Field effectiveness evaluation of advanced driver assistance systems

Area of Science:

• Automotive safety engineering
• Transportation safety research
• Human-computer interaction in vehicles

Background:

• Advanced driver assistance systems (ADAS) are crucial for enhancing vehicle safety through warnings and interventions.
• Real-world effectiveness data for ADAS is limited, despite promising laboratory results.
• This study focuses on new-model BMWs in the U.S. to assess ADAS impact.

Purpose of the Study:

• To evaluate and quantify the association between ADAS technologies and the likelihood of moderate or severe crashes.
• To provide real-world evidence on the effectiveness of specific ADAS features.
• To analyze crash reduction associated with ADAS in a large cohort of BMW vehicles.

Main Methods:

• A cohort of over 1 million model year 2014+ BMWs was analyzed using VIN-identified options data.
• ADAS features studied include frontal collision warning with autonomous emergency braking, lane departure warning, and blind spot detection.
• Cox proportional hazards regression modeled crash risk, utilizing BMW Automated Crash Notification data.

Main Results:

• Vehicles with autonomous emergency braking and lane departure warning were 23% less likely to crash (HR=0.77).
• Blind spot detection showed a 14% reduction in crashes (HR=0.86) when controlling for other ADAS.
• The effectiveness of combined AEB and LDW increased with vehicle model year, showing a 34% reduction for 2017 models.

Conclusions:

• This study provides robust evidence supporting the effectiveness of ADAS in reducing real-world crashes.
• The findings contribute to the growing body of evidence on ADAS safety benefits.
• Specific ADAS combinations demonstrate significant crash risk reduction, particularly in newer vehicles.