May 2017
8
I
NterNet
-
oF
-t
HINGs
Making the Internet of Things smart,
secure, and power-efficient
By Clayton Cornell,
Infineon Technologies
In the IoT, Intelligent devices are
interconnected and AI algorithms
are being used to process the vast
amounts of sensor data that is being
produced. This exciting marriage of
IoT and AI requires state-of-the-art
sensors, security, and power delivery
to make it all possible.
n
The Internet of Things (IoT) has become
one of the most important technology trends
of recent years. IoT devices are showing up
everywhere and radically changing how we do
business and interact with the world around
us. Our offices, our homes, our factories,
our automobiles, and the things we wear or
carry in our pockets are all being equipped
with embedded electronic systems, sensors,
and even Artificial Intelligence (AI) software.
These IoT devices that were once a science fic-
tion dream are rapidly becoming an irreplace-
able part of our daily lives.
Devices equipped with intelligent semicon-
ductor components are forming the founda-
tion of the smart, secure, and power efficient
IoT that is developing around us. The “smart-
ness”of these systems is achieved by stretching
the boundaries of current design and tech-
nology. The intelligent devices (which were
stand-alone in the past) are interconnected in
ways that were not possible even a few years
ago and AI algorithms are being used to pro-
cess the vast amounts of sensor data that is
being produced. This exciting marriage of IoT
and AI requires state-of-the-art sensors, secu-
rity, and power delivery to make it all possible.
IoT and AI in the real world
There are many varying definitions and opin-
ions of just exactly what IoT is and how it
is related to AI. When you filter through
all of the details, you end up with a simple,
basic explanation. IoT is the networking of
cyber-physical objects that contain embedded
electronics to sense, compute, actuate, and
communicate. The AI in IoT devices processes
the sensor data from the embedded electron-
ics, analyzes it, and attempts to understand
the surrounding environment and events.
Given the incredible potential behind this
concept, it’s hardly a surprise that companies
all over the world are taking steps to combine
IoT and AI in new and creative ways. One
interesting example is in the emerging “smart
city”. It is rapidly becoming apparent to city
administrators that smart networked infra-
structure is playing an increasingly important
role in defining the world we live in. Design-
ers are rethinking common everyday objects
and adding new functionality or completely
redesigning them. Out of this are coming new
IoT devices such as smart street lamps. These
new street lamps are evolving into a highly
functional, connected, energy-efficient portal
that is forming the backbone of the smart city
infrastructure. It is this marriage of IoT and
AI that provides street lamps that can sense
approaching vehicles and brighten or dim the
light accordingly. The street lamps can sense
empty nearby parking places and communi-
cate this to navigation systems in cars. These
street lamps can also do things like provide
E-Vehicle charge ports and even monitor and
analyze the performance of the light itself and
inform a central monitoring systems when
they are in need of preventative maintenance.
IoT Street lamps with built-in AI are not just
fantasy though. This is a very real concept that
has been born out of collaboration between
Infineon, Intel, and a Munich based startup
company, eluminocity. Street lights with this
feature are already in place in various cities
around the world such as Munich, Chicago,
Oxford, and Eindhoven, with more installa-
tions to follow. IoT and AI is opening up new
markets and even enabling markets that pre-
viously have not been possible or even imag-
ined. The invention of the home computer,
the development of the internet, the progres-
sion to cloud computing; each step along the
way has paved the road to IoT. Combining IoT
with AI is opening up a whole new world of
possibilities.
Challenges and Solutions
While the devices themselves and the AI
algorithms are key to this emerging technol-
ogy, the increasing number of devices cou-
pled with the growing power requirements
and the critical need for security is present-
ing designers with some big challenges. The
power levels required for this new AI tech-
nology for example are simply staggering.
In order to match the processing power of
a human brain which consumes roughly 20
Watts, an AI system would need to perform
more than 38 thousand trillion operations per
second (38 PFLOPS) consuming more than
15 Megawatts of power in the process. Simply
delivering these levels of power is challeng-
Figure 1. The Internet-of-Things and the flow of information
