At our fleet fuel card company, we think autonomous vehicles are a pretty neat idea, but we’re always wondering about the things that could go wrong if they become the norm. We came across this Forbes article by Sam Abuelsamid and found it was pretty interesting, so we thought we’d share:
“Over the past several years of testing, Google’s fleet of autonomous vehicles have been involved in a number of minor accidents, but up until now all of them were caused by human driven vehicles running into Google’s machines. That has now changed as an autonomous Lexus RX450h slowly ran into the side of Mountain View, Calif. transit bus. While autonomous vehicles like Google’s prototypes can often see a lot more than human drivers and respond properly, they can miss things that we do pick up on and that’s where vehicle-to-external (V2X) communications can help.
Take for example the classic problem of the four-way stop intersection. The basic rule for this common roadway encounter is that the vehicles stop and then cross the intersection in the order they arrived, first-in, first-out. However, when two or more vehicles arrive at the same time, the one furthest to the right goes first and the other’s follow going clockwise around the intersection. In many instances drivers seem to forget the second part of this rule and end up signalling their intent by giving gestures to each other to go if they are being polite.
If self-driving cars encountered the same scenario, chances are they would have a better chance of detecting small differences in arrival time that humans would miss and simply do the right thing. But what if two cars from different manufacturers with different software or sensors with different sensitivity had a disagreement about who was first? They could end up sitting there waiting for the other to go much as humans often do. However, if they were equipped with vehicle-to-vehicle communications, they could silently signal their intent to each and one could proceed followed by the other.
How about if a human-driven and autonomous vehicle were in that situation. Since humans are more unpredictable, the V2V communications could really help. As soon as the human released the brake pedal and went to move to the accelerator, a V2V signal to the autonomous car could indicate for it to hold position until the way was clear, even before the human driven vehicle had an motion that could be measured by the sensors.
Now consider the situation with the recent collision. The Google car which was apparently stopped by a construction barrier was preparing to pull out and assumed the bus would slow or stop, clearing the way. Unfortunately, the Google car has no way of detecting the bus driver’s intent from outside. Although there is no guarantee, a human driver might have looked over the shoulder and made eye contact with the bus driver which have might have made the decision different.
Recently, Google has apparently been adjusting the driving control algorithm to make it somewhat more aggressive after complaints that the system was too timid on the road, often getting stuck while trying to merge into traffic.
Had both the Google car and the bus been equipped with V2V communications, there is a distinct possibility the accident could have been avoided. Knowledge that the bus driver was not slowing or steering away could have been communicated wirelessly to the Lexus, providing that indication of intent, proactively.
Navigant Research projects that up to 80 million new vehicles annually will be equipped with V2X communications capability by 2025. Not everyone in the industry agrees that this is necessary, including some OEMs and some members of the Google cars team that have expressed skepticism in the past.
While there is a reasonable argument to be made that the reduction in accidents as a result of V2V won’t be as high as the 70 to 80 percent projected by U.S. Secretary of Transportation Anthony Foxx, the ability to share real-time information between vehicles, pedestrians and infrastructure can certainly help fill in some of the gaps in the view provided by other autonomous vehicle sensors. Later this year, General Motors GM +0.70% will be the first automaker to launch V2V capability on the 2017 Cadillac CTS and other automakers including Toyota, Honda and BMW are all bullish on V2X technology.
Will this help get the traffic accident and fatality rate down to zero? Probably not all the way, but it will surely be another incremental step in cutting down the more than 1 million people who die on the world’s roads every year.”
We’d love to hear what you think about autonomous cars. Share your thoughts in the comments section!