26
October 17
E
mbEddEd
C
omputing
Developments in control electronics
enable intelligent vending machines
By Fabrizio Petris,
Omron Electronic Components Europe
This article reviews
sensors for customer identification,
presence detection and reliable
note recognition in the mechanisms
of intelligent vending machines, and
also discusses remote reset switches.
Vending machines are a relatively unrec-
ognised, but fast growing, market opportunity
for the electronics industry. This growth is
being driven by a new class of unit: the intel-
ligent vending machine, which marries the
convenience and accessibility of instant, on the
spot transaction fulfilment with the tailored,
entertaining and connected retail experience
that customers now expect. According to a
March 2017 report from Grand View Research,
this market is expected to reach $11.84 billion
by 2025. Each and every one of these machines
will encompass electronic systems to facilitate
the interaction with the user, to control prod-
uct and cash dispensing and to manage energy
use. What solutions are available to create this
new breed of intelligent vending machines?
Essentially, this new breed of intelligent vend-
ing machines offers a purchasing experience
that is personal and individual for each user.
The machines themselves can be tailored to
the application, in terms of colour, size, space
and dispensing rate, which is essential as roll-
outs can be quite small. They also embrace
fully the Internet of Things paradigm, and are
fully connected offering cloud-based server
support, scanning tools, and value-added
services based on web-based interaction.
Vendors can use the Internet to update con-
tent for their interactive platforms and pres-
ent custom advertising. What are these new
vending machines selling? Whilst the range
of goods on offer is huge, food and beverages
is one of the largest markets: especially hot
drinks. Research and Markets found that the
hot drinks segment dominated the industry,
accounting for more than 62% of the over-
all volume due to the surging demand for
refreshment drinks. Many of these machines
are located in offices, as they facilitate employ-
ees grabbing a quick snack in between breaks
without the need of placing an order and
waiting for it to be processed.
The barista in your favourite coffee shop will
quite naturally greet a business executive, an
elderly lady and a teenager very differently.
The more engaged staff will try and tempt
the customer with a croissant, a cake or a
healthy snack depending on their experience
of what this kind of customer or even this
individual customer normally buys. Vending
machine manufacturers recognise that this
interaction is an essential part of the experi-
ence, and can even be monetised. For exam-
ple, their market research may show that a
specific snack appeals strongly to customers
aged 20-25 – so a tailored offer to these cus-
tomers might have an impact. A quite differ-
ent offer might appeal to the sixty-plus age
group. So how can a vending machine emu-
late something that a human does quite easily
and naturally? Fitting vending machines with
a camera is straightforward and inexpensive,
but the challenge is interpreting what it sees.
Fortunately, consumer electronics is already
providing a solution. Phones can recognise
their owners face, and vending machines can
make use of the same algorithms and tech-
nology. Implementing them is a challenge
though – vending machines are produced in
tiny volumes compared to phones, and have a
very different architecture. The success of the
design depends critically on the use of good
algorithms that deliver a reliable result with-
out using excessive system resources. Even
well-designed algorithms will be processor
and memory intensive, and adding vision will
demand additional system resource.
The Omron HVC module (figure 1) is aimed
at such applications, available in low vol-
umes and readily integrated by any designer
without any need to understand the com-
plex algorithms needed to analyse the image
or the optical design. HVC builds on the
Figure 1. HVC builds on the Omron OKAO vision software, a proven set of image recognition
algorithms used in over 500 million digital cameras, mobile phones and surveillance robots around
the world.
