Autonomous cars are testing on public roads. This is nothing new. They’ve been out there for years. Their track record isn’t perfect, but it’s pretty good. Despite that record, a recent crash involving an autonomous Uber car in Tempe, Arizona is getting lots of attention.
The crash looks dramatic with the autonomous Volvo on its side in the road next to another visibly damaged car. This wasn’t some minor fender bender resulting in nothing more than a few scratches, which has been the bulk of what we’ve seen in autonomous crashes.
It’s the rather dramatic first impression of the photos that has people talking about this one. It also has Uber taking quick action. Uber is suspending its autonomous car testing programs in both Arizona and Pennsylvania while the incident is investigated.
That’s decisive action and helps ensure people don’t see these cars as unsafe, although plenty will be unconvinced. Autonomous cars are still a little sci-fi for many people, so any crash gives more reason to be distrustful of the technology.
In this case, however, it looks like the technology wasn’t at fault.
Police in Tempe say the Uber wasn’t responsible. The crash occurred when another vehicle failed to yield. The only person in the Uber was the driver, but no one has said whether the car was operating in autonomous mode or if the driver was in control at the time of the crash.
That’s part of what Uber will be trying to figure out as quickly as possible. The sooner the details of how this crash happened are released, the better for Uber and autonomous cars overall.
People are upset by this crash because there’s the perception that a perfectly operating autonomous car shouldn’t be involved in a crash. People think these cars should be able to avoid what humans can’t, and isn’t that a large part of what makes them so great in the first place?
Yes and no.
Autonomous technologies do prevent accidents. Features like automatic emergency braking, which is found on numerous production cars that still require a human driver, do prevent accidents and save lives. The problem is an autonomous car or one driven by a human can only do so much to prevent a crash.
The real trick isn’t so much getting a car to stop before it hits another car, but getting both of these cars to talk to each other and avoid any kind of issue well before it occurs. This vehicle-to-vehicle (V2V) communication is a key component to eliminating crashes.
Even a perfectly functioning autonomous car can’t avoid every crash. Cars can only move in so many directions and sideways is not one of those directions. Another car can hit an autonomous car with no way for that little slice of the future to get out of the way in time.
But equip those cars with V2V communication and both vehicles will have plenty of notice about a potential crash. Instead of either car needing to take sudden evasive action, they could both calmly slow down and avoid the problem with passengers none the wiser.
The crash in Arizona is worth investigating and will provide plenty of data for engineers trying to perfect autonomous cars. It will lead to improvements in the future, but until V2V communication becomes a reality, even the smartest autonomous car can’t avoid every crash.