Three million people have died in automobile accidents during my lifetime, 3,004,031 to be exact.
Originally Published on Medium
Almost 40,000 people die in auto accidents every year in the US, and going back through annual records, 3,004,031 people have died in my lifetime. Many more people are killed every year worldwide, an estimated 1.2 million. In addition to fatalities, more than 2 million people are seriously injured every year in the US, and many of these have lasting problems. These deaths and injuries can be avoided, if we understand the causes and a potential solution.
Reckless Driving
Reckless driving is behind more than 90% of all fatal auto accidents. Excessive speed is a contributing factor in almost 30% of the fatal accidents. Drunk or impaired driving accounts for another 30%. Distracted driving is behind more than 10% of all fatal accidents, and increasing. Running red lights, ignoring stop signs, falling asleep, overly aggressive driving, and making improper turns are behind another 20% of fatal accidents.
In some cases, drivers may not necessarily be reckless, they may just may make fatal mistakes that could have been avoided. Such was the case in early 2020 in Florida, when the driver of a van with a college rowing team thought he had the right of way to make a left-hand turn, but didn’t. In Florida, there are left-hand turn arrows that give the right of way, but then they turn into green lights that mean you can turn but no longer have the right of way. The van continued to turn left after the green arrow turned into a green light, and it was hit directly by a truck, killing one passenger and severely injuring others.
90% of auto fatalities, more than 30,000 per year, can be prevented, and autonomous vehicles (AVs) can prevent these accidents. When you come across a news report of a fatal car accident, you can almost always see that it could have been prevented if the driver acted more responsibly, and AVs are more responsible.
Autonomous Vehicles Are Safer
Autonomous vehicles (AVs) aren’t reckless. They don’t drive drunk. They won’t speed excessively. They are never distracted. They won’t run red lights or miss stop signs. They will be careful to make safe turns. They won’t make the same type of mistakes that human drivers do too often. They may not be perfect, at least initially, but they will be safer than human drivers. I discuss AVs in my book — Autonomous Vehicles: Opportunities, Strategies, and Disruptions — but here are some of the reasons why AVs will be safer.
Perception
Human drivers are limited to their own perceptions of what they see around them. They can see ahead and in their rear-view mirrors. Sometimes, they need to narrow their perception to focus on a task at hand, like the driver of the van focusing on making the turn and not seeing the truck coming. AVs use more than a dozen sensors — cameras, radar, and lidar — all integrated into a computer system to perceive instantly everything that is happening all around them.Interpretation/Prediction
Humans, if they are paying attention, are pretty good at interpreting what they see. They can associate many factors and use instincts to to make quick interpretations and predictions on what will happen. AVs use artificial intelligence based on experience to make interpretations and predictions. They are getting to be as good as humans, and may soon surpass humans as they gain more experience.Experience and Learning
AVs have a much larger and broader base of experience than any individual human driver. 100 drivers have 100 different bases of experience. Rarely is driving experience transferred from one driver to another. AVs, on the other hand, all generally share the same extensive base of experience. They also drive more miles to gain that experience. Waymo, for example, has driven 20 million miles autonomously in 25 different cities. It has also “driven” 10 billion miles in simulation. A typical new driver may have a few hundred miles of experience before getting a driver’s license.Reacting/Acting
Drivers react instinctively, while AVs react using the computations of their computer systems. Although these are different approaches, the powerful computers in AVs are much faster than human in computing reactions. In may cases, the difference in reaction speed could be only 2–3 seconds faster, but this is enough to avoid many accidents.Following RulesAVs are programmed to follow rules such as fully stopping at stop signs, slowing to a stop at traffic lights, staying within speed limits, avoiding cutting off other cars, and yielding to pedestrians and bicyclists. Some drivers may find this boring, but it will be much safer.
Distractions/Impairments
There are many ways human drivers get distracted: using smartphones, talking to children, eating food, listening to music, talking with friends, looking at navigation systems, etc. AVs are never distracted. Human drivers also drive impaired with alcohol or drugs, leading to many fatal accidents. AVs are never impaired by drugs or alcohol.Being More Careful
AVs will be more careful than most humans. For example, instead of making a left turn to enter a busy road, the AV may take a safer right turn and then do a U-turn at the next traffic light. AVs will always travel a safe distance behind the vehicle ahead.
AV Accidents
There has been negative publicity about AV accidents, but most of it has been incorrect. For example, there have been several fatal accidents involving Tesla vehicles that are reported as accidents by autonomous vehicles. Tesla vehicles are semi-autonomous, not autonomous. As semi-autonomous vehicles, they only drive on their own in certain conditions and require continual driver attention to take over. All of these fatal Tesla accidents occurred because the drivers weren’t paying attention when driving semi-autonomously (unfortunately Tesla refers to its semi-autonomous mode as Autopilot). Tesla uses technology that is excellent at semi-autonomous driving, but it is very different technology from most other autonomous vehicle technology.
There was a extremely publicized and often cited Uber autonomous vehicle accident that fatally killed a pedestrian in 2018 in Tempe Arizona. If you read the police reports and watch the video of the accident from the vehicle’s camera, it shows that the accident could not have been prevented by a human driver. The woman crossed the road at night walking her bike, not at a crosswalk, out of the dark, appearing suddenly in front of the vehicle. She was visible for only two seconds before impact, and a human driver could not have reacted in time to avoid impact. It was so sudden that she was looking the other way and never even reacted to the vehicle with its headlights shining. Autonomous vehicles have been tested with tens of millions of driving miles without any other serious accidents, other than this one.
Why is there so much negative publicity about AV safety? One reason is that it makes more sensational press than human-caused accidents, that are so common that they aren’t even considered news. Another is that people just don’t understand autonomous driving yet. But perhaps the biggest reason is the special interests of what I refer to as the “auto accident industry”. This industry has a lot to loose when accident rates drop. It includes trial lawyers, insurance companies, auto repair shops, and many others. Their industry associations are tasked with trying to discourage autonomous vehicles.
It is inevitable that AVs will be involved in accidents, and some of these may be attributed to the autonomous driving systems. AVs record the video and digital data of all accidents, enabling them to learn how to avoid them in the future.
Cost/Benefit
The fight against COVID-19 required the US government to infuse several trillion dollars into the economy and shut down businesses that incurred incredible losses, perhaps another couple of trillion dollars when all is said and done. But these actions saved hundreds of thousands, perhaps millions, of lives.
Reducing auto accident deaths has a very different solution. Imposing regulations to get safer drivers hasn’t worked. Severely restricting driving the way businesses were restricted during COVID-19 is not realistic. The solution, as I referred to, is autonomous vehicles. In this case, there is little or no investment required by the US government. Approximately $100 billion is being invested in AV development by industry without much government financial support. There will be a cost to the “accident industry” but I think most people will find this an attractive trade-off for fewer fatalities. All that is required is a better understanding of the potential for AVs to save lives, which is the purpose of this article.