Unveiling the energy implications of automated driving: Evidence from real-world data

This paper examines the impact of automated driving on vehicle energy consumption. Specifically, it quantifies the additional energy demand to operate the automation kit, i.e., the sensing and computation systems which enable a vehicle to drive itself. The study leverages a unique real-world dataset collected from one of the first commercially available Level 3 vehicles, according to the Society of Automotive Engineers (SAE) taxonomy. Energy performance is evaluated across three automation levels: manual driving (no automation), SAE Level 2 system operating, SAE Level 3 system operating. Automation levels are assessed under equivalent operating and kinematic conditions to minimize dynamic-related bias in energy consumption. A methodological framework combining statistical modelling and high-fidelity simulations shows that, for the tested vehicle and analysed driving conditions, activating automated driving systems increases fuel consumption by approximately 10% compared to manual driving. Further analysis confirms that this increase is attributable to the automation kit's energy demand. These findings challenge the common assumption that considers the energy consumption of vehicle automation as negligible, especially in light of the expected energy benefits arising from the distinct driving behaviour of automated vehicles. Furthermore, these outcomes provide novel experimental insights into the ongoing debate on the environmental implications of vehicle automation and suggest that the energy consumption associated with the automation kit should be explicitly accounted for in simulation-based analyses and sustainability assessments.