Smarter Workzones |
“It might be a combination
of all three. We might find that
connectivity is not needed at all
or additional coatings won’t be
necessary, so we just want to
find out what are the most
important attributes.
“Ideally, we want
to find things that can
improve workzone
safety now, whether
we improve vehicles
with low-levels of
automation currently
available or contribute
to the technology that
is to come.”
Testing environments
After simulation testing,
the plan is to test in a closed
environment before doing
testing in the real-world
environment.
Pennsylvania State
University is a partner in
the project, who, along with
researchers at Carnegie
Mellon, will provide direction
for what is tested.
Professor of mechanical
engineering, Sean Brennan, will
head up Penn State’s contribution
and is well aware of the challenge
Open roads to the future
Penn State University believes open-source mapping
data is key to enabling more effective research into
connected and autonomous vehicles
Penn State has its own mapping
they face. When introducing the
subject in his lectures, Brennan often
highlights what he calls the ‘four D’s
– dirty, dangerous, dull and delicate
– as a summary of the situations
across all industries that benefit
most from automation. Driving
features two of them –
dangerous and dull.
“A key factor that supports
automation – especially for
situations that are dull – is
repeatability,” Brennan
explains. “For automated
driving, we want an
environment or working
situation that is highly
repeatable and also we want
it to be highly predictable. We
encourage that level of predictability
for human drivers by creating
markers and signage with repeating
patterns, while certain colours
have certain meanings, and this is
almost universally true across all
cultures on earth.
“Computer-based automated
driving in those kinds of ‘follow
A key factor that supports
automation – especially for situations
that are dull – is repeatability
Prof. Sean Brennan, mechanical engineering, Penn State University
the lane’ scenarios is kind of easy,
it’s more or less been a solved
problem since the 1950s or 1960s.
“The real challenges are in other
situations that are not repeatable,
not predictable, or where the driving
rules suddenly or unexpectedly
change, such as workzones.
“The roads will often be repainted,
traffic will be directed into oncoming
traffic, the signs may be uncertain or
in odd locations or even conveyed by
human gestures, and the road itself
may not be suitable for normal highspeed
driving.”
With this project then, Brennan
and his team are keen to develop
test cases for AVs in realistic
workzone situations that could lead
to the creation of testing standards
by which suitable vehicles can then
be certified.
With the project set to run for
the next three years, it is impossible
to say for certain how successful the
team will be in achieving their aim.
However, with more funding for
researchers to focus on the area, there
is certainly a great belief that, having
long been a neglected area for CAV
research, the time is now coming
when workzone safety and efficiency
can be improved in new ways that
will benefit entire communities.
vehicle that is capable of
adapting to a number of the
varied automation strategies that
exist and the expectation is that it
will play a useful role in the project
underway to enable AVs to effectively
navigate workzones.
“OEMs and mapping firms have far
more sophisticated systems but one
of the things we strive to do with our
vehicle is to have an open architecture
using software and methodologies
commonly in use within academia”,
says Sean Brennan at PennState.
“We really want to collect data in
a way that allows us to share it and
benefit further academic studies
and industry partners as well as
government organizations.
“OEMs would obviously not want
their production suite data to be in the
public domain as their strategies could
then be reverse engineered, but if our
data is made public we can create test
cases for others to build on.”
The vehicle features a suite of
sensors that include both defencegrade
capabilities and commercial/
automotive ones, as well as an
onboard lidar system for mapping
the road geometry.
Above: Sensors can
be integrated into
workzone bollards
10% Of all traffic congestion in
the USA is estimated to be
caused by workzones,
according to the FHWA
024 Traffic Technology International March/April 2020
www.TrafficTechnologyToday.com
/www.TrafficTechnologyToday.com