The National Highway Traffic Safety Administration (NHTSA) has conducted your first annual report about accidents caused when the driving assistance software included in some vehicles from firms such as Tesla, Honda and Toyota, among others, was in operation. After the reports of the different manufacturers, the administration has concluded that the autopilot of the models of the electric mobility company founded and directed by Elon Musk have been involved in 273 accidents of the 392 related to advanced driver assistance systems since last July.
Tesla vehicles, therefore, accounted for approximately 70% of all accidents in which the Level 2 ADAS (advanced driving assistance systems) system was in operation. The administration has also reported that 5 of the 6 deaths related to these types of accidents — some of them occurred in 2019, but not reported until now — are linked to their vehicles. With the data, the models of Elon Musk’s company represent 60% of the accidents in which one or more people were seriously injured, and approximately 85% in which someone died.
These data call into question the operation of the Level 2 ADAS system that Tesla includes in its vehicles. However, and despite the fact that they may seem worrying, are not conclusive. Although the NHTSA report is based on the information reported by the manufacturers themselves, there are a number of factors that must be taken into account and that could reflect that the Tesla autopilot, in reality, is not less safe than the system of other manufacturers.
NHTSA report does not guarantee that Tesla’s Autopilot is worse than other vehicles’ ADAS
Tesla first It is the only company that receives detailed accident reports electronically. Data that you then share with NHTSA. The rest of the manufacturers are also obliged to report the collisions in which their autopilot systems were active, but the way of collecting this data (which is not done electronically in all cases) could mean that the figures do not reflect all of them. real cases, as Tesla does when using an automated system.
On the other hand, the report does not detail the total number of vehicles analyzed or the distance that they have traveled with the ADAS Level 2 system active. Only in this way can the actual error rate of each system be determined or Tesla’s figures compared with those recorded by other brands.
The NHTSA report, on the other hand, does not specify whether the Tesla vehicle accidents are caused by standard Autopilot or the FSD, a more advanced autopilot system that is currently in beta. According to the regulation, remember, the two systems are considered Level 2, so it is likely that the report groups both FSD and the standard Autopilot.
And why is this relevant? It is not the same as accidents being caused by final software that is available in all Tesla cars (such as the standard Autopilot variant) that by the system Full Self Driving, which is accessed through prior approval of the brand and which is in the development phase. The first is a final system that should work as expected; the second, on the other hand, is a function in development that is not yet open to the general public.
Tesla, in fact, is the only manufacturer that has a system like this in a kind of public beta. Drivers must also accept a series of conditions before registering in this program to access the version beta. They must also ensure that they remain attentive to the road at all times in case the software makes any wrong decisions.
It is also not clear if the accidents have occurred due to the fault of the driving system itself – which has made a wrong decision – or due to an external factor —for example, something that happened on the road. Nor does it detail whether the driver could have avoided it.
What does all this mean? That determining that Tesla’s autonomous driving system is inferior or more dangerous than others solely based on the total number of accidents in which it was active is a mistake. There are many nuances or variables to take into account before making a judgment like that. And the US agency’s report, unfortunately, does not provide a sufficient level of detail to carry out such an analysis.