A new report from the IIHS has some interesting findings.
Last March 18 in Tempe, Arizona, an Uber modified Volvo XC90 SUV with self-driving technology struck and killed a pedestrian crossing a busy roadway at night. In the immediate aftermath of the crash, Uber suspended self-driving car testing on public roads. In the proceeding weeks that followed, we learned that not only was the driver not paying attention to the road (she was streaming an episode of "The Voice" on her smartphone) and could have manually hit the brakes to prevent the collision but also that Uber removed Volvo's collision-avoidance technology. Turns out that was a potentially big mistake.
According to Bloomberg, a new Insurance Institute for Highway Safety report has found that not only is Volvo's safety tech effective, but it likely would have prevented or at the very least reduced the seriousness of the crash. IIHS chief research officer, David Zuby, stated that "I think it's possible that, had the system been able to intervene, the fatality may not have occurred. I would argue that if developers of self-driving technology really intend to make our roads safer, they had better make sure they have the best crash-avoidance systems in place before they go out on the road."
The National Highway Transportation Safety Board also investigated the crash and in its preliminary report stated that the SUV's sensors detected the pedestrian, but Uber further informed those investigators that the automatic emergency braking maneuvers were not activated. Why? To reduce the potential for "erratic vehicle behavior."
In other words, it was the safety driver's job to handle any emergency braking situation, but Uber's system was designed to alert that driver if an object was detected. "Uber decided to forgo a safety net in its quest to teach an unproven computer-control system how to drive," Zuby said in IIHS's report.