At the stroke of midnight, as April Fools’ Day turns to April 2, California’s long awaited regulations for the testing and deployment of highly automated vehicles (“HAVs”) without human operators will go live. Though the regulations are over three years late, their implementation is no laughing matter when it comes to the future of HAVs in California and throughout the nation.
That’s because, with the arrival of April, “disengagement reports,” the widely recognized and poorly understood indicators of the technology’s progress made to the California DMV by HAV developers when their vehicles’ automated systems yield control to human operators, will be rendered meaningless. By the letter of California’s law, no “manual” driver means no disengagements to report.
The upside, of course, is that we will be spared the intermittent freakouts now routine when the California DMV issues its updates. The downside, sadly, is that the end of disengagement reports will offer critics an opportunity to call for new and imperfect state-based metrics by which to evaluate the progress of self-driving technologies.
Those calls should be resisted for a bevy of reasons, not least of which is that such regulations would be promulgated by regulators ill-suited to the task and, in practice, would be susceptible to manipulation in the field. But, more importantly, those calls should be resisted because a national framework for tracking progress is already quietly taking shape.
Last year, the National Highway Traffic Safety Administration (“NHTSA”), in addition to a number of other federal regulatory entities and the Governors Highway Safety Administration (“GHSA”), together promulgated the 5th Edition of their Model Minimum Uniform Crash Criteria “(MMUCC”). This document, voluntarily adopted by states, provides a uniform basis by which traffic records are collected and maintained. In doing so, the idea is that NHTSA and other vehicle safety regulators are able to assess the safety of vehicles and vehicle technology on the road using high-quality crash data. Ultimately, NHTSA uses the data generated by the MMUCC to protect the public.
While not mandatory in its entirety (meaning that states vary in their fidelity to the MMUCC’s standards), the guidance and its application are made palatable for states thanks to funded implementation assistance from NHTSA. This incentivizes states to adhere to the MMUCC, and that, in spite of its voluntary nature, provides the closest thing there is to a uniform data-set of crash information across the country.
In contemplation of the arrival of HAVs, the latest MMUCC includes explicit guidance for gathering information about crashes involving HAVs (p.121-124). And even provides a model crash report to collect information on HAVs involved in a crash (p.194). Usefully, the crash data generated by the latest MMUCC, and later analyzed by NHTSA, will provide HAV-watchers with a comprehensive basis by which to track the development of the technology.
In its current form, a MMUCC 5th Edition crash report involving a HAV would include standard information about the type of vehicle involved, the crash location, contributing circumstances, severity of injury and/or fatalities suffered, as well as uniform information about automation. Specifically, whether the vehicle was equipped with an automation system, what level of automation was involved, and what level was engaged at the time of the crash.
Thus, unlike California’s disengagement reports or any subsequently adopted state-specific requirements focusing on the operation of automated technologies in lieu of their physical safety performance, the information gathered via MMUCC 5th Edition crash reports will offer a ready point of comparison for analysis of not only the technology as a whole, but specific manufacturers and their specific models. Ultimately, this type of information will allow the public to gain a clear understanding of how self-driving technologies are maturing.
In light of NHTSA’s development and approval of the MMUCC’s 5th Edition, which is suggestive of the data that it is actually seeking for HAV safety analysis and rule-making, it behooves jurisdictions across the country to adopt the uniform crash reports offered therein.
Lamentably, gathering that crash data to quell the appetite of skeptics keen to sow concern may prove just as vital to the ultimate success of highly automated vehicles as overcoming the technical challenges that developers are in a race to surmount. With the imminent and welcome demise of disengagement reports, that’s a bad April Fools’ joke in itself.