Вы находитесь на странице: 1из 9

Cherpeski 1

Jonathan Cherpeski

Dr. C

ENG1101-219

13 November 2018

How would automated vehicles affect driving safety?

I always get asked, “What is your dream car?”, and I’ve always responded with a

shrug of my shoulders. I have put in a lot of thought in the past to try and answer that

question, but I never got far because nothing ever popped out at me. About a year ago,

I came across Tesla, a car company that only produces electric cars and they are one

of the leading companies in autonomous vehicle (AV) technology. Initially I was

interested in the quickness of the Tesla’s, however the autonomous driving aspect got

my curiosity gears turning. Ever since then I have wanted to investigate how

autonomous vehicles could affect overall driving safety, but also how and when they will

be implemented into society.

Automated vehicular technology has developed quickly in recent years with some

vehicles with semi-autonomous abilities driving on public roadways. AVs are supposed

to create a safer environment for drivers and reduce crashes worldwide; however, there

is minimal research to give evidence for that claim. Even though there have been over

50 companies that have got permission to test AVs on public roadways, which include

Tesla, Waymo, Apple, BMW, Honda, Ford, Uber, Intel and Nissan. Companies keep the

results rather than releasing them to the public, for no clear reason. However, there

have been recent studies performed by institutes worldwide to provide potential

evidence towards the big question “how safe are AVs”? (Peng)
Cherpeski 2

In April of 2018, there was a study that was co-conducted by the Department of

Civil Engineering, Monash University, Melbourne, Australia and School of Civil and

Environmental Engineering, Nanyang Technological University, Singapore. The study

was to investigate and estimate the potential safety impact AVs would have on driving

safety. This study was conducted using VISSIM (a traffic flow simulation software),

which was used as the traffic microsimulation platform, and also using Surrogate Safety

Assessment Model (SSAM), which is used to determine the number of potential

conflicts from the simulated data. Where “conflicts” means how many evasive

maneuvers are made by a vehicle to avoid a collision. This study was conducted on two

case studies, a four-way signalized intersection and a roundabout. Both of which are

located in the real world, the intersection is in Australia and the roundabout is in New

York. For both cases, they varied the AV penetration rate, where 0% would be no AVs

involved and 100% would be only AV involvement. The AVs are also set to have shorter

headways (the distance from between vehicles) to increase road capacity and reduce

relay. For the signalized intersection, AVs reduce the number of conflicts by 20% to

65% with the AV penetration rates of between 50% and 100% (statistically significant at

p<0.05). For the roundabout, the number of conflicts is reduced by 29% to 64% with the

100% AV penetration rate (statistically significant at p<0.05). These statistics do not

include the results when the AV penetration rates are between 0% and 50% because

they do not show a consistent decrease of potential


Cherpeski 3

Above: Total number of conflicts by AV penetration rate for Above: Total number of conflicts by AV penetration rate for
the roundabout (Morando) the signalized intersection (Morando)

conflicts across all cases. These results suggest that a high AV penetration rate may be

necessary to deliver the anticipated safety benefits. (Morando)

Given these results, AVs do seem to profoundly improve overall driving safety.

However, that is only the case if there were strictly AVs on the road, which may prove to

be difficult given society, as a whole, does not trust AV technology. There was a recent

study done by Kanwaldeep Kaur and Giselle Rampersad (two professors from College

of Science and Engineering, Flinders University, South Australia), that investigated the

key factors influencing the adoption of driverless-cars and how to allow the public to

trust them. They claimed that society does not trust AVs. Before conducting the study,

they decided to refer to technology adoption models (Theories that provide the relevant

factors influencing the public’s trust) to compare their results. They used a case study

where driverless cars are being implemented in a closed environment at Flinders

University, a mid-sized university in Australia. They would be used as


Cherpeski 4

commuting tools for staff and students. They decided to use a survey to collect data on

how open the staff and students are to this implication of AVs. The results revealed the

situations where people were most likely to adopt AVs, all those situations involved the

AVs being used under a closed circumstances. While the study was not open to

everyone, it does imply that given the right circumstances and opportunities, AVs can

be implemented in ways that will positively influence their public reputation (Kaur).

In 2015, the Boston Consulting Group performed a research project that

surveyed over 5,500 people in 27 different cities in 10 countries. At the time, it was the

biggest survey that was strictly for autonomous driving. Out of the wide variety of

questions and ideas being proposed, there is only two that are important to the topic at

hand. Fifty eight percent of people said that they would trust a self-driving vehicle, and

69% said that they would trust a partially self-driving vehicle. This result is surprising

based on the assumption made early by Kaur and Rampersad that the general public

consensus of AV technology is concern about safety. The other research question

presented the other 42% of people who would not ride in a self-driving vehicle with 10

reasons why. They then chose any option that applied to them. Fifty percent of people

claimed that they did not trust the car driving itself, and 45% said that they wanted to be

in control at all times. While the first result seems to contradict the idea assumed by

Kaur and Rampersad, there’s still 42% of people who would not trust an AV. Of those

42%, half of them agree that they would not feel safe. With that being said, Kaur and

Rampersad’s assumption that society does not trust AV technology can be refuted.

However, 42% is still a large percentage of the population (Lang)(Kaur).


Cherpeski 8

Above: 58% of people say that they would trust riding in a Above: From the 42% of people who did not trust riding in self-
self-driving car. (Lang) driving cars, 50% of them claim that they do not feel safe with
the car driving itself. (Lang)

Many people look at statistics and research studies to find a direct answer to the

impacts AVs may have on driving safety; however, Nicholas Evans has a different view

on AVs. He studies ethics of AVs decision-making processes as a philosophy professor

at the University of Massachusetts-Lowell, and he investigates a collision involving an

Uber self-driving test vehicle striking a pedestrian in Arizona. He proposed three

questions, the first being “could a human driver have avoided the crash?” He concluded,

based on visibility and the speed limit that it was easily avoidable. The second question

he proposed was “are self-driving cars still generally safer than human-driven cars?” He

used the federal estimate that 1.18 people die per 100 million road miles driven and

compared that to Uber’s cars only 3 million miles with 1 death. Evans goes on to explain

that there are roughly 32,000 people that die in vehicle-based accidents in the U.S.,

about 5,500 of them are pedestrians. Assuming AVs are made to prevent accidents with

other vehicles, that is may not translate into overall safety. Based on this assumption,

it’s reasonable to claim that AV implementation could cut total vehicle fatalities in half to
Cherpeski 8

16,000 while doubling the pedestrian vehicle fatalities to 12,000. This led to the third

and final question, “how safe is safe enough?” He made a comparison to drug

companies and how they test new pharmaceuticals. They put them through a series of

tests to prove its effectiveness rather than just putting it on the market hoping nothing

bad happens. He concluded by suggesting that AV technology should be able to do

much more than simply detecting pedestrians (Evans).

The ways AVs detect objects, lanes, and how to drive bring up another issue as

explained by Paul Wagenseil, the senior editor for Tom’s Guide, an online guide to tech

products (Wagenseil also talk and panels at technology conferences). Wagenseil

described an autopilot vehicle on a traffic-filled freeway, describing said vehicle to be a

“cybernetic grandma”. Saying how these vehicles are made to be too cautious. To

prevent that, companies would be forced to program the vehicles to break the law,

which no sane company would do. He then mentions all the other factors that people

have to take into account, such as weather and kids randomly running in the street. He

quotes Steve Shladover, director of the University of California’s Partners for Advanced

Transportation Technology (PATH) Program, “Even the most sophisticated of those test

vehicles is far inferior to a novice driver” (as quoted by Wagenseil).

Autonomous vehicles are expected to create a safer environment and reduce

crashes worldwide, and there is statistical evidence from research simulations that

concur with that expectation. However, that is only true when over 50% of vehicle on the

roads are AVs. This brings rise to how we can convince the public to trust AVs enough

to reach that target percentage, because recent surveys show that almost half the

population do not trust self-driving cars. AVs can show their worth if they are slowly
Cherpeski 8

made useful in closed environments around the world, rather than going straight onto

public roadways. That process will take time, as a result, AV technology will most likely

improve and become more capable of being safe enough to be exposed to public roads.

This way, the vehicles will be advanced enough and trusted enough to reach the target

percentage and live up to the high safety expectations.


Cherpeski 8

Works Cited

Evans, Nicholas G. “Self-Driving Cars Can't Be Perfectly Safe – What's Good Enough?

3 Questions Answered.” The Conversation, 18 Sept. 2018,

theconversation.com/self-driving-cars-cant-be-perfectly-safe-whats-good-

enough-3-questions-answered-92331

Kaur, Kanwaldeep, and Giselle Rampersad. “Trust in Driverless Cars: Investigating Key

Factors Influencing the Adoption of Driverless Cars.” Journal of Engineering and

Technology Management, vol. 48, 2018, pp. 87–96.,

doi:10.1016/j.jengtecman.2018.04.006

Lang, Nikolaus, et al. “Self-Driving Vehicles, Robo-Taxis, and the Urban Mobility

Revolution.” Https://Www.bcg.com, Boston Consulting Group, 21 July 2016,

www.bcg.com/en-us/publications/2016/automotive-public-sector-self-driving-

vehicles-robo-taxis-urban-mobility-revolution.aspx.

Morando, Mark Mario, et al. “Studying the Safety Impact of Autonomous Vehicles Using

Simulation-Based Surrogate Safety Measures.” Journal of Advanced

Transportation, Apr. 2018, pp. 1–11. EBSCOhost, doi:10.1155/2018/6135183

Peng, Tony. “Global Survey of Autonomous Vehicle Regulations – SyncedReview –

Medium.” Edited by Michael Sarazen, Medium, SyncedReview, 15 Mar. 2018,

medium.com/syncedreview/global-survey-of-autonomous-vehicle-regulations-

6b8608f205f9.
Cherpeski 9

Wagenseil, Paul. "Self-Driving Cars Are Not As Safe As Vehicles Operated by Human

Drivers." Opposing Viewpoints Online Collection, Gale, 2018. Opposing

Viewpoints in

Context, http://link.galegroup.com/apps/doc/ZLVJRK680638574/OVIC?u=dayt30

401&sid=OVIC&xid=ab3a185f. Accessed 15 Oct. 2018. Originally published as

"Pull Over, Robot! Self-Driving Cars Should Be Off The Roads," Tom’s Guide, 20

Sept. 2016

Вам также может понравиться