Comment
The future of smartphones lies in AI, mmWave
We could be forgiven for thinking
that the smartphone has run out of
room to innovate. However, in spite of
smartphones becoming a commodity
item in many instances, we are still
seeing areas where manufacturers can
differentiate their products. These include
sound, camera, video and biometrics.
In the future we can look forward to
mmWave communications with the advent
of 5G and advances in AI.
The latest iPhone X incorporates face
recognition technology (Face ID) taking
biometrics to a new level, which enables
the phone to essentially ‘recognise’ the
owner. Enabling all these new capabilities
is the A11 Bionic mobile processor, the
brains of the iPhone X, that not only
boosts performance and GPU capabilities
but incorporates a neural engine.
The neural engine is basically a pair of
processing cores for executing specific
machine learning algorithms that enable
advanced features in the iPhone such as
Face ID, Animoji, and augmented reality
apps. This is the beginning of a trend of
putting AI into the phone or at the edge
of the network rather than centralised in
the cloud. So far AI is done centrally in
the cloud, but this involves moving huge
amounts of data around, a lot of it being
private and sensitive. By putting AI on the
smartphone, Apple is taking the approach
where AI is done at the edge – protecting
sensitive data while reducing data flow
across already overloaded networks.
Apple is not alone here. At IFA 2017
in September 2017, the CEO of Huawei
Consumer Business Group Richard
Yu revealed the company’s vision for
the future of artificial intelligence in
smartphones, combining the power of the
cloud with the speed and responsiveness
of native AI processing.
Key to this is the Kirin 970 mobile
processor which uses a 10nm advanced
process and is powered by an 8-core
CPU and a new generation 12-core GPU.
The 5.5 billion transistor chipset, only
one square centimetre, also features
a dedicated Neural Processing Unit
(NPU), making it Huawei’s first mobile
AI computing platform. Compared to a
quad-core Cortex-A73 CPU cluster, the
Kirin 970’s new heterogeneous computing
architecture delivers up to 25x the
performance with 50x greater efficiency,
claims the company. In a benchmark
image recognition test, the Kirin 970
processed 2,000 images per minute.
Qualcomm is also targeting mobile
AI. Recerntly Qualcomm Technologies,
Inc., and SenseTime Group Limited
announced their collaboration on AI and
machine learning (ML) for future mobile
and IoT products.
SenseTime will contribute ML models
to the parnership while Qualcomm
will offer algorithms with Qualcomm®
Snapdragon™ premium and hightier
platforms, which offer advanced
heterogeneous computing capabilities for
client based AI. The companies expect to
drive the popularity and development of
on-device AI in areas such as innovative
vision and camera-based image
processing.
PLASMA ANTENNAS
Turning to 5G, mmWave on the
smartphone will poise huge challenges.
A lot of progress is being made but one
stumbling block is the antenna. Here,
a company based in the UK, Plasma
Antennas, the inventors of plasma
silicon technology, recently announced
its mmWave Plasma Silicon Antenna
(PSiAN) for use in smartphones and
other consumer electronics, delivering
high throughput with low-latency and
utilizing directional beams that generate
less interference and maximize energy
efficiency.
Millimetre wave connectivity for
smartphones faces significant problems
as the signals are easily blocked by
fingers, hands, heads and bodies.
Plasma Antennas patented Plasma
Silicon technology has a range of
unique properties that mitigate these
effects. When used in combination with
distributed radiating elements, Plasma
Silicon Devices (PSiD’s) can be used as
a combination switch and beam former
to utilise only elements that are able
to receive and transmit line of sight or
reflected signal. The result is, whether held
in the left or right hand, against the user’s
head, in portrait for video calls, surfing and
email, in landscape for video, and even
with both hands, for gaming the system
will cope and continue to provide the low
latency and high throughput necessary for
photo realistic AR/MR/VR, 3D gaming and
similar demanding applications.
Plasma Silicon Antennas (PSiAN) are
steerable and dynamically reconfigurable
and use low power with a tiny footprint
(10 mm at 28 GHz). The beams are
formed and steered from a single device
eliminating phase shifters and reducing
computation with no need for calibration.
The technology also boasts very low loss
reducing and consolidating amplification.
Plasma Antennas recently modelled
Plasma Silicon corner antennas as
replacements for array modules for our
device manufacturer customers and their
silicon supplier.
SMARTPHONE MARKET GROWING
Is the smartphone market really
stagnating? Market research firm GfK has
found that global smartphone demand
reached 367 million units in 3Q17, up
three percent year-on-year. This is the
highest 3Q tracked by the firm, in terms
of absolute demand, despite moderating
growth. The record demand was primarily
driven by Latin America (up 11 percent
year-on-year) and Central and Eastern
Europe (up nine percent year-onyear).
Global average sales price (ASP)
continued to increase in 3Q, rising by a
record seven percent year-on-year.
Arndt Polifke, global director of
telecom research at GfK, comments,
“Although unit sales may be down in
some regions, the increase in ASP reveals
the fantastic opportunity to grow the
value of the smartphone market.”
“At the same time, premium
features are increasing in importance
to consumers, so we expect to see
more emphasis on water and dust
protection, battery power and memory,
high resolution sound, camera and video
capabilities, bezel-less design and even
biometric sensors on new launches.”
By Jean-Pierre Joosting, Editor, MWEE
6 MW November - December 2017 www.mwee.com
