It is obvious that car crashes will be abundant with AV's. It's not due to AV errors, its due to two totally different thought processes working together in the same place. Humans crash every day, the reasons are numerous, and for some reason, AV technology companies think that their cars are impervious, blaming everything on the human driver. However, what the AV techs dot consider is that every decision made by a human driver is based on the perception of reality, and humans are not LiDAR's, our perception is filtered not accurate. In this instance, what AV tech is not including in their calculations are the fuzzy logic algorithms accounting for how a human will react in any given situation, all AV's do, are calculate distances between solid objects and review scenarios. These scenarios do not include "the hand of God," which in this case is the hand of man on another vehicle steering wheel. That is why AV will only work properly with 99.5% accident-free driving in a human-free ecosystem, anything less will generate an increase in accidents, not reduce them.
Here are a few examples of what has happened with AV's in the small time they have had on public roads.
Pittsburgh, PA. February 24th the case in question is a left turn. Left turns require that the driver make sure that there is no oncoming traffic and that they integrate into the lane as quickly and as safely as possible. Jessica McLemore was the human driver in this incident, the place was the four-lane Liberty Avenue, heading Northeast. McLemore claims that the Uber driver came from the opposite direction and stated that there were no more oncoming cars, she had her left turn signal on, and the Uber AV had its right turn signal on. As McLemore started to make the left turn, the other car didn't turn right, but continued straight on and slammed into the side of McLemore's vehicle.
According to McLemore, she and the Uber safety driver in the AV stopped and got out to confront each other. The driver agreed that both cars had their turn lights flashing and that his AV was in control of the car. What the Uber driver said was "I was not expecting someone to turn from the far-left lane into my lane." McLemore has everything recorded on a dashcam video which she maintains in her car.
Uber released a brief statement claiming that the AV in question did have its turn light on since it was planning to turn. McLemore claims that Uber has not replied to her request for damages and will pursue the issue further.
In another crash incident that happened in 2017, GM's AV Cruise project had an incident when one of it's AV's crashed into a motorbike in San Francisco. According to the reports, the AV was getting ready to change lanes, and as it started to make the change, a bike came up alongside it, most probable to overtake it, and then the AV decided that it did not have enough clarity to make the change and corrected its position. This caused the AV car to swipe the bike. GM Cruise claimed that the bike rider was at fault, since he should not have come so close to the car.
However we think that it doesn't matter where the bike was. If the Cruise LiDAR worked properly, then it would have noticed the bike and would not have come back. This is a serious "blind" issue for the Cruise LiDAR and points to flaws in its development. Having stated this, the two incidents go on to prove that an AV environment will only work in a 100% human-free system. Once AV's are on the road, they can communicate with each other and the traffic system, this will ensure and assure full safety protocols to be active. The moment you add a human into the equation, you add chaos, and this is the reason why AV's cannot be allowed to drive on human roads., or for humans to drive on AV roads. The number of accidents will only increase, not decline