NHTSA Revisits the 5-Second Rule

We have written about Event Data Recorders (“EDRs”) and their limitations before, noting “most significantly, an EDR only records data in a crash (or crash-like) event, and even then provides only a snapshot of the seconds leading up to the event.” Earlier this year, NHTSA published a proposal to address this limitation by amending its regulations regarding EDRs. Specifically, the amendment would “extend the EDR recording period for timed data metrics from 5 seconds of pre-crash data at a frequency of 2 Hz to 20 seconds of pre-crash data at a frequency of 10 Hz (i.e., increase from 2 samples per second to 10 samples per second).”

In requesting public comment on the proposed amendment, NHTSA explained that the existing 5-second recording duration (for speed, engine throttle, service brake, engine RPM, ABS activity, stability control, and steering input) was selected when regulations were promulgated in 2012 “because the agency concluded that it would be long enough to ensure the usefulness of the data in crash reconstruction while also minimizing the risk that the data capture process would over-tax the EDR’s microprocessor, which could cause a malfunction that could lead to a loss of data.”  Three years later, the Fixing America’s Surface Transportation (FAST) Act of 2015 required NHTSA to consider whether 5 seconds is sufficient “to investigate the cause of motor vehicle crashes.” Without spending time on the extraordinary amount of time the rulemaking process can take in this country, NHTSA’s recently-proposed amendment is the culmination of its efforts in response to that 2015 mandate.

According to NHTSA’s evaluation (in conjunction with Virginia Tech), 5 seconds of data turns out to be inadequate for determining the cause of a crash—largely because it fails to capture pre-crash maneuvers—as compared to 20 seconds, as reflected in the tables below.

This is interesting. More interesting to us, however, is that in evaluating EDR recording time for causation analysis, NHTSA also found that “a better understanding of the driver’s pre-crash behavior will also assist in the evaluation of emerging crash avoidance systems (e.g., lane departure warning, lane keeping assist, forward collision avoidance, automatic emergency braking, and intersection safety assistance systems).” While undoubtedly true, NHTSA’s continued focus on EDR data alone actually ignores an entire category of data that would be even more useful in assessing these advanced crash avoidance systems as well as accident causation—i.e., Vehicle Performance Data (VPD).

We have discussed the value of VPD here and here, and were encouraged to see that a number of significant commenters made similar observations.* To name a few:

• The National Association of Mutual Insurance Companies, while generally supportive of the amendments, urged NHTSA to “consider proposing more wide and extensive auto safety data recording and reporting,” to take into account the fact that “many vehicle manufacturers collect literally thousands of data elements for their vehicles hourly about the driver actions and about the vehicle performance and condition” and because “[t]here is a universe of relevant safety data being recorded in these vehicles and at the manufacturer that could increase safety that remains outside the collection and reporting of EDRs.”

• The National Transportation Safety Board expressed concern that the proposed amendment is not “sufficient to enable full understanding of crashes involving vehicles with advanced driver assistance systems (ADAS), particularly those with SAE International (SAE) Level 2 (L2) capabilities,” and reiterated its prior recommendations that NHTSA require and standardize ADAS data collection for causation analysis and improved safety.

• Advocates for Highway & Auto Safety remarked that the data elements captured by EDRs “continue to be insufficient to capture the performance of advanced driver assistance systems (ADAS). . . . Many vehicles on the road today are already producing voluminous data, and the amount and type of data will only increase as the technologies evolve. Many stakeholders need that data for numerous and varied important reasons including safety. . . .  Making additional data about the on-road performance of vehicles with ADAS more widely available will increase understanding of the capabilities and limitations of ADAS and contribute to consumer confidence.”

• State Farm expressed support for NHTSA’s proposed amendment, but also advocates for an “ADS EDR,” that would register “at a minimum: • Individual sensor data (static and moving obstacles) • Forward looking and in-cabin views • Sensor fusion outputs (obstacle classification) • Localization – where is the vehicle in terms of latitude and longitude • Driving context – at intersection, lane change, speed, etc. • Decision-making – traffic light status, slowing down • Current and planned path and speed profiles • Health status of hardware and software tasks • V2X communications.”

• General Motors opposed the proposed amendment to increase recording duration, in part because “EDR data is only one piece of information used in the complex analysis necessary to determine crash causation. This data, in and of itself, cannot provide the answer.” Instead, GM said that EDR data must be considered in conjunction with other elements including “other vehicle data, such as ADAS event data.”

• Honda wrote that NHTSA’s approach “inherently focuses crash causation on the driver actions, and vitally misses valuable data that could drive improvements for other driver assistive safety technologies. There is far greater value that can be obtained from an EDR by incorporating crash avoidance data elements, in contrast to the [amendment] proposal to extend the time duration and sampling frequency of the current data elements.”

Perhaps, however, The Center for Auto Safety (CAS) best summed up the issue, stating:

The [proposed amendment], while providing slightly more useful data for crash investigators, appears to be unnecessarily constrained by the perception that traditional parameters are sufficient to respond to the needs of modern and evolving ADAS and ADS-equipped vehicle crash investigation and root cause analysis. That approach is not adequate. While CAS agrees that an update of EDR requirements is timely, NHTSA should require EDR collection and recording of data that enables post-crash assessment of safety critical sensor, data processing, OEDR, vehicle control, motion, and gross performance data for the broad scope of technology that is allowed to be used in current and anticipated production vehicles, notably those equipped with ADAS or ADS capabilities, as recommended by the NTSB. NHTSA should require that those EDR data are provided to investigators unfettered by proprietary interests of manufacturers.

* Our affiliated company, QuantivRisk, submitted its own comment regarding the proposed amendment and value of VPD. It can be accessed here.

Copyright Nelson Niehaus LLC

The opinions expressed in this blog are those of the author(s) and do not necessarily reflect the views of the Firm, its clients, or any of its or their respective affiliates. This blog post is for general information purposes and is not intended to be and should not be taken as legal advice.