Safety Building Blocks of Highly Automated Vehicles

By Zach LaCelle , Christopher Hill, Ph.D.

Autonomous vehicle technology is expanding, but the cutting-edge technologies driving these systems are difficult to qualify as safe and market ready. A new paradigm is needed to ensure that the promises of highly automated vehicles become a reality.

Download Resources

In ground vehicle transportation, the decade of the 2010s has shown the enormous potential of highly automated or autonomous vehicle technology. From initial automated systems such as adaptive cruise control and blind spot monitoring, innovators have expanded to the testing of fully autonomous systems that complete routes with no human input. As many researchers have highlighted, these technologies present opportunities in safety, accessibility, and efficiency for the entire transportation system. An entire new industry focused on these technologies has bloomed, with billions of dollars invested in non-traditional tech companies centered on autonomous vehicle technology.

However, the technology required to accomplish these tasks is cutting edge and difficult to qualify as safe and market-ready. In a system where humans expect highly reliable machines to keep them safe, these new innovations require advancements in evaluation methods. At lower levels of automation, such as Advanced Driving Assistance Systems like adaptive cruise control or lane keeping assistance, human operators have had the task of ensuring safe performance of the system. However, research has shown that this method is flawed for more sophisticated systems; untrained drivers make poor safety operators.

As the technology evolves, relying on a human backstop for safety is insufficient. Furthermore, each year more systems capable of higher levels of autonomy, such as Tesla’s Autopilot or Waymo One, are deployed—and further delay in policy approaches or regulatory frameworks means falling further behind the technology. A new paradigm is needed to ensure that the safety, accessibility, and efficiency gains promised by highly automated vehicles become a reality.

This new paradigm must be both flexible and holistic, recognizing that some fundamental challenges remain unanswered. However, now is the time to engage proactively and effectively to provide a clear and unambiguous set of recommendations, promising practices, requirements, and regulations around autonomous and automated driving systems (ADS). Indeed, recent actions such as the National Highway Traffic Safety Administration’s (NHTSA) advanced notice of proposed rulemaking (ANPRM) regarding safe adoption of ADS underscore a sentiment throughout the industry: Now is the time for a clear safety approach, cognizant of these unique challenges, that will remove environmental and regulatory uncertainty.