12.1 Introduction

Internet of Things (IoT) is novel standard that improves the traditional living style and convert it into advanced and technological life style [1]. This new emerging pattern simplifies our daily life by enabling the communication between sensors and electronic devices through the internet [2, 3]. IoUT is a new form of IoT, and this can be defined as underwater network of smart interrelated underwater objects. The concept of IoT was introduced in 1985 and IoUT was introduced in 2012 [4]. IoUT is a worldwide network of smart interrelated underwater objects that helps in exploring and monitoring the area below the water [5, 6].

Underwater Wireless Sensor Network (UWSN) has been provided as an efficient and promising network system for the Internet of Underwater Things (IoUT) [7]. There are several components which mainly exist in UWSN; one of its main components is sensors which are usually deployed under the water. These sensors can also be termed as nodes having acoustic modems distributed randomly in the shallow water. Different sensors are used for sensing different information about environment condition such as quality of water, water temperature, and pressure in the water and various phenomenon changes in the biological and chemical materials or elements [8]. Another essential component that is the part of UWSN is a sink node. The sink is responsible for receiving data from the sensor nodes present on the water surface [9]. Surface sink nodes have both radio and acoustic modems properties to deal with satellite connection and underwater nodes, respectively. These sinks might be any type of ship, Autonomous Underwater Vehicles (AUVs) or surface buoy. After successful reception of data by the sink, it forwards the data to the control monitoring center or any onshore base station with the help of radio channels. This monitoring center is responsible for controlling all the activities that can be performed in the shallow water and also control the working of the underground sensors. All the information gathered from the water areas is analyzed and collected by the monitoring center. Apart from this, AUVs are deployed in the water which majorly performs collecting and forwarding related tasks in the deep water [10].

The state-of-the-art IoUT lend a hand the researchers to explore water reservoirs such as large oceans, rivers, and lakes through numerous communication approaches which correlate stationary and mobile nodes below the water, on the water surface, and in the sky, as shown in Figure 12.1. Using embedded sensors and internet facility, these devices sense the environment, understand, and then respond accordingly. Each object which is deployed underwater can be accessed virtually that helps to get information like physical properties and historical information about the object. Due to the ubiquitous nature of the information, it can be accessed or managed in real-time through many different ways of communication like Human to Thing (H2T) and Thing to Thing (T2T).

This book chapter gives us brief introduction about the IoUT and organized as follows. Section 12.2 familiarizes us with definite characteristics of the IoUT. Following, Section 12.3 proposed IoUT architecture that is explained from technical aspect. Section 12.4 discusses about the different challenges in IoUT. After this, application scenarios are described in Section 12.5. Later, in Section 12.6, there is discussion about the new evolving technology machine learning (ML) and how this ML is helpful in study or analyses of underwater environment. Simulation comparison of three different localization techniques in IoUT is presented in Section 12.7. Section 12.8 is the conclusion of the chapter.

12.2 Characteristics of IoUT

Communication Technology: Radio and electronic waves do not work properly so acoustic waves are used for underwater communication. Radio waves does not propagate well below the water, and for using electronic magnetic waves, large-sized antennas are required and these waves gets attenuated with distance so acoustic waves are preferred in comparison of these waves.

Battery Power: Battery power is limited for underwater sensor nodes as it is difficult to charge these nodes or replace them. The need of batteries can be eliminated by using supercapacitors with ambient energy harvesting. This technique eliminates the need of batteries [11].

Network Density: A large number of devices communicate with each other in IoT for the proper functioning but it is very difficult in case of IoUT. As a smaller number of interconnecting digital devices are used in IoUT because deployment of devices is very tough in underwater. Localization Technique: In IoT, Global Positioning System (GPS) is used to find the location of devices but GPS does not work below the water so there exist different localization techniques categorized into range free and range based [12].

12.3 Architecture of IoUT

The proposed architecture of IoUT is very simple, as shown in Figure 12.2. It is a three-layer system architecture. Basic functionalities of these layers can be explained as follows.

12.4 Challenges in IoUT

The research in IoUT is slow because of the unique behavior of UWSN and so many challenges are associated with this IoUT which are explained below:

Transmission Media: In TWSN for communication, use of radio waves is found to be very helpful. But UWSN mainly depends on the communication based on acoustic rather than communication done by radio. Acoustic communication does not absorb quickly in water, but in the case of radio communication, they are very quickly absorbed in water. Therefore, the properties of both the communication are quite different, and hence, the properties of TWSN cannot be applied directly to the UWSN. Hence, the medium of transmission is one of the major challenges for the IoUT [4, 15].

Propagation Speed: The speed of propagation in UWSN is 200,000 slower than that of TWSN. In TWSN, speed of propagation of the radio channels is 300,000,000 m/s, but, as compared to UWSN which is around 1,500 m/s, that is quite slow as compared to TWSN. This raises a challenge which is called end-to-end delay and can be a challenging task for the IoUT [4].

Transmission Range: The transmission range in UWSN is 10 times longer as compared to TWSN. Environment under deep water to decrease the absorption level of signals by water the transmission has to be done at a low frequency level as if the frequency is low the transmission range is long. Due to this long transmission, there might some difficulties of interference and collision, and data packets may be loss due to that. Therefore, it is quite important to tackle these collisions and interferences that may be the one of the challenges for the IoUT [4].

Latency: As acoustic waves are used for communication in IoUT and this is thousand time slower in comparison of terrestrial IoT networks. So, real-time communication is restricted due to this low speed in IoUT. Though, the study on the growth of optical modems is still in the academic research phase [15].

Network Life: Underwater communication in IoUT is entirely depend upon the sensor communication which are battery operated. Strict environment below the water do not allow these batteries to be recharge so this affects the network life time [15].

Self-Management: This is one of the biggest challenges as network in IoUT have to manage all network operation itself without any human intervention [5].

12.5 Applications of IoUT

For last years, several researchers have proposed the various types of applications in IoUT, as shown in Figure 12.3. All the application [16] has been categorized into basic five types: (1) monitoring the environment, (2) prevention of disaster, (3) military, (4) exploring underwater, and (5) others.

Monitoring the Environment: Environment monitoring is the furthermost communal IoUT application that is used. This type of application is useful to monitor numerous parameters like quality of water, monitoring the temperatures, monitoring [17] thermal pollution, and monitoring pollution from various biological and chemical compounds. In addition, UWSN helps in monitoring of gas and oil [18]. It has been noted that the application of environmental monitoring has become more popular especially in demand for smart cities.

• In this paper, author provides a suitable method and architecture network for monitoring the quality of water; basically, the water quality is determined by the pH value of the water. This paper also presents sensor node design which helps in monitoring the pH value of the water [19]. It has been noted that pH value is a significant factor that helps in understanding the quality of water. Using Zigbee communication, this sensed pH value is transmitted wirelessly to base station.
• Ad hoc sensor networks have been projected for the monitoring of ocean pollution. This new approach applicable to short-range underwater communication instead of other long rage strategies [20]. The best part of the short-range communication is that it has the capability to reuse some of the acoustic networks and also helps to tackle some of the challenges generated by long-range communication. Also, modems used in short-range communication are cost-effective as compare to long-range [21] communication. In this paper, author also proposed synchronization protocol helps to ameliorate the quality index of the service (QoS) and pragmatically enhances the underwater sensor network.
• The authors discussed the latest inventions in wireless communication technologies which have formed a space for advancement in every domain of industrial applications associated to natural resource exploration, especially oil [22] and gas processing. Through this approach, the oil and gas pipeline monitoring solution are the brainchild of the author. The proposed approach is the most pragmatic solution for flawless examination [23] as well as provide pipeline health-related statistics with the help of wireless sensor networks.
• Many underwater monitoring applications were introduced by different authors. The authors outline underwater surveillance as a network of sensors commissioned undersea to observe the marine environments, features, characters, or any objects of interest [24]. The authors segregated and the underwater surveillance applications in three different categories, namely, (i) monitoring underwater explorations, (ii) water quality monitoring, and (iii) habitat monitoring and explained them.

Underwater Exploration: The IoUT, new approach can be used to discover lost-treasure beneath the water. It has also noticed that, in 1985, the titanic discovery is being benefited from the usage of AUVs [25]. Further, IoUT can also be used in fish tracking. Moreover, discovery of natural resources in water like metals, coral reefs, and corals can be benefited from the UWSN infrastructure [26] as well.

• The main aim is to create or develop an efficient technique for communication that can investigate in this large area of uninhibited oceans. A large-scale system that uses the UWSN have been proposed to study [27] the environment of the ocean. Also, various methods have been used like deploying of sensors having attached video capturing devices. With the help of this system, we can able to study the resources [28] present underwater.
• Authors also claim that all the existing advance technology have to be used in a careful manner, without disturbing the existing one, basically, we have to understand the basic functions of this technology in order to control the advance one for the discovering of oceanic world [29].

Disaster Prevention: Disaster prevention applications that saved thousands of precious lives are one of the most principal IoUT applications

[30]. Natural calamities, especially water-based are actually threatening. For instance, the Fukushima Daiichi nuclear disaster in Japan was commenced fundamentally by the tsunami following the Tohoku earthquake. IoUT is the best weapon to detect and predict flood, tsunami and earthquake [31].

• Perez et al. present a layout of a real-time dimensions for flood monitoring and prediction. Even, researchers arranged a real-world surveillance network with various real-time computation tools in Spain. The simulation [32] outcomes prove its energy efficiency and robustness for the long-term.
• The author explained the importance of an early threatening generation system founded on UWSNs to perform rescue operation for human lives. The warning disasters such as earthquake and tsunami [33]. The author addressed the principal physical layers challenges in establishing a UWSN system for early warning generation. The authors achieve that communication reliability, efficient resource management and low-power design will remain the significant obstacles for the UWSN-based warning created systems [34].
• Several strategies were addressed by authors for tsunami forecasting and recommend an effective layout for such applications. Seismic pressure sensor is utilized by suggested architecture to prognosticate tsunamis and transmit the sensed data by the diffusion routing protocol as well as explained modus-operandi behind this strategy.

Military: The defence forces do regular exercises to protect the motherland from external as well as natural threats. Underwater discoveries and monitoring scheme is also most important for military purposes. These apps have a tremendous capability for [35] further marine troops.

• The author investigates sensor distribution pattern and a pragmatic approach to enhance the monitoring domain for submarine identification. By fetching the water deepness, transmitting scope, and reduction into account the [36] suggested scheme utilizes Particle Swarm Optimization (PSO) to identify the precise locations of sensor nodes. The authors used both the NS3-based images and mathematical analysis to validate the performance of the suggested system.
• Submerged mine identification method patterns were investigated by Khaledi et al. Especially, they examine five towing-vehicle substitutes (i.e., one underwater vehicle substitute, two surface vehicles [37] substitutes, and two airborne vehicle substitutes) and two sonar options. Computing pattern to imitate the procedure of demonstrating submerged mines, and examine every of the pattern substitutes. The conclusion displays that the submerged automobile substitute absorbed the minimum amount of energy. They also accept the security, speed, fuel-saving, and chance of detection into count and offer the related examination to show the finest substitute.
• The authors suggest a unique UWSN framework for marine environment monitoring methods. In the suggested design, the sensor nodes are primarily located in face buoys; then, the sensor [38] nodes are decreased to various depth. The depths are selected by the recommended scheme for maximizing the monitoring coverage. Additionally, each node is implemented with various distinct varieties (e.g., radiation, acoustic, magnetic, and mechanical) of micro-sensors. Based on the interpretation of the data gathered from the micro-sensors, the researchers display a data mining–based design for the categorization of star ships, minefield, and divers.

Some Other Applications: Owing to the latest research and innovation in this domain, IoUT, nowadays, is playing a niche role in many applications such as localization applications, navigation, and sports. Especially, the localization application as GPS is not applicable for marine surveillance environment [39] that provides a valuable location to ships, underwater vehicles, divers, and swimmers that can be provided by using underwater sensors as location allusion points [40].

• The authors present a precise analysis, spotlighting the bio-mechanical investigation of swimming utilizing the inertial sensors underwater. In current years, sport [41] scientists, coaches, and swimming athletes have been struggling to succeed a “fraction of a second” improvement. The sensors perform an imperative role in sustaining performance intensification from a biomechanical point of view. This paper presents diverse existing techniques equipped with wearable [42] inertial sensors such as magnetometers, gyroscopes, and accelerometers to evaluate the biomechanics performance. The outcomes specify that the underwater inertial sensors are appropriate and compatible means for swimming biomechanical investigations.
• The gathered information could play a niche role only when referred to the location of the sensor for a wide range of applications; for example, navigation is such an application [43]. The authors recommend a multi-stage AUV-aided localization approach for UWSNs. They conducted a complete series of simulations to assess the performance of the supposed scheme based on few QoS parameters like localization coverage, accurateness, and communication overheads. The outcomes shows that good performance can be accomplished by appropriately selecting the communication range.
• The authors introduce an Anchor-Free Localization Algorithm, called AFLA, for underwater conditions. As utmost traditional underwater localization algorithms based on special underwater nodes [44] equipped with special types of equipment or AUVs can be used as “anchor nodes” to assist in finding the location underwater, while the anticipated AFLA does not require data from any particular anchor nodes. AFLA utilises the adjacent connection of normal sensor nodes to discover the underwater location [45].

12.6 Machine Learning

ML subset of artificial intelligence is widely used spectrum and used in many application areas. Based on data, ML is able to take decisions. Nowadays, ML can be used in different areas; it also helps us in oceanography or underwater activities. Different applications that include ML in oceanography are habitat monitoring, prediction of ocean weather and climate, coastal water monitoring, species identification, detection of oil spill and pollution, and marine resources management. Several real-world problems are solved by ML itself because of its capability. Oceans are gigantic, complex, and dynamic in nature. So, ocean data structure is increasingly becoming large and complex. Coastal area is also exposed to many natural calamities like coastal flooding and sea level rise. To avoid these calamities, an accurate and reliable tool is required that helps in forecasting seashore evolution. Traditional methods were costly and time consuming and sometime these methods are not helpful in analysis, so ML can be used. ML outperforms traditional methods as this is fast, high precise, and robust.

ML helps in many application areas mentioned above in IoUT. Neural network can be used to predict the sea level waves and surface temperature. Pollution monitoring can be done with the help of ML. MLP NN model was used by Del Frate et al. to detect oil spill on sea surface from synthetic aperture radar (SAR) images. Many types of species exist below the water and identification of these species is tough task. ML helps us in identification of these species. ML algorithms are trained for this identification using images, videos, and other type of data. These trained algorithms can interpret data and identify the different type of species exist in water. For management of underwater activities, habitat monitoring is required [46]. Algorithms in ML can be trained to find the matching variables, and using this algorithm, it is easy to identify appropriate habitat for particular species at any location. Meteorological forecasting can be performed using genetic algorithms in ML which helps in modeling rainy vs. non-rainy days. Sea level pressure can also be predicted with this MLP NN model. This ML can help in many application areas and can improve IoUT research in future by making use of precise models. Appropriateness of data set used for training plays a big role in success of ML technique. Quality, preciseness, and amount of data are critical components for the success of these ML applications [47].

12.7 Simulation and Analysis

In this section, performance analysis of different localization techniques Cluster-Based Mobile Data Gathering Scheme (CMDG), Diffusion Logarithm Correntropy Algorithm (DLCA), and Large-Scale Hierarchical Localization (LSHL) has been compared in IoUT based on two quality of service parameters: energy and average communication cost, as these both should be minimum for a good localization technique. Figure 12.4 shows the average communication cost of three different localization techniques with node mobility effects.

Figure 12.5 shows the energy consumption in each of the three techniques with respect to number of sensor nodes.

From these two parameters, we conclude that CMDG perform better from other two techniques. In communication cost, it performs 17% better than DLCA and 24% better than LSHL. In terms of energy, CMDG performs 14% better than DLCA and 19% better than LSHL.

12.8 Conclusion

This chapter familiarized with different basic technological concepts of IoUT. For the development of practical IoUT network, IoUT architecture is explained in detail. Few important challenges are also summarized here. IoUT behavior will results in establishment of this technology. Also, in end of the chapter, ML concepts were discussed and how this ML helps in many applications for underwater environment. Diverse methods exist in ML that can be useful for oceanographic applications that totally depends upon the data. Quality of oceanographic research can be improved by developing new accurate models in ML, and this will help in discovering unseen forms and learnings.