Grace Johnson
8 January, 2018
Robots are no longer a thing of science fiction. Instead of it being outlandish to dream of
receiving aid from machines, now goods are manufactured by automatons on assembly lines, and
A.I. programs answer trivia and conduct searches from within smart phones. With these
mechanical and technological advances already entering so many different areas of life, robots
According to research by Google, 94% of all vehicular accidents are caused by human
error. This could refer to many circumstances, whether drivers are distracted by smartphones or
disobey laws through ways such as accidental, or intentional, speeding. Because humans are
prone to making these careless choices, they potentially endanger the lives of themselves, their
passengers, other drivers, and even pedestrians. Of all the ways that this hazardous behavior has
been combated in recent years, one of the most popular avenues still being explored, and
The development of A.I. systems has been a popular topic for research in recent years,
and the idea of putting these systems into autonomous modes of transportation is similarly big
with many companies. Mercedez-Benz, a car manufacturer, and Waymo, an offshoot from
Google's parent company Alphabet Inc., are some of the most prominent. Examining the ways in
which the two companies choose to handle the matter of safety, enhancing it with the use of A.I.
technology and automotive engineering, is useful in allowing the public to gain confidence in
passing control of the wheel to another, as well as informing technicians of areas they can
improve on in the future. Self-driving vehicles have the potential to greatly increase both
figuring out what can, and does, commonly go wrong on the road, vehicle engineers and
designers learn what exactly to augment and pay special attention to; noting the different kinds
of hazards is essential.
As recorded by the National Highway Traffic Safety Association, the NHTSA, the most
common causes for car accidents are driver related errors (NHTSA). Distracted driving in
particular is one of the most prevalent ways a person can increase the chance of injury. The
window of action as a car swerves into the wrong lane or some inattentive pedestrian walks into
traffic is small for a driver, and yet the quick decisions made in that window can mean the
difference between life and death, or at the very least costly injury. One is unfortunately unable
to react quickly when his or her focus is split between the road and elsewhere.
Texting while driving is a common form of casual safety negligence, one of the "most
alarming [distractions]" (Currin). Cell phones have become increasingly popular in the last
decade, especially with the newest generation of drivers, and so too has the habit of reading and
sending messages while behind the wheel. Having one’s head down for even a few seconds
while traveling at high speeds, left unaware of light changes or sudden stops from the vehicles up
ahead, is a disaster waiting to happen. The idea that a person can multitask in such scenarios is
Aside from phone usage, drunk driving and driving while drowsy are also forms of error.
Drunk driving has long been an often-discussed cause for vehicular accidents. Intoxication,
though acceptable while in a bar surrounded by friends or in similar settings, dulls the senses and
impairs a person’s awareness while in a vehicle, making reactions to accidents slower which
endangers any possible passengers. Fine judgement may also be at negatively impacted; common
laws like obeying the speed limit or wearing a seatbelt might be forgone by a hazy mind and
make injury in a possible collision worse. Although many don’t realize, drowsiness can cause
many of the same effects that drunkenness does. A tired driver would have a mind that’s just as
clouded depending on the level of sleep deprivation, in addition to having the added risk of
drifting off while en route to a destination. Thus, he or she would also be unfit to sit behind the
wheel of a vehicle.
There are, of course, other ways that accidents can occur without the driver directly being
at fault. Problems with the vehicle itself are a more indirect issue in that maintenance should be
sought after prior to pulling out of one’s drive-way. In some cases, road conditions are less than
favorable. Debris from previous wrecks or wayward trash becomes a hazard. In addition, weather
creates danger in the form of wind and precipitation, which can be especially dangerous in hilly
areas and roads with multiple turns; warmer climates face literal roadblocks in the cases of sleet
or snow as drivers have less experience in those conditions. Driver error, however, remains the
most controllable and most prevalent aspect of vehicular crashes, and thus is the aspect multiple
Google's self-driving car company Waymo has safety as one of its main goals. The
company states that they are building a car that "is always alert and never distracted" () on its
website, both directly acknowledging driver error and stating that it will be subverted with the
vehicle. To this end, the company has turned to artificial intelligence capable of learning road
patterns and four different types of sensors. Its system ranks as a five, or “Full Automation”, on
the scale of autonomous vehicles created by the Society of Automotive Engineering (SAE
International) and adopted by the NHTSA for the sake of manufacturing clarification. A five on
this scale, which measures the amount of human involvement needed to drive the vehicle, means
In the interest of transparency, Waymo recently put out a many paged “Safety Report”
detailing its methods to the public. It details, among other things, the four types of sensors that
the cars use to perceive the environments around them. The Light Detection and Ranging system,
or LiDar System, uses lasers to determine the distance from surrounding objects to the vehicle by
measuring the time it takes for lasers to hit something and return; the Vision System uses 360-
degree cameras to get a view similar to that which human driver would see, capable of detecting
color and thus recognizing signs, flashing lights, and certain vehicles; the Radar system detects
objects and their movements, even through precipitation and visibility-reducing weather such as
fog; and the other supplementary systems include applications such as GPS and audio receivers
which can pick up emergency sirens. Together the sensors “paint a detailed 3D picture of the
world” (Waymo) to be used by the “brain” of the vehicle in traffic analysis, prediction, and
decision making on the road. These methods make the Waymo self-driving car significantly
more aware of its surroundings than a human driver, beyond even the potential to be distracted
by texting or tiredness. The car’s advanced capability is aided by its A.I.’s collective prior
experience; it’s gained through practice driving on the road, on simulated roads, and on test
tracks that put the vehicles through multiple hazardous scenarios and shared collectively amongst
the artificial intelligences. The fully autonomous system that Waymo employs not only removes
the chance for a person’s distraction to negatively affect othe6rs, it provides passengers with a
more well-suited driver in their stead, and this method is only one way to get driving safety out
of human hands.
Mercedes-Benz has its own automotive systems that it plans to save lives with. In
contrast to Waymo, the car they’re featured in, the Mercedes-Benz E-Class First Drive, still
requires a level of human involvement to run smoothly (Golson). As stated previously, Waymo's
systems operate at a level five on the NHTSA’s self-driving scale; Mercedes-Benz’s autonomous
augmentations are closer to a level three, “Conditional Automation.” Vehicles at this level can
control certain operations in certain instances, “but the human driver must be ready to take back
control when the automated system requests,” (NHTSA). As reported by Autoblog, E-Class can
drive continuously in a single lane at up to 130 mph, correcting its steering to keep in said lane
using sensors that detect markers on the road, due to its Drive Pilot system. Failing said
indicators, Drive Pilot can still have a vehicle follow traffic at up to 80 mph. Other “Driver
Assistance” systems within the car can prevent against possible accidents by alerting the driver
of pedestrians in the car’s blind-spot and breaking automatically to avoid collisions. Despite the
inability to handle driving by themselves, the autonomous systems can still handle a large
majority of the driving. To cover for the parts that they can’t accomplish, a beeping occurs at
thirty-second intervals to keep the driver focused on the road and prevent he or she from getting
too comfortable with the A.I. having control (Golson). Mercedes-Benz has developed many
systems that are capable of, if not removing the threat of driver error entirely, augmenting the
threat to a point where humans are responsible for fewer aspects of driving, and thus fewer
Though many companies that have invested years into developing autonomous
technologies, most of them are yet to create a finished product that will be road worthy for many
years; there are ways that the safety of self-driving vehicles can thus improve before they are
widely sold commercially. The way self-driving cars will handle unavoidable crashes is one area
of contention, specifically concerning ethics. The Trolley Problem is a moral dilemma that many
are looking at seriously. It proposes a situation in which a person is in control for the rail switch
for a trolley, and both the current track and alternative have people standing on them, the first
with a group while the second has only one. The problem then begs the question that, in a
situation where lives will end despite the choice made, should one make the decision to save the
many at the cost of the few? While it may seem irrelevant, this dilemma will inevitably matter
when it comes a time for a self-driving car’s A.I. to choose between protecting the passenger or
the civilian. How the engineers and developers choose to answer the question matters immensely
in these scenarios.
Mercedes-Benz does, in fact, have a level five Fully Autonomous car at the concept level
of development, the F 015 Luxury in Motion, planned to be released in later decades. Although
Mercedes does not yet have the system to run it, the German company still has plans for the way
it should handle potential crashes. In an interview for Car and Driver magazine at the Paris
Motor Show in 2016, Driver Assistance Systems manager Christian von Hugo claimed that the F
015, as well as all future level 4 and 5 cars, would always protect the passenger because, "If all
you know for sure is that one death can be prevented, then that’s your first priority” (Dodgson).
This spiked a big controversy in the public, and the statement has since been rescinded, with a
spokesperson claiming that “neither programmers nor automated systems are entitled to weigh
the value of human lives” (Orlove). While this may feel more morally correct, one of the policies
organization of the NHTSA, says that the ability to make moral decisions in cases of such a crash
and others is necessary for the vehicle to be approved (NHTSA). A positive to this issue is that,
as claimed by many, introducing more self-driving vehicles to the road would make it more
likely that accidents simply wouldn’t occur in the first place. This is no substitute for an answer
to the Trolley Problem, as the reckless actions of pedestrians can be uncontrollable and
inevitable in many cases. However, mitigating the reckless actions of those behind the wheel
now is the first step to reducing the amount of people hurt overall and in the future.
Ultimately, self-driving vehicles have the potential to vastly improve the safety of
everyone who uses the road, directly as travelers and indirectly as pedestrians. Human error in
drivers, the habit of being tired, intoxicated, or otherwise distracted while behind the wheel, is
what causes most roadway accidents. Self-driving technology can prevent these accidents by
replacing humans as the ones in control. Unlike humans, they can never stop paying attention to
their surroundings. Sensors systems like the LiDar and the Radar used by Waymo allow level 5,
fully autonomous vehicles to be incredibly perceptive of road conditions and the locations of
near-by vehicles or cyclists, regardless of the weather or the time of day. Even less automated
vehicles of third or fourth level can lessen the burden on a human driver, as proven by Mercedes-
Benz. The company’s E-Class Drive Pilot can continuously follow traffic and detect and stay
within lane boundaries, and human intervention is only necessary in certain circumstances like
tight turns. The fact of the matter is that the systems they utilize, and the engineering involved in
creating their designs, makes them better equipped to drive than even the most diligent person. It
is true that the ethical question involved in giving a machine control during an accident, and
programing exactly how they should act in no-win situation, is an issue that must be given due
consideration. However, that does not change the fact that increasing the number and the use of
self-driving vehicles would do less physical harm than not deploying them in the first place. The
F 015 Luxury in Motion is being created with the idea that autonomous transportation will
completely transform the experience of a car ride, allowing passengers to regain the time lost to
transit and focus work or each other. Whether or not this is really the way driving culture will be
shaped in the future, self-driving vehicles will allow people to stop worrying about theirs’ and
Austin, Michael. “2017 Mercedes-Benz E-Class First Drive.” Autoblog, Autoblog, 8 Mar. 2016.
driving/distracted-driving.
Dodgson, Lindsay. “Why Mercedes' Decision to Let Its Self-Driving Cars Kill Pedestrians Is
Probably the Right Thing to Do.” Business Insider, Business Insider, 12 Oct. 2016.
Golson, Jordan. “Mercedes Will Give Tesla's Autopilot Its First Real Competition This
Orlove, Raphael. “Now Mercedes Says Its Driverless Cars Won't Run Over Pedestrians, That
one.nhtsa.gov/people/injury/research/udashortrpt/summary.html.