Professional Documents
Culture Documents
Chapter 3 - 5G + New Technology Innovative Application
Chapter 3 - 5G + New Technology Innovative Application
In the information age, the Internet has profoundly revolutionized the way
people work and live and injected strong impetus to global economic
development. The world is now stepping into the age of the digital economy
— a new form of economic and social development that uses digital
knowledge and information as a primary production element. Driven by
digital technology innovation, it relies on modern information networks,
HCIA-5G Learning Guide
enabling the deep integration of digital technologies and the real economy to
improve the level of digitization and intelligence in traditional industries and
accelerate the reconstruction of economic development and governance
models.
Over the next two to three decades, an intelligent society will become
reality, where all things are aware, connected, and intelligent.
1
"5G + Cloud + AI": Engine for the New Era of Digital Economy, CAICT
HCIA-5G Learning Guide
All of this will be made possible as ICT technologies continue to advance. ICT
infrastructure will be the foundation of an intelligent world and based on the
device-network-cloud IT architecture.
Devices include mobile phones, cameras, and sensors, and they sense the
physical world. Networks can be mobile (such as NB-IoT, 4G, and 5G) or fixed
(such as broadband and private lines). Cloud refers to cloud computing. In
the future, all device-sensed data will be transmitted to and pooled on the
cloud through the networks to form big data, based on which AI analytics will
be reality. For example, cameras and sensors in cities will collect various data
and send the data to cloud to form big data that can enhance smart security
while improving urban management efficiency.
This means that 5G enables smart city to usher in a new opportunity of
development by providing connectivity to everything anytime and anywhere.
With such connectivity, each person, thing, and organization in digital-twin
cities will be connected in real time, making them seamlessly integrated and
interactive with the physical cities and enabling all intelligent connections to
act as distributed super brains. The cities will become more intelligent to
HCIA-5G Learning Guide
fulfill various personalized needs 2. None of these will be possible without the
high bandwidth, low latency, solid reliability, and massive capacity enabled
by 5G.
In the future, 5G, cloud, AI, and IoT technologies will enable vertical
industries. 5G and cloud will become the foundation for information-driven
development. The device-network-cloud architecture will further enable top-
level applications required to improve overall efficiency. 5G capabilities will
play a significant role in achieving these goals.
IoT is an Internet for thing-to-thing connection. This means that its core and
foundation are still the Internet, with the connections extended and
expanded to things for communication and information exchange between
them4.
The concept of IoT dates back to Bill Gate's 1995 book titled The Road Ahead,
in which he mentions the idea about the Internet of Things. It attracted little
attention due to the development of wireless networks, hardware, and
sensors. In 1998, MIT creatively proposed an IoT-like concept, which was
then called the EPC system. In 1999, the Auto-ID Center in the US first
proposed the concept of IoT based on item coding, RFID technology, and the
Internet.
China launched the Made in China 2025 initiative in 2015, vowing to promote
the deep integration of digital and smart manufacturing with information
technologies such as IoT, cloud computing, AI, and smart manufacturing as
the main path for future development required to upgrade China from a
2
Source: 5G + Smart City White Paper
3
Source: International Telecommunication Union (ITU)
4
Source: baike.baidu.com
HCIA-5G Learning Guide
Layer Function
The device side belongs to the sensing layer, where sensors collect data and
are connected to the access and transport networks, such as the 2G, 3G, 4G,
NB-IoT, and 5G networks, over edge IoT gateways. Data is transmitted to the
cloud to form a big data cloud platform, and the platform uses the data to
HCIA-5G Learning Guide
1. Sigfox uses the Ultra Narrow Band (UNB) technology to maintain stable
data connectivity at a low power consumption. It supports a maximum
distance of 1,000 km and a capacity of up to 1 million IoT devices per
base station.
2. LoRa, short for Long Range, is maintained and managed by the LoRa
Alliance. It supports two-way data communications over a long distance
based on the physical layer, and features a high capacity and long
battery life. LoRa is best suited for automatic metering, smart home,
building automation, wireless warning and security protection,
industrial monitoring and control, and remote irrigation.
3. NB-IoT is a cellular narrowband IoT. Built on cellular networks, it
requires a cell bandwidth of only around 180 kHz and can be deployed
on top of legacy GSM, UMTS, and LTE networks, which is favorable to
reduce costs and ensure smooth upgrades. NB-IoT is an emerging
technology that focuses on global use cases requiring low power
consumption and wide coverage. It is built for low-speed services and
features wide coverage, huge connectivity, cost effectiveness, low
power consumption, and excellent architecture. NB-IoT has been
incorporated into 5G standards.
HCIA-5G Learning Guide
With NB-IoT incorporated into 5G standards, let's take a further look at the
four features of NB-IoT.
● Low costs
Huawei SingleRAN solution facilitates the upgrade and
reconstruction of legacy devices, helping cut construction and
maintenance costs. NB-IoT chips are specifically designed for IoT
narrowband and low-speed demand, supporting only single-
antenna transmission, half duplex mode, and simplified signaling.
As a result, NB-IoT chips cost only a few dollars.
● Low power consumption
NB-IoT uses the power saving mode (PSM) and extended
discontinuous reception (eDRX) for IoT services where small
packets are occasionally transmitted. With these features, IoT
devices enter the dormant state immediately after sending data
packets and wake up only when data transmission is required
again. As a result, IoT devices can be kept dormant for up to 99%
of their service time, achieving ultra-low power consumption.
1.2.1.4 5G+IoT
Based on market statistics by IDC, the global IoT connections had
approximated 30 billion by 2020, with the IoT market size expected to grow
by 16.9% per year to reach USD1.7 trillion by 2020. The breakthrough of 5G
provides a new opportunity to the IoT industry. Compared with 4G, 5G has
more powerful communication and bandwidth capabilities, meeting the
requirements of IoT use cases for high speed, stability, and wide coverage.
With 5G, many IoT applications that are still in the theoretical or
experimental stage may see brand-new opportunities to be quickly
implemented and exploited.
5G helps clear away the hurdles to the transmission speed and massive
connectivity of IoT. mMTC, as a 5G use case, will support one million
connections per square kilometer, achieving connectivity of everything. As
such, 5G IoT will enable a massive number of devices to access networks,
supporting smart cities, smart metering, and smart parking among many
other applications.
5G supports 1 ms of end-to-end latency required for remote control in
autonomous driving and industrial production. With 3G for remote control,
the braking distance is 3.3 m, given that a car travels at a speed of 120 km/h.
With 4G, the braking distance is 1.67 m, which is still not short enough to
ensure safety. By contrast, 5G reduces the braking distance to 0.033 m,
markedly improving safety.
HCIA-5G Learning Guide
Cloud VR requires higher mobility, larger bandwidth, and lower latency. For
example, entry-level VR needs a bandwidth of 100 Mbps and a latency of 10
ms, while to ensure ultimate VR experience, the bandwidth will be increased
to 9.4 Gbps but the latency reduced to 2 ms. Only 5G networks are able to
meet the requirements of ultimate VR experience.
5G+Cloud VR is only one application, and more will be explored in the future.
What is the size of big data? 2.9 million e-mails are sent worldwide in one
second, which would take one person 5.5 years to finish reading them day
and night, assuming that one email consumes one minute. 28,800 hours of
videos are uploaded to YouTube every day, which would take one person 3.3
years to watch them day in and out. On the Internet, a large amount of data
is generated every day. So, what is the use of such a large amount of data?
Let's look at big data technologies first.
Big data technologies are an umbrella name for technologies involving the
collection, storage, analysis, and application of big data, as well as those
related to dealing with the massive amount of structured, semi-structured,
and non-structured data through various tools to produce analysis and
prediction results. For example, GE uses a large amount of aircraft engine
running data for analysis to predict when the engine will be likely to
encounter faults. In the financial industry, Citibank uses IBM Watson to
recommend products to wealth management customers, and Bank of
America uses customer clicking data to recommend featured services.
HCIA-5G Learning Guide
The industry has not reached consensus on the definition of AI. Intelligent
machines are generally classified as four categories: "think like a human",
falling into the field of weak AI, with examples including Watson and
AlphaGo; "act like a human", falling into the field of weak AI, with examples
including humanoid robots, iRobot, and Atlas (by Boston Dynamics); "think
rationally", belonging to the field of strong AI, which is still not yet available
due to the bottleneck in brain science; and "act rationally", which also falls
into the field of strong AI.
When it comes to AI, one can never ignore machine learning and deep
learning. AI is a new science that studies and develops the theories,
methods, techniques, and application systems to simulate and extend human
intelligence. Machine learning studies how computers acquire new
knowledge or skills by simulating or performing the learning behavior of
human beings, and how they reorganize the existing knowledge structure to
improve its performance. It is one of the core research fields of AI. Derived
from the research of artificial neural networks, deep learning is a new field in
machine learning that simulates human beings to interpret various data,
such as images, sounds, and texts. For example, multilayer perception is a
type of deep learning structure.
HCIA-5G Learning Guide
The overall process of machine learning starts from data collection. There is
a famous saying in the industry: "Data determines the upper limit of machine
learning, and models and algorithms are just used to approach this upper
limit." Therefore, data is critical for the entire machine learning project as AI
modeling requires a large amount of data cleansing.
Data loss, uneven distribution, exception, and irrelevant impurities occur
more or less in data sets. This requires the collected data to be cleansed
through processing of missing and deviated values, data normalization, and
data conversion among other methods.
Data cleansing aims to ensure data normalized for subsequent feature
extraction — a process of extracting features of data sets and reducing data
dimensions.
After data processing, a proper machine learning model is selected for data
training. During model selection, different models are used to train the data
and the output results are compared to choose the best model for
evaluation and testing.
After the best model is selected from a model class, model evaluation is
performed to determine whether the model is over-fitting or under-fitting. If
HCIA-5G Learning Guide
the data fitting is not proper, parameters are adjusted to optimize model
deployment and integration.
1.2.4.2 5G+AI
AI is undergoing a third wave of development triggered by deep learning,
with remarkable progress achieved in data, computing power, algorithms,
and platforms. 5G and AI are growing into strategic technologies driving new
technological and industrial revolution, emerging as a high priority of new
infrastructure. The two emerging technologies enable and benefit each
other, accelerating the digital transformation together in both economic and
social areas.
Rapidly developing 5G networks and maturing synergy of 5G with cloud,
edge, and devices will help achieve full connectivity and data convergence
and reduce barriers to AI applications while boosting AI integration into
social and economic development. According to the forecast of an
authoritative organization, the AI market will be worth more than USD6
trillion by 2025, confirming that AI will become ubiquitous. Turning into a
core application on 5G networks, AI will accelerate the intelligent
transformation of 5G networks as well as cloud, edge, devices, and other
HCIA-5G Learning Guide
5
Excerpt from the speech by Dong Xin, the general manager of China Mobile, at the
WAIC Cloud Summit 2020
HCIA-5G Learning Guide
5G, cloud, AI, and IoT are the basic elements of future infrastructure. The
convergence of 5G and other new technologies is already changing society,
providing the basis for the digital economy.