Self-Driving Cars Become a Reality

Yesterday, the New York Times announced that self-driving cars have received backing from the federal government.

According to the article, last year, nearly 40,000 people died in automobile accidents in the US. In theory, self-driving cars have the potential to change that. If all cars on the road are self-driving, no one can have accidents, right?

Unfortunately, it's not quite that simple. In fact, self-driving cars have created some of the most difficult ethical puzzles for our generation. Programmers must now determine whether cars should attempt to minimize a loss of life in general, or prioritize the life of the car's passengers. For instance, if an autonomous car is heading toward a crowd of people, should it continue forward, risking many lives, or spin off the road, risking its passengers' safety?

This problem has recently become much more relevant. According to the New York Times article, the first fatal self-driving car collision occurred this past May, when an automated Tesla crashed into a tractor-trailer in Florida.

Based on the results of the crash report from the National Transportation Safety Board, I calculated the speed at which the car was traveling when it crashed into a utility pole on May 7, 2016. At the time of impact, the car was traveling at 74 mph (33 m/s), nine mph above the posted speed limit on the Florida highway. After the collision, the car coasted off the road for approx. 297 feet (90.5 m) before colliding with a utility pole. The car then broke the pole and continued for an additional 50 feet while rotating counterclockwise, but I only calculated the speed at the time of impact with the utility pole.

During this 297-foot journey post-collision, the only force acting on the car in the x-direction (disregarding drag) was the force of friction. The weather conditions were dry, so the coefficient of kinetic friction was approximately 0.7. Since ∑F = m*a = force of friction = μ*normal force, and normal force is equal in magnitude to m*g, a = μ*g, or about -7 m/(s^2). (Final velocity^2) = (initial velocity^2) + 2*a*(distance traveled).Therefore, after traveling 90.5 m, the car had slowed to a stop by the time it collided with the utility pole. (Note: I ignored the effect of the collision with the truck itself on the car's initial velocity. I also ignored the fact that the car did not travel in a perfectly straight line after the collision, but veered off the road in a curved path according to the diagram above.)

Should we be glad that the government is supporting self-driving cars? Like most things in science, it depends. Perhaps, with more research, we can reduce the number of automobile accidents in the US. Ethicists, physicists, public health experts, and programmers will have to come together to make this pipe dream a reality. For now, though, we are a long way from a highway dominated by autonomous cars.

*Note: All sources are linked in text.