Evaluating General Motors Super Cruise system field effects in the context of strategic system use

Emerging Level 2 (L2) systems combine Adaptive Cruise Control and lane centering functionality. Evaluating the safety effects of L2 systems is currently more challenging than quantifying the effectiveness of Advanced Driver Assistance System (ADAS) features due to lower crash samples, limited system availability (resulting from Operational Design Domain constraints), selective system use, and overlapping ADAS content with proven safety benefits. This paper aims to address these challenges before the accumulation of further crash data. The General Motors Super Cruise system evaluated allows hands-free driving on GPS-mapped, system-compatible roads. It employs a camera-based driver monitoring system, along with a series of escalating alerts, to prompt the driver to pay close attention to the road ahead and take steering control when takeover requests are issued.

By matching feature content data from Super Cruise-equipped and matched comparison vehicles to police-reported crashes from 10 states, 4,238 analysis-relevant crashes were identified. Three complementary analyses were conducted to account for crash exposure, the appropriateness for small samples, and the ability to account for potentially confounding factors. One analysis used telematics data to understand actual Super Cruise use surrounding crashes.

In two analyses, Super Cruise-equipped vehicles reduced single-vehicle road departure crashes. However, these analyses did not account for system use, and the observed benefit sizes suggest that drivers would need to use the system highly strategically to achieve these reductions, and that additional crash data are needed to stabilize these effects. Regarding potential unintended L2 effects, results across analyses indicated no increases in lane departure, rear-end striking, lane change, or rear-end struck (control) crashes for Super Cruise-equipped vehicles relative to comparison vehicles.

The analysis approach used here should generally apply to L2 safety evaluations. Additional telematics data could be used to further understand system use conditions and clarify the addressable crash population.