5G and Beyond

The Future of IoT

Parag Chatterjee, Robin Singh Bhadoria, Yadunath Pathak, Parag Chatterjee, Robin Singh Bhadoria, Yadunath Pathak

• Publisher: Chapman and Hall/CRC
• Year: 2022
• eBook ISBN: 9781000531862
• Topic: Computer Science
• Subtopic: Computer Networking
• Index: Computer Science

The Internet of Things (IoT) has seen the eventual shift to the "Internet of Everything" in the recent years, unveiling its ubiquitous presence spanning from smart transports to smart healthcare, from smart education to smart shopping. With the 5G rollouts across the different countries of the world, it raises newer perspectives toward the integration of 5G in IoT. For IoT-based smart devices, 5G not only means speed, but also better stability, efficiency, and more secure connectivity. The reach of 5G in IoT is extending in multifarious areas like self-driving vehicles, smart grids for renewable energy, AI-enabled robots on factory floors, intelligent healthcare services . . . The endless list is the real future of 5G in IoT.

Features:

• Fundamental and applied perspectives to 5G integration in IoT
• Transdisciplinary vision with aspects of Artificial Intelligence, Industry 4.0, and hands-on practice tools
• Discussion of trending research issues in 5G and IoT

As 5G technologies catalyze a paradigm shift in the domain of IoT, this book serves as a reference for the researchers in the field of IoT and 5G, proffering the landscape to the trending aspects as well as the key topics of discussion in the years to come.

Part I Fundamental Architectural Concepts for 5G and IoT

1 The Impact of Artificial Intelligence on 5G-Enabled IoT Networks

Ranjana Sikarwar

Amity University, Gwalior, India

Parth Wazurkar

Indian Institute of Information Technology (IIIT) Nagpur, Nagpur, India

DOI: 10.1201/9781003045809-2

Contents

1. 1.1 Introduction
   • 1.1.1 Artificial Intelligence: State of the Art and Prospects
   • 1.1.2 Important Subsets of AI
   • 1.1.3 Background of AI
   • 1.1.4 Current Research in 5G
2. 1.2 Role of AI and 5G in Digital Transformation across Industries
3. 1.3 Impact of Machine Learning for a 5G Future
   • 1.3.1 Categorization of Machine Learning Models for 5G Deployment
4. 1.4 Potential and Limitations of AI and Machine Learning for 5G
   • 1.4.1 Potential of AI
   • 1.4.2 Limitations of Using AI and ML
5. 1.5 Requirements and Key Enabling Technology in 5G IoT
6. 1.6 Artificial Intelligence Driven Cases for Real-Time Business and 5G IoT
   • 1.6.1 COVID-19, Digital Healthcare and the Role of 5G
   • 1.6.2 Real-World Business Use Cases for AI
7. 1.7 Conclusions
8. References

1.1 Introduction

The Internet of Things (IoT) technology, constantly growing to accommodate the needs of future IoT applications, currently uses existing 4G networks. IoT-centric applications such as augmented reality, which work through intelligent connections, require higher data rates, large bandwidth, increased capacity, low latency and high throughput in order to function quickly [15]. The exponential growth of IoT devices that generate massive amounts of data wirelessly has led to widespread investigation of high speed, enhanced bandwidth and low latency 5G cellular networks. 5G is widely expected to achieve enhanced end-user quality of experience (QoE) and higher data rates than 4G, as well as 1000 times greater system throughput and ten times greater spectral efficiency. Current IoT arrangements face many technical challenges, including security, compatibility and longevity, the large number of node connections and new standards. Higher data rates, low latency, efficient use of the spectrum and seamless connectivity between different networks are the most debated topics in IoT [15, 21].

Artificial intelligence (AI), with its basis in the disciplines of computer science, mathematics, and engineering, is a concept that simulates human intelligence in machines, using new methods, theories and the latest technological application systems. AI is used to analyze the bulk data produced by numerous IOT devices and make intelligent decisions accordingly. The aim of AI research is to solve real-world complex problems without human intervention so that machines can replace humans. Among the useful tasks that can be performed by smart systems using AI technology are diagnosing diseases such as cancer more effectively than doctors, or safe driverless vehicles. Artificial intelligence can be used to develop smart systems for office, business or home use [9, 19].

AI aims at developing an automated computer with software enabling it to think intelligently in the same way as humans can reason in an intelligent way. Machines can also learn and solve problems in an intelligent manner, hence the terms ‘machine learning’ and ‘machine intelligence’. Artificial intelligence has made real progress in numerous areas such as image recognition, natural language processing, medical image analysis, game playing and many more [14].

1.1.1 Artificial Intelligence: State of the Art and Prospects

Before developing AI programs, we need to know what intelligence is. Intelligence comprises perception, reasoning, learning, problem -solving, and so on. Figure 1.1 uses a Venn diagram to show elements of intelligence, with intelligence in the middle, surrounded by learning, perception, reasoning and problem solving.

Figure 1.1 Elements of intelligence.

Thus, the aims of program development are:

• To correlate perception and action smartly with huge knowledge of the world
• To know how to solve a real-world problem accurately and using cognitive functions
• To apply high-level logic to solution finding,
• To solve knowledge-based tasks and learn from mistakes
• To mimic human intelligence.
  1. Perception: The process of inferencing, acquiring, selecting, and organizing valuable information from unprocessed input. Humans use their experience, sense organs and the environment for perception. Intelligent machines undertake perception in a logical manner [3].
  2. Learning: The process by which new understanding can be acquired from different sources, such as books, life experience, teaching by experts, knowledge and skills gained through study. Learning increases a persons knowledge in new fields and areas.
  3. Reasoning: The act of thinking in a logical way to predict something, to judge or make decisions on real-world cases [1].
  4. Problem solving: The method of discovering a problem, analyzing it, then finding the best solution in the minimum time using generic or ad hoc methods [20].

1.1.2 Important Subsets of AI

The subfields of AI (see Figure 1.2) make machines work smarter, ranging from estimating the price of your ride in car booking apps through to self-driving cars [5]. These subfields are contributing to 5G by helping wireless networks to become proactive and predictive in nature.

Figure 1.2 Subfields of AI.

Machine learning (ML) uses different algorithms to make predictions from data. The 5G infrastructure will include many networks operating at different frequencies resulting in complex scenarios with multiple variables. ML tools will play a key role in outlining the patterns in these scenarios and finding unknown patterns by using machine-learning algorithms [14]. A principal use is image analysis in healthcare, including for the diagnosis of diseases. Popular machine-learning tools used today include Python, R, MATLAB, Spark and TensorFlow [16].

Deep learning learns all the human characteristics and behavioral databases to carry out supervised learning. Models, images, text or sound are classified through the processing and analysis of input data using different methods or algorithms to get the desired output. The classification or grouping of similar datasets (such as images or documents) to predict future events is known as predictive analysis. Among the popular algorithms used are neuro-evolutionary methods and gradient descent [6]. Models are trained continuously with huge labeled datasets along with multi-layered neural network architectures.

Artificial neural networks (ANNs) are the real brains of AI. They help to process huge bytes of data using edge computing in 5G networks. ANNs consist of layers of neurons called perceptron which are biologi...

Table of contents

• Cover Page
• Half Title page
• Title Page
• Copyright Page
• Contents
• Preface
• Editors
• Contributors
• Part I Fundamental Architectural Concepts for 5G and IoT
• Part II Applied Scenarios of 5G and IoT
• Index