Surprisingly, the National Highway Traffic Safety Administration (NHTSA) recently began collecting some extensive data on vehicle crashes involving SAE Level 2 Advanced Driver Assistance Systems (ADAS) – driver assistance technology such as cruise control, lane keeping and automatic lane change. The first report on the collected data was published last month. Immediately, a rather shocking figure was pulled from it: “Tesla cars accounted for about 70 percent of the 392 accidents involving advanced driver assistance systems reported since last July, and most of the deaths and serious injuries — some of which are in the past. for more than a year.”
Immediately, this number made no sense to me. However, instead of seeing any real criticism of the result, I saw occasional references to bash Tesla Autopilot and FSD Beta. I have Tesla Autopilot and FSD Beta, and I know they’re not perfect, but they’re as good as they get for current ADAS. Think that more despite the fact that accidents occur due to Tesla ADAS more It defies basic logic that non-low-Tesla ADAS systems are on the road. Yes, you might assume that Tesla drivers expect more from their ADAS and drive more recklessly, causing more accidents, but 70% of all ADAS-related crashes in the US? Really? Of course not.
No, there’s something else playing with this data, and NHTSA actually discloses it in various disclaimers – apparently no one wanted to pay attention or pay much attention to the disclaimers. The bottom line and the simplest: you only have the data you have. This is the assumption that NHTSA collects and records all ADAS-related crashes road off the mark. But let’s turn to the agency’s own language — which was included at the beginning of its first report on the subject.
NHTSA Information on the 1st Report of ADAS-related crashes
Here is the first line of the report: “In June 2021, the National Highway Traffic Safety Administration (NHTSA) issued a Permanent General Order (General Order) requiring specified manufacturers and operators (reporting agencies) to report certain vehicle accidents to the agency. gave. Equipped with SAE Level 2 Advanced Driver Assistance Systems (ADAS).” In other words, collected crash data is data Car manufacturers have released information about this. One can assume that some automakers have ulterior motives for hiding such crash data and not reporting their numbers to NHTSA, but I think the problem is simpler and less ugly than that. Apparently NHTSA does too.
First, the total number of ADAS-related crashes in about a year from automakers: “As of May 15, 2022, NHTSA has received 392 incident reports involving vehicles equipped with Level 2 ADAS.” That’s right – 392 accidents in total, 70% of which were reported by Tesla. that’s it a total of 119 accidents reported by other automakers. If you think there have been a total of 119 accidents involving non-Tesla ADAS in almost a year, I have a bridge to sell you.
NHTSA goes on to state that it never learned about such accidents until the aforementioned General Order (which is actually kind of surprising). This is exactly what NHTSA wrote, but the most interesting and relevant points are not until the next section (“Information and Limitations”): “Prior to the implementation of the General Order, NHTSA’s sources of timely accident notifications were limited (e.g., media, vehicle owner questionnaire , direct contact with a voluntary company and Early Warning Report) and are generally inconsistent between manufacturers, including developers.
“The General Order allows NHTSA to receive timely and transparent notification from manufacturers and operators of real-world accidents involving vehicles equipped with Level 2 ADAS. With this information, NHTSA can respond to accidents that raise safety concerns about ADAS technologies through additional investigation and enforcement. As an example, the Executive Order informed the agency about specific accidents involving ADAS-equipped vehicles. NHTSA contacted several manufacturers to learn more about these crashes, and that information informed the agency’s investigation into the driver assistance system in late 2021. The information also provides transparency into the safety of vehicles equipped with this technology. currently operating on public roads. It also provides important information necessary for research and the development of policies to improve the safety of these technologies. Safety is critical to the deployment and public acceptance of new technologies, and the General Order is an important step in assessing the safety of ADAS and ADS technology.”
It all sounds fine and dandy if you’re going to get really comprehensive information from automakers. The problem is stated somewhat subtly, but clearly enough, in the next section of the NHTSA report, titled “Information and Limitations.”
NHTSA Information Limitations on ADAS-Related Crashes
The first line of this section tells the reader that really, really, seriously consider these restrictions before drawing any conclusions: “Understanding the terms and requirements of the General Order, the information it is intended to obtain, and the limitations of that information is critical to accurate interpretation and analysis. When considering this and other information, the following should be considered.” Then this section has 5 subsections:
- Access to Accident Data May Affect the Accident Report
- Incident Report Information May Be Incomplete or Unverified
- Redacted Confidential Business Information and Personally Identifiable Information
- There may be multiple reports of the same accident
- Summary Event Report Data Not Normalized
You can get the gist of it from these subheadings alone, but the short and sweet side of it is that automakers are often not equipped to collect data on ADAS-related crashes and therefore cannot provide NHTSA with the data it requests. Without most of the industry data, it is of course impossible to draw conclusions comparing different ADAS systems or car manufacturers.
Photo: Zach Shahan | CleanTechnica.
Here are some quotes from NHTSA from the first subsection of this section (emphasis added):
- “Crash data recording and telemetry capabilities can vary greatly by manufacturer and by driver automation system.”
- “There may be many Level 2 ADAS-equipped vehicles data recording capabilities are limited related to driving automation system activation and crash conditions.
- “The car the ability to remotely transmit this information to the manufacturer may also vary widely for notification purposes.
- “Furthermore, vehicles equipped with Level 2 ADAS are generally privately owned; consequently, in the event of a reportable accident, manufacturers may not be aware of this unless they contact the car owner.”
- “Manufacturers of vehicles equipped with Level 2 ADAS, which have limited data recording and telemetry capabilities, can only receive consumer reports of driver automation system involvement in a crash, and if the manufacturer is notified, there may be a delay. all.”
- “Furthermore, some crashes involving Level 2 ADAS-equipped vehicles with limited telematics capabilities cannot be included in the General Order unless the consumer reports that the automation system was activated within 30 seconds of the crash or there is no other available information. indicates a Level 2 ADAS designation due to limited information obtained from the crashed vehicle.
- “In contrast, some manufacturers have access to greater amounts of accident data immediately after an accident due to advanced data logging and telemetry.” (Read: Tesla)
It’s likely that most of this is NHTSA rewriting what automakers have said about why they don’t really have any information to hand over to NHTSA. Here’s my hypothetical example of the lawyers or executives of the automakers telling NHTSA: “Well, we don’t have real data on the road about this because our cars aren’t equipped to collect that kind of data, and certainly aren’t equipped to send it automatically. come back to us. “The only way we know if an advanced driver assistance system (ADAS) was used at the time of the accident is if the vehicle owner tells us.”
It’s great that NHTSA is now working to collect data on this issue and investigate the safety of various systems, but right now the agency simply doesn’t have the data. Car manufacturers don’t have the information and therefore can’t send it.
Perhaps as cars get smarter and smarter and automakers incorporate more and more advanced ADAS, the data will begin to flow and NHTSA will have a useful job benchmarking and improving the systems. On the other hand, with this research and reporting system in place now, automakers can also improve their internal data collection and avoid risking regulatory or legal problems.
The NHTSA report has more. Check it out to investigate further, but at least don’t jump to conclusions about the grossly incomplete “US ADAS crash data”.
Do you value CleanTechnica’s originality and clean tech news? Consider becoming a CleanTechnica Member, Supporter, Tech or Ambassador, or a Patron on Patreon.