As the analysis of traffic accidents in the USA by the United States Insurance Institute for Highway Safety (IIHS) showed, self-driving cars that developers advertise as a way to get rid of road incidents are likely to prevent only a third of all accidents.
According to the IIHS study, the remaining two-thirds of crashes were caused by errors that autonomous driving systems are unable to handle better than drivers. Traffic experts say that approximately nine out of 10 accidents result from human error. Last year, about 40 thousand people died in the USA as a result of car accidents.
Companies involved in the development of self-driving cars are positioning fully automated driving as a tool to significantly reduce road deaths, eliminating the human driver from the equation. But the IIHS study provided a more detailed picture of driver errors, showing that not all errors can be eliminated with a camera, radar, or other sensor-based autonomous driving technologies.
During the study, IIHS specialists analyzed more than 5,000 typical accidents throughout the country, recorded in police reports, and identified factors related to human errors that contributed to the accident. Only one third of all accidents was an exceptional result of control and perception errors or limitation of driver’s legal capacity.
But most crashes were the result of more complex errors, including an incorrect assessment of the possible maneuvers of other road users, movement too fast or too slow for road conditions, or incorrect evasion maneuvers. Many accidents occurred as a result of a combination of several errors.
“Our goal was to show that if you do not solve these problems, autonomous cars will not provide significant safety benefits.“Said Jessica Cicchino, IIHS Research Vice President and Research Co-author.
If you notice an error, select it with the mouse and press CTRL + ENTER.