Bicycle Friendly America Summer 2018 - 13

and presented at the 2017 National Bike Summit, provides a
good start for the types of behaviors that could be subject to
reporting or testing.
Direct outreach to automakers on their behavior towards
bicyclists, pedestrians, and other outside of their vehicles may
also help. BikePGH did direct outreach to Uber after interactions between bicyclists and Uber automated vehicles in
Pittsburg, leading to a public email address for reporting those
interactions that had not previously existed.
A SOLUTION TO DISTRACTED DRIVING
Why should we expect a solution to distracted driving?
If you are not driving a vehicle then you can't be distracted
from driving it. Highly automated vehicles, SAE Level 4 and 5,
which do not require driver attentiveness are offered as a solution to distracted driving by removing the driving component.
This is also a motivation for technology companies investing
in automated vehicles as the 290 hours that the average American spends driving7 is time not spent using online services.
What barriers to solving distracted driving exist?
There are two barriers that I'd like to mention.
The first is that lower level automated vehicles presume that
human drivers will be attentive and engaged, ready to take over
the responsibility for driving in a seamless and safe manner.
Whether it is the multiple deaths of drivers over-estimating
the ability of Tesla's autopilot or the performance of the safety
driver when Elaine Herzberg was killed, it seems apparent that
people will have problems living up to the expectation of vigilance as the vehicle mostly drives itself.
Ngani Ndimbie, a Women in Transportation Fellow at Traffic21 Carnegie Mellon's learning lab for transportation technologies, spoke at the 2018 National Bike Summit about the
dangers of assuming driver attentiveness in SAE Level 2 and
3 automated vehicles.8 She proposed that bicycle advocates
should push for higher levels of automation to leap past a possibly unsolvable human issue of shared responsibility with automated systems.
The second barrier is that enforcement of distracted driving
may become more difficult if the use of handheld devices are
allowed in automated vehicles, but there is no feature that allows law enforcement to distinguish between automated and
standard vehicles.
What can be done now to ensure the automated vehicles solve distracted driving?
There do not appear to be any current proposed bills that
would address adopting SAE level 4 or 5 vehicles rather than
levels 2 or 3. While direct outreach to manufacturers could be
useful, it is likely that their views on these levels is driven by
their business models for automated driving system deployment.
At this time, reaching out to law enforcement and asking
them about whether they will be able to enforce existing disSUMMER 2018

tracted driving laws in mixed automated and human-operated
traffic or if they need additional legal or other tools might be
more promising.
BETTER BIKE INFRASTRUCTURE
Why should we expect better bike infrastructure?
Automated vehicles rely on sensors, machine learning, and
algorithms to determine how the vehicle should act. Reports
say that better roadway markings help these systems understand the world. Roadway markings are easier to read and
react to than humans. This could lead to automated vehicle
manufacturers and policymakers supporting separated bicycle
infrastructure, clearer bike lane markings, and other improved
bike-related markings that clearly convey where bicyclists
should be and provide a safe and comfortable experience for
bicyclists so that they want to be there.
What barriers to better bike infrastructure exist?
One issue to be aware of is that this infrastructure, if it is
due to deficiencies in the ability of automated driving systems
to safely detect and interact with bicyclists, may be something
that bicyclists are required to use. While most bicyclists would
not object to using well designed, well maintained, and well
connected bicycle infrastructure, the experience of bicyclists
and pedestrians between the 1930s to present when government policy enabled roads to be shifted from public spaces to
spaces for cars says that we should be cautious of the promise
of better infrastructure if it goes hand-in-hand with a requirement to use it - as the requirement happens at once but the
infrastructure and its upkeep is speculative.
What can be done now to ensure the automated vehicles go hand-in-hand with better
bike infrastructure?
The National Association of City Transportation Officials
has published its "Principles of Autonomous Urbanism" which
sets out a vision for how cities can improve bicyclist, pedestrian, and transit infrastructure while adapting to automated
vehicles. Among the principles are "rebalancing the right of
way" through which they hope that cities will "reallocate street
space to active, sustainable modes." The entire blueprint is
available at: https://nacto.org/publication/bau/blueprint-forautonomous-urbanism/.
Now is a great time for cities and/or regional transportation authorities to consider adopting policies that will guide
the development and adaptation of their streets to automated
vehicles.
MORE BIKE INFRASTRUCTURE
Why should we expect more bike infrastructure?
A widely publicized study of the potential impact on traffic
of driverless cars in Lisbon, Portugal found that a shared automated vehicle fleet available through a ride hailing app compleBICYCLE FRIENDLY AMERICA 13
https://newsroom.aaa.com/2016/09/americans-spend-average-17600-minutes-driving-year/ https://2018nationalbikesummitgrassroots.sched.com/event/D7EB/complete-streets-in-the-age-of-automated-vehicles https://www.nacto.org/publication/bau/blueprint-for-autonomous-urbanism/ https://www.nacto.org/publication/bau/blueprint-for-autonomous-urbanism/
Table of Contents for the Digital Edition of Bicycle Friendly America Summer 2018
