Examining the Performance of Automatic Emergency Braking Systems

Automated vehicles are expected to significantly reduce traffic crashes and the resultant injuries and fatalities. Although there is ambiguity as to the timeline for market-ready fully automated vehicles, lower levels of automation have already demonstrated some of this significant safety potential. This includes technologies such as automatic emergency braking (AEB), which will be a mandatory feature in all new vehicles from September 2029. This study involved an evaluation of AEB test data from the Insurance Institute for Highway Safety. These tests covered various scenarios, including those in which the test vehicle encounters a balloon car, as well as tests involving “dummy” pedestrians that were walking either parallel or perpendicular to the road. These tests were conducted under various speeds and lighting conditions. The test vehicles ranged from model year 2013 to 2025 and included a diverse range of sensor configurations. A series of random-effects logistic regression models were estimated to evaluate the efficacy of these vehicles across the test scenarios. The results showed that AEB performance has improved considerably over time, reflecting improvements in the underlying sensor technology. However, performance was significantly worse at higher testing speeds, under nighttime conditions, and in other scenarios that represented greater challenges for the sensing technology. This study provides important insights as to the potential and limitations of these systems in their current form.