ECAI 2016 – International Conference 8th Edition [600858]

ECAI 2016 – International Conference – 8th Edition
Electronics, Computers and Artificial Intelligence
30 June -02 July, 2016, Ploiesti, ROMÂNIA
Smart Refrigerator
A next generation refrigerator connected to the IoT

Aurel -Dorian Floarea
Computer Science and Engineering Department
University POLITEHNICA of Bucharest
Bucharest, Romania
[anonimizat] Valentin Sgârciu
Computer Scienc e and Engineering Department
University POLITEHNICA of Bucharest
Bucharest, Romania
[anonimizat]

Abstract – The Internet of Things (IoT) refers to the set
of devices and systems that interconnect real world
sensors and actuators to the Internet. T his includes
many different systems, including smart objects, smart
monitoring devices, home automation systems,
smartphones (which are increasingly being used to
measure the world around them) and many more. This
paper presents the concept, architecture, building
process and functionality of such a device that
incorporates the systems stated above: a refrigerator
that is exposed as an IoT object and interacts with the
items stored within, gathers information about them,
process this information into releva nt data that is later
conveyed through an IoT platform to its owners, in
other words, a smart refrigerator.
Keywords: internet of things; cyber -physical systems;
RFID; smart refrigerator
I. INTRODUCTION
The growth of the number and variety of devices
that ar e connected to the internet and are collecting
data is incredibly rapid. In [1], Cisco conducted a
study in which it estimates that in 2010 the number of
devices connected to the internet overcomes the
human population and that there will be over 50
billio n devices connected to the internet by 2020.
Today, more than two billion consumers
worldwide use an internet connection in order to
browse content, send and receive information in the
form of email or instant messages, access various
multimedia resources and, among others, make social
networking. Furthermore, it is expected that the
internet will also serve as a global platform for
enabling new ways of working, interacting and living
by interconnecting physical objects or "things"[2].
The Internet of Thing s revolves around an
increased trend of machine -to-machine
communication. Its core philosophy is built on cloud
computing and networks of sensors gathering data.
The Internet of Things is mobile, virtual, and based on
instantaneous connections. Very soon i t is going to
add a "smart" prefix to everything in our lives.
In this paper, we propose adding the "smart" prefix
before an electrical appliance that can be found in
almost every house: the refrigerator. The contributions
of this paper are summarized as f ollows:
 We provide an in -depth analysis on how a
future household appliance would function  We identify the interrelations among the
device capabilities, physical layer, link layer
and network layer, and explain how these are
captured in our solution.
The r emainder of this article is organized as
follows. In section 2 we revise the related work in
terms of other attempts of smart refrigerators and other
smart things relevant to the idea proposed in this
paper. In section 3 we describe the concept and design
approach around our idea of a smart refrigerator. In
section 4, we review the system architecture. In
section 5, we evaluate the performance and utility of
the proposed solution by means of simulation. Finally,
we conclude the paper in Section 6.
II. RELATED WORK
The purpose of any smart appliance is to provide
an awareness of the users' needs, providing them with
a better home life experience without overpowering
them with complex technologies and intuitive user
interfaces.
Over the last years , the number of r esearch
activities on the topic of Internet of Things has been
constantly increasing, as perspectives of the
technology become more and more obvious. Below
we give an overview of the works, which are most
relevant to the topic of this paper.
In 2000, LG la unched the world’s first internet
refrigerator named Internet Digital DIOS, also known
as R -S73CT. The technology was the result of a
project that started in 1997 and staffed by a team of 55
researchers with a budget cost of almost $49.2 million.
The refri gerator had a TFT -LCD (thin -film transistor –
liquid crystal display) screen with TV functionality
and Local Area Network (LAN) port. It included a
LCD information window that featured an electronic
pen, data memo, video messaging and schedule
management fun ctions and provided information, such
as inside temperature, the freshness of stored foods,
nutrition information and recipes. Other features were
a webcam that was used as a scanner and tracked what
is inside the refrigerator, a MP3 player and a three –
level automatic icemaker. In addition, the electricity
consumption was half the level of conventional
refrigerators at the time and the noise level was only
23 decibels. This refrigerator was an unsuccessful
product because the consumers had seen it as
unnece ssary and expensive (costing more than
$20,000 mostly due to the huge investment made in

Aurel -Dorian Floarea, Valentin Sgârciu

2
the technology used) and that the problems solved by
it were obscure [3, 4, and 5].
Another intelligent refrigerator was patented in
2007 and refers to a refrigerator system to be used in
the pharmaceutical environment. Its main use was
storing pharmaceutical product containers. RFID was
used for product identification to uniquely identify
containers as they are added or removed. A variety of
details such as information about the manufacturer,
expiration date, inventory levels and mor e were
automatically retrieved from a data storage system
consisting of a local or remote database. A message or
warning was automatically delivered towards an
operator if the details indicate that there is a problem
with a particular pharmaceutical. It wou ld also monitor
and report any faults a ssociated with the refrigerator
[6].
A cost -effective space sensing prototype for an
intelligent refrigerator was described in [7]. The
proposed system was cost -effective and automatically
monitored the amount of space, and indirectly the
usage, in a refrigerator compartment. An extension to
the proposed design allowed the system to
automatically alert the homeowner of the refrigerator's
status through short message service (SMS). This
added dimension of automatic detection and
communication provides an attractive enhancement to
available commercial models of "smart" refrigerators.
III. CONCEPT
The proposed system revolves around the core
concept of product identification based on RFID
technology. We have considered in this paper a use
case that in the nearby future, all or most products
bought from the store will have a tracking RFID tag,
with information stored in a globa l level database by
all or most manufacturers, that will serve two
purposes:
 Firstly, the manufacturers will easily track
their products from the assembly line,
transportation route and finally to the store
shelf, giving them more information about
assembl y and transportation costs along with
information about product visibility in the
store itself
 Secondly, the consumer will be able to easily
access the above information, in time, for
himself, information that is not easily or at all
accessible to this day .
The concept itself is built on a set of well -defined
functionalities. Below we give an overview of them.
A. Identifying new products
Similar to the intelligent refrigerator system for
storing pharmaceutical product containers described
above, the proposed s ystem will scan all products at a
fixed time interval and identify whether or not a
certain product was already marked as found in a
previous scan. If not, then the product is considered
new and a web request is made to a global database
from which informa tion like product name, manufacturer and expiration date is fetched and stored
locally.
B. Identifying removed products
Using the same fixed interval scanning mechanism
described above, the system identifies products that
are missing from a scan, although the y were present
inside the refrigerator in previous scans. If this
scenario is repeated consecutively three times then the
system assumes that the product has been permanently
removed from the refrigerator and, based on the
functionality stated below, it wi ll add it to a shopping
list described in the following below.
C. Inventory and shopping list
Not everyone buys the same products over and
over again but there are those few products that are
essential in every household’s refrigerator (e.g. milk,
eggs, chees e, etc.). For this purpose, an Essential
Inventory was created in which the refrigerator owners
can add new products that they wish to have available
at all times or save an already available product in the
Essential Inventory. A product that is marked for the
Essential Inventory and has been permanently
removed from the refrigerator, as described above,
will be added to the shopping list.
The Shopping List is a web application that the
refrigerator owner can access on any internet
connected device (smartph one or tablet). The list of
products is a live feed because it does not interact
directly with the refrigerator and the data is stored off
site, in a remote database.
D. Triggered alerts
There are two types of alerts implemented in the
system so far. Firstly and most important is the alert
that signals if and when a specific product is about to
pass its expiration date. Secondly is the alert that
signals if a specific product has not been used in a
long period.
E. Information output
One information output source is the Shopping List
application, but this is not the only one. Inventory is
another web application has been developed in order
to easily see what products are in the refrigerator at
any time and even display what products were
available in past scans. Al ong with the available
product list other information is displayed, such as
product type, manufacturer and expiration date. Also,
any product can have multiple tags assigned to it, such
as a person based product reservation (e.g. "Dorian's
milk") or an eve nt based product reservation (e.g.
"Cakes for after dinner"). Similar to the Shopping List,
the Inventory application stores its data off site in a
remote database so this information is accessible from
any internet connected device in any remote location.
IV. ARCHITECTURE
As stated above, in our system we focus on two
main quotidian activities related to any refrigerator:
adding and removing products. Figure 1 provides a
functional block diagram of the proposed system. The
monitoring system consists of 5 main components, 4
of which are hardware components and the fifth is a

Smart Refrigerator

3
software component: the RFID kit which can be
broken down into the RFID reader and the RFID
antenna, the Arduino Breakout Kit, the Intel Edison
and finally the remote IoT platform used to a ggregate
the data received from the local system.
Figure 1. The proposed system architecture

A. The RFID kit
The Nano -UHF (figure 2) reader is a small, low –
power, low -cost RFID reader module that is designed
for short -range battery -powered mobile embedded
applications, such as cell phones, handheld readers,
printers, consumer electronics, or smart shelves. An
external antenna can be connected [8, 9].
The Nano -UHF is a single -antenna reader module.
The backscatter signal is demodulated to baseband
through a standard I/Q mixer stage and is filtered and
amplified. The baseband signal is then sampled and
process ed by a nanocontroller [8, 9].
The Nano -UHF reader is designed to interface with
another circuit board and communicate using a simple
3-wire serial interface [8, 9]. The reader's main
features and hardware specifications can be found in
[8, 9].
Figure 2 Nano -UHF Reader Module
Along with the Nano -UHF Reader Module, the
Full Patch UHF Antenna [10] was used in order to
improve the signal strength of the reader, and thus its
range covered up to 1 meter (figure 3).
Figure 3. Full Patch UHF Antenna

B. The Intel Edison and Arduino Brea kout Kit
A lot of recent IoT projects use a Raspberry Pi
module instead of a dedicated IoT designed computing
module. The problem with that is the fact that the
Raspberry Pi offers a lot more than the functionalities
actually needed for the device. It offers video and
sound output, network connection, U SB ports. This is
because the Pi aims at providing a computing terminal
for teaching purposes. The Edison, on the other hand,
is meant to be a deeply embedded IoT computing
module.
The Intel Edison module is a SoC (System on
Chip) that provides a 500MHz dual -core processor and
a 100MHz microcontroller.
Figure 4 Front and back of Intel Edison chip mounted on a Arduino
Breakout Board

Aurel -Dorian Floarea, Valentin Sgârciu

4
Its key features are the fact that it has integrated Wi -Fi,
BLE (Bluetooth Low Energy) and support for Linux,
Python, Node.js and Wolfram [11].
The Intel Edison Kit for Arduino (figure 4)
provides 20 digital input/output pins . It is designed to
be hardware and software compatible with Arduin o
shields designed for the R3 model of Arduino Uno .
C. The IoT platform
An IoT application platform, the backbone of any
IoT system, is centered on connectivity. All devices
that are components of an IoT system and are
connected to the Internet and each other share a
connection with an IoT application platform as it
provides a common link between those devices and
their data [12].
Nowadays , many companies offer IoT Platforms
and all of them offer some level of analytics reports
but the similarities end there as they are entirely
different software applications. For someone new in
this field it may not be easy to understand that this
term refers to a complete and mature IoT cloud
platform. More so, there are some software
applications that have been stretched to the point of
being called IoT platforms even when they describe
just an ele ment of a platform or even something
completely different [13].
There are 4 main types of platforms that are often
referred to as "IoT Platform" and are not because they
serve some other purpose :
 Connectivity / M2M pla tforms : as the name
suggests, these platforms focus on connected
IoT devices via telecommunication networks
but do not offer processing of data .
 IaaS (Infrastructure -as-a-service ) back ends:
they are more an interface to a platform than a
platform; they offer controls to configure the
hosting space provided and processing power
for different kinds of applications .
 Hardw are-specific software platforms: they
are proprietary software applications that
come along with some devices .
 Consumer /Enterprise software extensions:
they are usually enterprise software programs
of packages that offer some functionalities of
an IoT platform such as Microsoft Windows
10.
As stated above, the main goal of an IoT platform
is enabling and facilitating device communication.
Besides this main feature, an IoT complete platform
may feature alot of other importa nt functionalities that
are meant to improve either the IoT system
performance and capabilities, or the quality of life of
the platform consumer or client. A complete IoT
platform consists of the following eight features:
 Connectivity & normalizati on: unifies
multiple protocols and data formats under a
software interface ensuring continuous and
accurate data streaming and communication
with all devices .  Device management: monitors the connected
devices and ensures proper connectivity
between devices .
 Database: usually consists of scalable storage
of device d ata, implementing data volume,
variety, velocity and veracity requirements .
 Proce ssing & action management: consists of
event or rule based triggers enabling execution
of specific actions based on sensor data .
 Analytics: generates reports based on data
clustering extracting the essence out of the
data-stream from the IoT devices .
 Visualization: offers a human readable
graphical representation of patterns and trends
through different kind charts .
 Additional tools: consists of implementation
examples, tests and prototypes .
 External interfaces: offers ways to integrate
with 3rd -party systems via application
progra mming interfaces; also can consist in
software development kits in order to expand
the IoT implementation .
For the proposed system described in this paper,
Google Cloud Platform was used because it offers
tools to scale connections, gat her and make sense of
data, and provide the reliable customer experiences
that hardware devices require.
V. IMPLEMENTATION
The system presented has been implemented at a
proof of concept level, meaning that a refrigerator was
not actually used in the process and the system was
built and assembled on working bench, partly because
of the necessary funds required but also because in this
case, it would not bring an added value to this paper.
However, the full implementation was part of the
design process and was thought of. In order to have
the widest coverage inside the refrigerator the antenna
was to be plac ed and the top or bottom of the
refrige rator. Using a vertical implementation will
result in not requiring multiple antennas, one for each
level compartment, thus reducing the costs and the
entire system complexity. Figure 5 better describes the
hardware implementation inside the refrigerator.
Figure 5 Implementation of the RFID antenna inside the
refrigerator

Smart Refrigerator

5
VI. EVALUATION AND RESULTS
The goal of the proposed architecture is to enhance
the sustainability of IoT applications by exploiting
smart and reliable (networks of) things and by being
able to utilize a big number of heterogeneous device
platforms. The proposed architecture enables the
development of an environment for IoT applications
through cross -platform channels that incorporate
technologies for Data, Information, Things and
Decentralized Management [14].
As briefly described previously at a conceptual
level, the proposed system 's workflow is as follows.
The Nano -UHF Reader Module coupled with the
Full Patch UHF Antenna and triggered at 15 minutes
intervals by the Intel Edison processing chip scans the
refrigerator contents. If the scan reveals additional tags
that were not uncov ered by the previous scan then the
new tag codes are analyzed. For every new tag code, a
HTTP request is made to a global database that
contains information about all tag -equipped products.
If the request is successful then information about the
products i s returned, for instance the product type and
name (which is the main information displayed in the
Inventory and Shopping List web applications), the
manufacturer name and the expiration date.
Figure 6 Brief flow of the RFID scanni ng logic

This information is then sent to the IoT platform
for storage and dispersion into the two mentioned web
applications. If however the request is not successful
for any reason, an alert is triggered by the Intel Edison
and send to the IoT platform to be displayed in the
Inventory web application alerts dashboard. Within the
same scan cycle, if there are tags that are not present
and were present in the previously scan results then
these are marked with a removal flag in the local
Edison's database. These products are not removed
from the inventory at this point though. Only after the
fifth consecutive scan when these marked products are still not found inside the refrigerator, will these
products be marked as permanently removed from the
Inventory. At this point, the essential product
inventory described above is consulted and, if the
product type is found on the list then the removed
product is added to the Shopping List.
Another action that takes place while scanning the
products is the expiration date validation. If the
expiration date is within 3 days or has already passed
an alert is triggered and send to the Inventory web
application.
Figure 6 illustrates the graphical user interface of
both Inventory and Shopping List web applications.
Figure 7 Inventory and Shopping List web applications user
interface

VII. CONCLUSION
The dynamic rapidly changing and technology -rich
digital environment enables the provision of added –
value applications that exploit a multitude of devices
contributing services and information. As IoT
techniques mature and become ubiquitous, emphasis is
put upon approaches that allow things to become
smarter, more reliable and more autonomous.
This paper presented the concept, architecture,
building process a nd functionality of a smart
refrigerator, a future IoT component.
The ideas presented in this paper can be further
developed in several directions. First of all, further
research is needed to effectively overcome one of the
RFID's biggest weakness which represents the radio
wave s propagation through metal and liquids.
Secondly, the proposed solution did not touch any
of the security and privacy concerns that currently
affect the IoT paradigm development.

Aurel -Dorian Floarea, Valentin Sgârciu

6
Lastly, the proposed solution does not cover the
products stored in a refrigerator that are not container
in a RFID -equipped packaging like fruits of
vegetables.
Although the solution presented in this paper
works well in the context of a smart refrigerator, the
concept can be adapted and implemented in all
contexts that require managing and tracking storage
items or inventories in enclosed small and medium
sized areas. The added value of this paper is
represented by more than the technical solution and
context that were chosen, the concept itself being new
and bringing new opportunities and ideas to the IoT
industry. As a short example, a similar solution can be
implemented inside a wardrobe in order to assess what
clothes are available and what are the optimal matches
that can be wore in a specific day, base d on what was
removed from the wardrobe the day before, outside
weather forecast, desired color scheme and other
situational parameters. Many more implementations
can be found even outside of a home. This is why the
concept adopted in this paper is part of the added
value.
Realizing the vision of sustainable IoT applications
requires the enhancement of IoT technologies with
new ways that will enable things and objects to
become more reliable, more resilient, more
autonomous and smarter.
REFERENCES
[1] CISCO. Th e Internet of Things, Infographic. Available online
at http://blogs.cisco.com/news/the -internet -of-
thingsinfographic , 2011.

[2] D. Miorandi, S. Sicari, F. Pellegrini, I. Chlamtac, I nternet of
things: vision applications & research challenges, Ad Hoc
Networks (Elsevier) Journal (2012).

[3] http://www.telecompaper.com/news/lg -unveils -internetread y-
refrigerator –221266

[4] Andreas Jacobsson, Martin Boldt, Bengt Carlsson, A risk
analysis of a smart home automation system, Future
Generation Computer Systems , Volume 56, March 2016,
Pages 719 –733

[5] Corinne Belley, Sebastien Gaboury, Bruno Bouchard,
Abdeno ur Bouzouane, Nonintrusive system for assistance and
guidance in smart homes based on electrical devices
identification, Expert Systems with Applications, Volume 42,
Issue 19, 1 November 2015, Pages 6552 –6577

[6] https://www.google.com/patents/US7775056

[7] P. K. K. Loh, D. Y. H. Let, A cost -effective space sensing
prototype for an intelligent refrigerator, Control, Automation,
Robotics and Vision Conference, 2004. ICARCV 2004 8th,
798 – 803 Vol. 2

[8] http://www.tagsense.com/images/stories/products/uhf_readers
/Tagsense -Nano -Technical -Specification.pdf

[9] http://rfidstore.myshopify.com/collections/passive –
rfid/products/nano -uhf-reader -module

[10] http:/ /rfidstore.myshopify.com/collections/antennas/products/
full-patch -uhf-antenna

[11] https://www.arduino.cc/en/ArduinoCertified/IntelEdison
[12] Manuel Díaz, Cristian Martín, Bartolomé Rubio, St ate-of-the-
art, challenges, and open issues in the integration of Internet
of things and cloud computing, Journal of Network and
Computer Applications, Volume 67, May 2016, Pages 99 –
117

[13] IoT Analytics GmbH, White paper, IoT Platform, The central
backbone f or the Internet of Things, 2015

[14] Dimosthenis Kyriazis, Theodora Varvarigou, "Smart,
autonomous and reliable Internet of Things", 2013 (442 –
448), International Workshop on Communications and Sensor
Networks (ComSense -2013)