The Road Ahead

Examing and Predicting the World of Autonomous Vehicles

Auto Pilot in the Sky and on the Road?

Over two million people take to the skies every day in North America be it for business or leisure. The ability for one to travel across a continent in a matter of hours is astounding, and may be one of the most important and impressive technological advances of the last century. Yet research has shown that 40% of Americans still experience a fear of flight. While it is manageable for some, many are incapable of taking advantage of this miracle of transportation that, to many, has become routine.

Where does this fear stem from, and specifically what are the causes? While there are many, the most popular responses are due to fear of heights, fear of crashing, claustrophobia, fear of panic attacks and perhaps the most intriguing, the fear of not being in control. It is not irrational to have some trepidation of putting your life in the hands of pilots. Yet many do without hesitation every day.

Or do they?

A study conducted by Sarter and Woods (2009) looks at how certain technological advancements led to the automation of a multitude of functions in the cockpit of commercial flights. Essentially, most passenger flights are in the pilots’ hands for takeoff and landing but once the plane enters the skies, the autopilot takes over. Autopilot, which communicates with numerous sensors that surround the aircraft can control speed, elevation, anticipate turbulence and make the necessary adjustments.

This often leads to an interesting comparison with the future of self driving cars. Many imagine in a not so distant future where human pilots will be responsible for getting their car to the freeway, and then, just as an aircraft pilot flips the switch, will enable autopilot mode and cruise down the freeway in a designated lane. While this may seem feasible to many, it is important to note complications that will certainly arise. What exactly will those challenges be? Interestingly enough they are hurdles the FAA (Federal Aviation Administration) are presently dealing with.

Will Cars’ Autopilot Follow the Same Trajectory of Passenger Aircraft?

With autopilot accounting for the vast majority of flight time, how can we ensure that a pilots’ fundamental skill: flying a plane, is not lost? Currently the answer lies in pilot training, testing and retraining. Is that enough? Many recall the heroics of Capt. Chesley Sullenberger  when he landed US Airways Flight 1549 in the Hudson River January 15, 2009. After hitting a flock of birds, Capt. Sullenberger had to rely on the skills he was taught and practiced throughout his career in the Air Force and as a commercial pilot, to land the plane safely in the river, miraculously without any casualties to the passengers or crew. One has to wonder how a younger pilot, who has spent more time flying on autopilot than not, would fare in a similar scenario. While keeping pilots vigilant is an important part of the equation, ensuring that their skills do not deteriorate over time is a problem that needs to be addressed.

The Aftermath of US Airways Flight 1549 After an Emergency Landing on the Hudson River

When examining and designing the world of autonomous vehicles it is important to draw on similar industries and case studies so the same mistakes are not made. Retaining vigilance and skill amongst drivers is paramount until level 5 autonomy (full autonomy) is a global standard.

Levels of Automation

If we take another look back at individuals’ fears of flying on planes, it is important to note that there is one interesting. While lack of control is emphasized as a common fear, the hesitation on dependancy of an autonomous system controlling the flight is rarely mentioned. However, it it often cited as one of the most significant obstacles in the way of consumers’ willingness to adopt autonomous vehicles.

Why is this?

One belief is that it is due to a degree of separation from the cockpit to one’s seat. Passengers do not have the chance to interact and experience the intricacies of an aircraft like they do in their cars. The familiarity one has with vehicles, formed over a span of years is hard to disrupt.

While this may seem bleak, it is important to note scientific studies that examine human”s trust in self driving cars, notably a study by MIT AGELAB on autonomous vehicles and trust. Among other findings the study shows that more than 50% of every age group studied is willing to have vehicles that assist the driver, with a large subset of the 25-44 age range willing to accept full autonomy.

Abraham et al. (2016)

It should also be noted that a majority of almost every subset of the population studied would want automation features that: prevent collisions, help with speed control, help with steering and even periodically take over control of the vehicle. This is interesting as each feature has the potential to contribute to a gradual development of trust among these features. Trusting cars with small tasks, such as automatic parking and collision assistance can help prove the efficiency and reliability of autonomous vehicles.

Abraham et al. (2016)

No matter how sound the technology, it is vital to consider the hurdles facing the consumer, both in terms of education and willingness to use new technologies. Discovering the most effective ways of communicating how autonomous vehicles can solve many if not all of the pain points and reassure people’s hesitations is as important to the success and adoption of autonomous vehicles as the algorithms that run inside them.

Alexander Wulkan is a Junior at Tufts University studying Engineering Psychology (B.A. Human Factors Engineering) and Entrepreneurial Leadership from Montreal, Qc, Canada. 

References:

Pilot Interaction With Cockpit Automation II: An Experimental Study of Pilots’ Model and Awareness of the Flight Management System, (Nadine B. Sarter & David D. Woods) link

Autonomous Vehicles, Trust, and Driving Alternatives: A survey of consumer preferences (Hillary Abraham, Chaiwoo Lee, Samantha Brady, Craig Fitzgerald, Bruce Mehler, Bryan Reimer, & Joseph F. Coughlin) link

 

1 Comment

  1. Alex, I really liked your parallels between airplanes and self-driving cars. One other reason, I think, that people have a much higher trust of airplanes is that they are unaware how little the pilots do. I was talking about this to my parents a few months back when I asked them if they would trust self driving cars. My mom pretty much told me that she wouldn’t trust any vehicle that wasn’t primarily driven by a human. I was a little taken aback as she has no fear of air travel. So I proceeded by asking her how much she thinks pilots actually do in airplanes. When I explained that nearly the entire time the plane is in the air, the plane pilots itself, she was really shocked. I think a lot of older people in particular are unaware of how much autonomous technology is used in our daily lives. Planes are not advertised as fully autonomous, so the passengers don’t need to know or worry about how they are driven. There is this idea in computer science of data abstraction. The user never really needs to worry about how tasks are completed as long as the tasks are completed correctly and efficiently (so developers can use different algorithms that lead to similar outcomes). Data abstraction currently occurs with planes because, as you pointed out, there is so much separation between the cockpit and the passengers. However, you can’t be in an autonomous car without understanding that there is no human driver. I’m curious as to how the industry will deal with this and how this will affect public/government adoption. I’m also curious if cars in the future will try to show the passengers more or less information about what is going on in the computers. I guess this is just stuff we will have to watch play out!

Leave a Reply

© 2017 The Road Ahead

Theme by Anders NorenUp ↑