Evaluation of Level 2 Automated Driving Safety on Curved Sections

Ahn, Seyoung
DOI: https://doi.org/10.1007/s12205-024-2178-x
IF: 2.115
2024-06-11
KSCE Journal of Civil Engineering
Abstract:Automated vehicles minimize driver errors and can help reduce traffic accidents. If the automated driving system encounters an unexpected problem, control is transferred to the driver and driving safety is significantly impacted. 63.3% of disengagements caused by road structures are due to recognition errors on curves. However, the criteria for curves are not clear, and there is a lack of research on automated driving disengagement and curve radius. This study attempted to analyze the correlation between driving speed and curve radius, which affects the disengagement of Level 2 automated driving on curved sections and thus is a risk factor for automated driving. By analyzing data from field driving tests, we evaluated the safety of Level 2 automated driving along curves and compared the curve driving performance of automated driving systems applied to different car models. As a result, the disengagement rate varies with the curvature, and for a smaller curve radius, it is more difficult for the Level 2 automated driving system to maneuver to take control and drive safely. The driving speed in the curve has the largest impact on the disengagement of the ADAS. Hence, the driving speed in curves must be managed to ensure the safety of Level 2 automated driving systems.
engineering, civil
What problem does this paper attempt to address?