INSIGHT Jason Chester - InfinityQS
“Yet even in these high-end automated environments, it is not uncommon to see ‘information’ about that
production environment being captured and recorded manually by operators.”
Jason Chester, Director of Global Channel Programs - InfinityQS
Industrial Automation vs.
Cognitive Automation
Manufacturers are seeing increasingly diminishing returns from industrial automation. Here, Jason Chester,
Director of Global Channel Programs, InfinityQS explains why
The nascent area of cognitive
automation (where we
apply the same level of
drive and innovation to
automating our information and
decision-making processes as we
do to our physical processes),
is starting to gain momentum
across manufacturing.
Automation has undoubtedly
been the tour de force of
manufacturing for many
decades. Ever since we crossed
the chasm from cottage
industry to the mass production
of goods, we’ve continually
sought new ways to replace
physical tasks with automated
alternatives. From the birth of
mechanised production lines to
the revolutions of electrication
and digitalisation, we’ve become
masters at leveraging technology
within manufacturing. is was
always inevitable. Humans have
always invented new things to
extend their physical capabilities
ever since the dawn of Stone Age
tools. While our focus has been
on migrating physical activities
from workers to automated
processes, the same cannot be
said of our cognitive activities.
Imagine this scenario. e
physical production process
from the input of raw materials
to the packaging and storage
of nished packaged goods
is entirely automated. Digital
control solutions are used to
keep the production process
within dened parameters,
such as temperatures, pressures,
line speed and ow control etc.
High-speed visual and sensor
monitoring systems work in
concert with robotic systems
doing everything from removing
anomalies to selecting and
sorting products.
Yet even in these high-end
automated environments,
it is not uncommon to see
‘information’ about that
production environment being
captured and recorded manually
by operators. ose operators
often have to remember to
record critical parameter
measurements periodically. For
example, quality stations pull
product o the line at regular
intervals and perform manual
inspection and recording
of quality characteristics.
Production counters are
captured for measures such
as production rates, units
produced/rejected or line up/
down time.
is process of manually
monitoring, recording and
analysing information from
the production environment
is fairly consistent throughout
manufacturing and even more
prevalent in lesser sophisticated
manufacturing environments.
e information is captured and
stored in a variety of formats -
from paper-based forms to Excel
spreadsheets. In some instances,
statistical techniques may be
used to manually plot data on to
statistical process control charts
to visually represent trends over
time, although that is certainly
the exception rather than the
rule. Having this information is
great, but it is how we collect it
and what we do with it (or what
we don’t do with it) which is
where our quest for automation
ends.
e data that is collected
in eect becomes the way in
which we sense what is going
on within our production
processes, in the same way
that our eyes and ears help us
sense what is going in the world
around us. Presumably, the
reason we collect and record
that data in the rst place is so
we can use it to make decisions.
If it is not, then I would argue
that the cost of collecting and
recording that data is not worth
it and is instead being driven
by some misguided belief that
it is just ‘the right thing to do’.
ese data decisions may be
based on operational questions,
such as is the product we are
making within specication?
Are we on target for production
volumes?
Transforming raw data into
a form that can meaningfully
support those decisions takes
a lot of eort. We need to do
things with the data to make
it conceptually meaningful
like perform calculations on
it, aggregate it, put it into
management reports, create
charts and graphs, and then
distribute it to those that want or
need it. e resultant information
then needs to be interpreted
before making the actual
decisions for which it is intended.
at all takes eort, time and
resource. In the vast majority of
manufacturers, it is a manual
process to a greater or lesser
extent. Because of that eort,
we tend to only apply it to those
decisions we know we need to
make, or think we need to make.
Cognitive automation is not
exclusively concerned with
automating decision making in
of itself, but rather automating
the actions and processes
of acquiring the knowledge
and understanding of our
manufacturing operations and
production processes in order
to support more ecient and
eective decisions making.
Leveraging this relatively
untapped opportunity will be
nothing short of a panacea for
the future of manufacturing
performance.
40 | Comms Business Magazine | July 2020 www.commsbusiness.co.uk
/www.commsbusiness.co.uk