-speed and high-speed was a variable parameter for different data sources. 3) The energy consumption of AVs may be higher than that of HVs under complex traffic conditions in the high-speed range. Based on a vehicle-specific power distribution model, this study developed an AV energy consumption model under real traffic conditions at average speeds of 10 to 100 km/h. Potential energy consumption savings and losses were determined using two parameters: the turning point between low-speed and high-speed and the rate of change in vehicle-specific power standard deviation in the high-speed range. Under real traffic conditions, the results implied that light-duty AVs had the potential to reduce energy consumption by 3.78% to 26.47% under stable conditions but could also increase energy consumption by 3.25% to 48.93% compared with light-duty HVs under unstable conditions.