Название: Machine Learning Modeling for IoUT Networks: Internet of Underwater Things
Автор: Ahmad A. Aziz El-Banna, Kaishun Wu
Издательство: Springer
Год: 2021
Язык: английский
Формат: pdf (true), epub
Размер: 10.9 MB

This book discusses how Machine Learning (ML) and the Internet of Things (IoT) are playing a part in smart control of underwater environments, known as Internet of Underwater Things (IoUT). The authors first present seawater’s key physical variables and go on to discuss opportunistic transmission, localization and positioning, machine learning modeling for underwater communication, and ongoing challenges in the field. In addition, the authors present applications of machine learning techniques for opportunistic communication and underwater localization. They also discuss the current challenges of machine learning modeling of underwater communication from two communication engineering and data science perspectives.

The first aim of this book is to shed light on the leading variable physical properties of water that affect the transmission of the core carrier for underwater communications, i.e., acoustic waves, and what leads to changes in its characteristics, e.g., intensity, resulting in inaccurate production of the deployed technologies that build on the assumption of fixed speed transmission of sound in underwater environments.

The second major aim is to investigate the application of machine learning (ML) techniques in settling diverse challenges that are encountered during deployment of underwater technologies, such as the Internet of Underwater Things (IoUT) and multi-modal underwater networks, and to capitalize on their merits. In addition, ML has the capabilities to treat the traditional underwater model-driven problems by considering the enormous measured data into appropriate data-driven problems and handle them to design a proper and adaptive behavioral modeling of these problems. This overcomes the main underwater problem where there is still no generic model that exists for the underwater environments because of the extremely harsh and fluctuating nature of such ambiance over the spatio-temporal domains since the ML techniques, unlike the theoretical systems, do not rely on explicit or certain propagation models or assumptions.

Various underwater environments are promising areas to deploy recent innovative applications and technologies such as IoUT and multimodal underwater networks. IoUT is a recent category of the Internet of Things (IoT) technology that extends the operation of UWSNs. In addition, IoUT could employ cloud computing platforms to assist in the communication process between other components of the network. Moreover, IoUT is expected to be extensively deployed in various underwater environments to cover numerous underwater monitoring and actuation applications toward smart worldwide networking of underwater devices.

The book provides IoUT network and node structure and the ML modeling for underwater communication in Chap. 1, the key physical variables of water and their interrelationships in Chap. 2, an example of channel modeling and an adaptive transmission framework for underwater networks in Chap. 3, two examples of the positioning systems in Chap. 4, and application of the decision tree as a classifier and dynamic modeling using neural networks for underwater techniques. The book also proposes in Chap. 6 some challenges faced by underwater communication and some glimpses of the solutions from both communication and data science perspectives.