Cruise has teamed up with The University of Michigan Transportation Research Institute (UMTRI) to present a pivotal study to establish a safety benchmark for human driving. This collaboration involved industry stalwarts General Motors and the Virginia Tech Transportation Institute (VTTI).
Why It Matters
As autonomous vehicles (AVs) edge closer to widespread deployment, there’s an increasing need for precise safety performance metrics. Such benchmarks offer a standardized measure against which autonomous vehicles can be assessed, ensuring they not only meet but exceed the safety performance of human drivers.
- The research was centered on ridehail drivers in San Francisco over a span of two years.
- From the study, the following statistics emerged for Cruise’s autonomous vehicles when compared to human ridehail drivers:
- 65% reduction in overall collisions
- 94% fewer accidents where they were the primary contributors
- 74% less chance of accidents leading to significant injury risk
- The study considered a whopping 5.6 million miles of human ridehail driving data in San Francisco, with an exclusion for high-speed roads (those with posted speeds over 35 mph).
- This research combined data from UMTRI’s extensive fleet study and VTTI’s precision instrument study, allowing a comprehensive understanding of the human ridehail crash rate.
Dr. Carol Flannagan, the paper’s lead author and a research professor at UMTRI, commented on the importance of the study. She emphasized the gap in autonomous vehicle research for benchmarks mirroring actual human driving. Flannagan noted, “We delved into the real-world driving patterns of ridehail drivers in intricate urban settings, enabling a significant benchmark for analogous human driver performance.”
Louise Zhang, Cruise’s Vice President of Safety & Systems, highlighted this research’s insights into human ridehail driving behaviors. She believes this is a significant stride toward the ultimate goal of safer roads with the large-scale introduction of safe self-driving vehicles.
This collaborative research provides a crucial benchmark against which autonomous vehicles can be gauged, laying a foundation for ensuring AVs’ safety is beyond par. As self-driving technology gains traction, such benchmarks will be invaluable in shaping a safer future for road users.