NHTSA Begins to Study Automated Driving Crashes

After 4 years with an administration that created a more lax regulatory environment, the National Highway Traffic Safety Administration (NHTSA) is starting to study automated driving crashes. NHTSA is part of the U.S. Department of Transportation, which is responsible for setting and enforcing a wide range of vehicle safety regulations that cover everything from occupant protection to lighting systems. After sitting on the sidelines as automakers and other companies have developed and deployed a range of driving automation features, NHTSA now wants to determine if these systems are causing or contributing to crashes. 

Increased Pressure Prompts NHTSA to Act

The agency’s largely hands-off stance on automated driving systems (ADS) dates back to 2016 when, under the Obama administration, it issued the first ADS guidance document with voluntary recommendations for ADS developers. While 2016 was probably too soon to start active development of ADS regulations, the agency could have mandated data sharing to help it better understand how the technology was developing and to consider how performance standards might be drafted. Instead, under former Secretary of Transportation Elaine Chao during the Trump administration, the voluntary guidance was reduced further. 

In the years since Joshua Brown’s fatal accident while using the first generation of Tesla’s Autopilot in a Model S, there have been numerous similar crashes that involved the advanced driver assist system (ADAS). There have been at least three investigations by the National Transportation Safety Board (NTSB), and NHTSA recently acknowledged 30 investigations, many of which are still open. 

NTSB made numerous recommendations that NHTSA did not act on. These recommendations include mandating active driver monitor systems, ensuring that ADAS and ADS are geofenced so that they can only be used within their operational design domain, requiring more robust data logging, and requiring greater oversight of these systems. Earlier this year, NTSB, which is an investigatory body with no regulatory or enforcement authority, sent a letter to NHTSA, criticizing the agency for failing to act.

New Data Sharing Requirements Promote Safety and Accountability

With numerous automakers planning to launch dozens of vehicles with hands-free, partially automated capabilities over the next few years and higher level automation on the way, NHTSA has decided to act. The agency has issued a general order to all automakers and ADS developers that they must share data related to all crashes that involve Level 2 or higher ADAS and ADS.

Level 2 systems combine lane centering and automatic speed control functionality but require the driver to remain attentive, watching the road and ready to take control at any time. Level 3 eliminates the requirement to watch the road and supervise but still requires a human to take over within 10 seconds, and Level 4 requires no human intervention but is limited to operation within constraints, such as location or weather. 

Under the new order, companies will have to submit a preliminary report to NHTSA within 24 hours of learning of a crash that occurs while the ADAS or ADS is active or if it was active within 30 seconds prior to the crash. This last part is important, because there have been many crashes where drivers claimed they were using Tesla Autopilot but the automaker said the data showed they were not. However, Tesla has never revealed when Autopilot has disengaged before a crash. 

The agency should learn a lot about the real efficacy of these systems from this data and it hopefully will lead to both improved safety and manufacturers being more honest about the true capabilities of ADAS and ADS technology. NHTSA is responsible for road safety, and automated driving is becoming another facet in the changing road ecosystem.