independent conclusions from reams of data about cargo routes,
historic fl ight and arrival times, and forecasts for traffi c and
weather. Over time, the predictions become more accurate as the
amount of data proliferates. When a forwarder books a shipment,
algorithms created by machine learning will suggest an optimal
route. For carriers, Digistics supports dynamic pricing based on
AI’s evaluation of historical pricing and shipper information.
Online capabilities
As part of its commitment to AI and machine learning,
Unisys has launched an online Artifi cial Intelligence Centre of
Excellence to develop capabilities in advanced data analytics.
In addition, Unisys is establishing centres of excellence at a
number of airports, most recently at Bengaluru Kempegowda
International, in India. The BIAL Analytics Centre of Excellence
will provide 150 business intelligence reports using AI and
machine learning to forecast trends for both passenger and cargo
movements.
“We are replacing the old system because it was so slow,”
reveals Kohli. “Traditionally, the reports were available via Excel
and it would take four or fi ve days to create them, then they
would be sent for review. We want to create automatic reports for
both passengers and cargo that give immediate feedback.”
Setting up centres of excellence at individual airports is not
as straightforward as it sounds, however. Kohli says that the
data systems are usually a complex blend of legacy methods and
more modern technologies. Creating an interface that integrates
them all takes time at the outset. But once the data is available,
machine learning is able to provide meaningful reports and
Unisys’ teams of data scientists can use the insights to develop
superior algorithms. “I describe air cargo as a broken system
because there are pockets of automation, but so much shipment
data is still manual,” says Kohli. “For example, too often the
AI is speeding up
processes
Dheeraj Kohli, Vice President
of Travel and Transportation, Unisys
It’s a great example of how
modern computing can help
to monitor the emotions of
cargo,” explains Kohli.
Digi-Pets is one of nine
cloud-based services offered
on the Digistics platform,
which won Best Software
Architecture at the ICMG
USA Architecture Excellence
Awards, in 2017. The Digistics
service is especially valuable
for optimising routes when
precious cargo has to arrive on
time.
“If it’s blood, or perishable
goods, or animals, a delay
would be an emergency,”
admits Kohli. “The algorithm
on Digistics even includes
weather data when it’s
suggesting routes. The booking
agent can inform the shipper
that there’s a hurricane
coming to Miami, and it
would be better to send the
goods via New York this time.
It might cost a few dollars
more, but it will defi nitely
arrive on time,” Kohli says.
Security services also rely
on Digistics’ cloud-based data.
Kohli is unable to give specifi cs
because the information is
classifi ed, but he says that
security agents are monitoring
cargo movements using the
Digistics platform in order to
search for potential threats.
Behind the Digistics
analytics lies the advanced
AI computing method of
machine learning. These
programmes use “neural
networks” that function
similarly to neurons in the
brain, allowing them to draw
customs forms are manual
and every country has to have
one. This is the old-fashioned
system we’re fi ghting against
and it will take time, as air
cargo has lagged behind in
these areas.”
Kohli’s long-term vision is
to connect airports together
at district, then regional,
then national level, to enable
them all to benefi t from each
other’s analytical insights.
There would need to be
Unisys analytics centres at
a number of large and small
airports, but some airports
could simply be connected on
the cloud. “We’re in highlevel
discussions to automate
the Digistics operations for
an entire country. I can’t
say which one yet as it’s a
competitive bid - but it’s a
large nation,” he says.
Benefi ts to the cargo
community
There are a multitude of ways
in which AI has the potential
to benefi t air cargo operations.
A contrasting example to
Digistics is that of IAG Cargo,
where there are ongoing
experiments with self-driving
vehicles designed by Oxfordbased
Oxbotica. Last year, IAG
Cargo carried out a 25-day
trial at Heathrow airport. The
CargoPod ran autonomously
along a pre-determined cargo
route around the airside
perimeter, gathering data
about its own performance.
“We collected more than
200 kilometres’ worth of route
information about how
www.airlogisticsinternational.com June 2019 23
/www.airlogisticsinternational.com
/www.airlogisticsinternational.com