The idea of autonomous vehicles is based on both the convenience and the potential to eliminate hazardous, fatal, and costly human errors. According to the National Highway Traffic Safety Administration, the majority of vehicular accidents, more than 90 percent, involve man-related errors.
Road tests of these self-driving cars are already being conducted and their usage and deployment are being encouraged through updates to current regulations and a variety of new regulations. What remains to be determined is how much better these vehicles should be compared to human drivers before they are widely implemented and offered to consumers.
A new study by RAND Corporation has reported that autonomous vehicles require being fairly better, around 10 percent, than human drivers before they are widely engaged in the United States. More importantly, the study claims that this approach of early implementation, that is, even before the technology is perfected, could prevent thousands of vehicular accidents and fatalities over the next 15 years. This is in comparison with waiting for 75 or 90 percent improvement in the technology.
“Our work suggests that it is sensible to allow autonomous vehicles on America’s roads when they are judged to be just moderately safer than having a person behind the wheel. If we wait until these vehicles are nearly perfect, our research suggests the cost will be many thousands of needless vehicle crash deaths caused by human mistakes. It’s the very definition of perfect being the enemy of good,” explains Nidhi Kalra, co-author of the report and director of RAND’s San Francisco office.
Although the self-driving vehicles are proven safer than the average human driver, researchers recognize that they would still produce crashes as they are still susceptible to other hazards, such as inclement weather, complex traffic situations, and even cyber-attacks. These uncertainties on their performance were the basis of the calculations made by the study. The number of road fatalities over time was estimated considering hundreds of various plausible futures and safety requirements.
“This may not be acceptable because society may be less tolerant of mistakes made by machines than of mistakes made by people. But if we can accept that early self-driving cars will make some mistakes – but fewer than human drivers – developers can use early deployment to more rapidly improve self-driving technology, even as their vehicles save lives,” said David Groves, co-author of the study and co-director of RAND’s Water and Climate Resilience Center.