IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 16, NO.

7, JULY 2020 4297

Underwater Internet of Things in Smart Ocean:

System Architecture and Open Issues
Tie Qiu , Senior Member, IEEE, Zhao Zhao, Tong Zhang , Member, IEEE,
Chen Chen, Senior Member, IEEE, and C. L. Philip Chen , Fellow, IEEE

Abstract—The development of the smart ocean requires characteristics in many fields, such as the smart home [3], [4],
that various features of the ocean be explored and under- transportation [5], [6], healthcare [7], [8], industrial automa-
stood. The Underwater Internet of Things (UIoT), an ex- tion [9], [10], and emergency response [11], [12]. For human, the
tension of the Internet of Things (IoT) to the underwater
environment, constitutes powerful technology for achieving ocean not only has a large amount of oil, gas, and fish resources,
the smart ocean. This article provides an overview of the but is also an important channel for the international trade and
UIoT with emphasis on current advances, future system ar- a potential energy resource, such as tidal energy and the kinetic
chitecture, applications, challenges, and open issues. The energy of the ocean currents [13]–[15]. However, there remains
UIoT is enabled by the most recent developments in au-
a lack of intelligent and convenient applications for the ocean
tonomous underwater vehicles, smart sensors, underwa-
ter communication technologies, and underwater routing because of the complexity of the underwater environment and the
protocols. In the coming years, the UIoT is expected to expense of the underwater equipment. As an extension of the IoT
bridge diverse technologies for sensing the ocean, allowing in the underwater environment, the Underwater IoT is increas-
it to become a smart network of interconnected underwater ingly becoming a powerful technology for developing the smart
objects that has self-learning and intelligent computing ca- ocean [16], [17]. Here, Underwater IoT is defined as UIoT that is
pabilities. This article first provides a horizontal overview of
the UIoT. Then, we present a five-layer system architecture a smart network with self-learning and intelligent computing ca-
for the future UIoT, which consists of a sensing, commu- pabilities. It can sense, monitor, and identify underwater objects
nication, networking, fusion, and application layer. Finally, by wired or wireless communication for building smart ocean.
we suggest the current challenges and the future UIoT re- Although the unique characteristics of the ocean can bestow
search trends, in which cloud computing, fog computing, many benefits on humans, they also restrict the development of
and artificial intelligence are combined.
the UIoT. Hence, the model and design of the land-based IoT
Index Terms—Artificial intelligence, cloud computing, cannot be adopted directly by the UIoT. For instance, ocean
fog computing, smart ocean, system architecture, Under- currents constitute a nonnegligible issue for the deployment of
water Internet of Things (UIoT).
the UIoT [18], [19]. Underwater node movement in the UIoT
caused by ocean currents often affects network coverage and data
I. INTRODUCTION transmission quality [20]. Moreover, the mode of the underwater
acoustic communication (UAC) seriously restricts the efficiency
N INCREASING number of physical objects are rapidly
A being connected to the Internet, realizing the idea of the In-
ternet of Things (IoT) [1], [2]. In recent years, the IoT has devel-
of data transmission because of the UAC characteristics such as
high cost, narrow bandwidth, high bit error rate, slow transmis-
sion speed, and high energy consumption [21], [22]. In addition,
oped rapidly and enabled numerous applications with different
the battery power of UIoT sensor nodes is severely limited, and
node batteries cannot be conveniently recharged owing to the
Manuscript received June 12, 2019; revised September 9, 2019; ac- limitation of seawater corrosion and seawater pressure [23],
cepted October 1, 2019. Date of publication October 10, 2019; date of
current version March 17, 2020. This work was supported in part by
[24]. Therefore, the future UIoT should connect underwater
the National Natural Science Foundation of China (NSFC) under Grant entities intelligently, as does the existing Internet of Underwater
61672131 and Grant 61571338, in part by the Innovation Foundation of Things [25]. More importantly, artificial intelligence and fog
Tianjin University, and in part by the Tianjin Key Laboratory of Advanced
Networking (TANK). Paper no. TII-19-2481. (Corresponding author: C.
computing need to be combined to provide the UIoT network
L. Philip Chen.) with certain self-learning and intelligent computing capabilities
T. Qiu and Z. Zhao are with the School of Computer Science and so that it can be adapted to the complex underwater environment
Technology, Tianjin University, Tianjin 300350, China (e-mail: qiutie@
ieee.org; zhaoz526@foxmail.com).
and meet the different needs of marine applications.
T. Zhang and C. L. P. Chen are with the School of Computer Science An UIoT model, shown in Fig. 1, usually includes underwater
and Engineering, South China University of Technology, Guangzhou sensing and transmission modules (underwater sensor nodes
510006, China (e-mail: tony@scut.edu.cn; philip.chen@ieee.org).
C. Chen is with the State Key Laboratory of Integrated Service
and surface nodes), underwater computing and transmission
Networks, Xidian University, Xi’an 710071, China (e-mail: cc2000@ modules [autonomous underwater vehicles (AUVs)], surface
mail.xidian.edu.cn). computing and transmission modules [surface base station (BS),
Color versions of one or more of the figures in this article are available
online at http://ieeexplore.ieee.org.
surface ships, and surface nodes], and coastal control modules
Digital Object Identifier 10.1109/TII.2019.2946618 (seashore BS and seashore control center). A simple operating

1551-3203 © 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.

Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.
 4298 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 16, NO. 7, JULY 2020

Fig. 1. Model of UIoT.

process of the UIoT is described later. A large number of The challenges and open issues of the UIoT are presented in
valuable ocean data are collected by underwater sensor nodes Section IV. Section V concludes this article.
at first. Then, the data are completed corresponding data fu-
sion and intelligent computing when transmitted by underwate
nodes and AUVs. Underwater nodes and AUVs use a variety II. CURRENT ADVANCES OF THE UIOT
of underwater communication technologies to transmit data to The functions and structure of the UIoT are similar to those of
surface computing and transmission modules. Some underwa- the land-based IoT. In addition, compared with the IoT, the UIoT
ter valuable data are analyzed by edge servers on surface in also has its own unique network characteristics; for example, it
surface for the underwater network and marine applications is a dynamic network, its density is sparse, it is a large-scale
(exploration platform and deep range, etc.), and the rest are network, and its underwater communication channels are of poor
transmitted to Internet cloud servers or seashore control centers quality. These characteristics make the UIoT a new-type net-
via radio communications. Seashore control centers generate work, which requires further study. In this section, we introduce
a series of intelligent decisions based on the data collected and summarize the current advances in the development of the
from the ocean and Internet cloud servers to facilitate human Internet of Underwater Things to help researchers understand
activities in the ocean. Furthermore, underwater wireless sensor the UIoT research focus more concretely.
networks (UWSNs), which can provide ocean information and In [21], the main differences between the UIoT and the IoT
improve users ability to monitor and forecast events in the were introduced. Furthermore, a three-layer system architecture
underwater environments, are also an important part of the UIoT for the UIoT was described, which includes the application
[26], [27]. layer, network layer, and perception layer. However, this system
The outline of the contributions of this article relative to the architecture is too simple for forming an actual UIoT because
recently published literature in the field can be summarized as it lacks the consideration of the characteristics of the under-
follows. water communication and the role of cloud computing and fog
1) We propose a five-layer system architecture for the future computing in the UIoT. In addition, the authors proposed some
UIoT based on the characteristics of the IoT and the important application scenarios of the UIoT that illustrate the
requirements of underwater applications. interaction of its components. Finally, the challenges that the
2) We summarize the challenges faced by the UIoT based UIoT presents were identified and analyzed. The authors of [28]
on our analysis of the underwater environment, as well as also provided an overview of the UIoT. They introduced and
on IoT issues. classified the practical underwater applications. In addition, they
3) We discuss the research trends in underwater communica- noted the differences between UWSNs and terrestrial wireless
tion, underwater networking, and cooperative computing sensor networks. Finally, they examined and evaluated the model
technology for the UIoT. of underwater acoustic channels. In [29], the challenges and
The remainder of this article is organized as follows. Cur- routing algorithms of the UIoT are summarized and analyzed.
rent advances related to the UIoT are discussed in Section II. In these studies, cloud computing, fog computing, and arti-
Section III describes the future UIoT system architecture. ficial intelligence were not combined with the UIoT. However,

Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.
QIU et al.: UNDERWATER INTERNET OF THINGS IN SMART OCEAN: SYSTEM ARCHITECTURE AND OPEN ISSUES 4299

these technologies are beginning to be widely used in the IoT.

It is universally known that the UIoT employs a large number
of heterogeneous devices, such as underwater nodes, surface
ships, and underwater robots, and generates a very large amount
of underwater perceptual data. Furthermore, the extraction of
knowledge from this data requires fast and accurate complex
computations using cloud computing and fog computing [2].
Compared with the terrestrial environment of the IoT, the com-
plexity of the ocean environment renders the topology of the
underwater network of the UIoT vulnerable to damage; hence,
artificial intelligence is required to optimize the network topol-
ogy. That is, the UIoT can be more practical for the underwater
network construction by using cloud computing, fog computing,
and artificial intelligence.
Specific studies of the UIoT are as follows. In the research
described in [16], the MAC-layer packet scheduling for the UIoT
was studied. Using spatial-temporal reuse of the spectrum, in
this article, a sender–receiver role-based scheduling protocol
with high computation utilization and high energy efficiency
was designed to support the transmission of the data of smart un-
derwater objects. For monitoring the underwater environment,
in [25], the design of an integrated smart environment system
framework based on the UIoT and big data was presented.
This framework includes five subsystems: remotely operated
vehicles equipped with water quality sensors, a portable water
quality monitoring system, a coral reef monitoring system, a
big data analysis system, and wireless mesh network access. Fig. 2. The system architecture of UIoT.
In [21], the authors proposed an enhanced channel-aware routing
protocol to optimize the data transmission and energy efficiency
five-layer UIoT system architecture, which consists of two parts,
in the UIoT network layers. This routing protocol can reduce
underwater and nonunderwater, as shown in Fig. 2. The UIoT
the UIoT energy consumption and prolong the network lifetime.
system architecture includes an application, fusion, networking,
To realize efficient and secure utilization of network resources
communication, and sensing layer. Each layer of the system
in the UIoT system, a secure cloud-based solution for UIoT
architecture has an independent function and scalability. In
real-time monitoring and management was provided in [30].
addition, we will discuss the application of cloud computing,
First, the design of an energy-aware efficient framework for
fog computing, and artificial intelligence in this UIoT system
the UIoT system was described. Then, the solution was defined
architecture. These technologies have great impact on the devel-
as an enhanced attribute-based encryption scheme for effective
opment of the IoT. For cloud computing, in [35], it is defined as a
user attribute management. In [31], the management system role
model for enabling ubiquitous, convenient, on-demand network
and the integration of the terrestrial system with the constraint
access to a shared pool of configurable computing resources that
environment were presented, as well as an analysis of the
can be rapidly provisioned and released with minimal manage-
management system based on the high-level system architecture
ment effort or service provider interaction. As an extension of
model usage for the UIoT.
cloud computing, fog computing is more time sensitive, which
Although studies have been conducted on the UIoT and
is a highly virtualized platform that provides compute, storage,
UWSNs have been studied extensively, the research and design
and networking services between end devices and traditional
of the UIoT remain in the initial stage when compare with the
cloud server [36]. Machine learning algorithms in artificial
IoT. We hope that these studies on the UIoT can provide more
intelligence can solve several challenges of the UIoT, such as
references for future researchers of the UIoT.
objects targeting, event detection, data transmission, network
security, and quality of service. Machine learning is described
III. FUTURE UIOT SYSTEM ARCHITECTURE as the adoption of computational methods for improving the
machine performance by detecting and describing consistencies
The UIoT is a complex system comprising multiple hetero-
and patterns in training data [37]. In the following, we describe
geneous networks; hence, a flexible layered system architecture
the five layers of our proposed UIoT system architecture.
is critically needed. Many different IoT system architectures
based on the analysis of the needs of researchers and industry
A. Sensing Layer
exist [32]–[34]. However, few UIoT system architectures have
been proposed. A basic system architecture of the UIoT was In the sensing layer, the various sensors, such as tempera-
proposed, which has three layers: the application layer, network ture, flow velocity, salinity, and motion sensors, and cameras
layer, and perception layer [25]. In this article, we propose a provide meaningful sensing data and identification data for
Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.
 4300 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 16, NO. 7, JULY 2020

various underwater applications to supply records and assist designed by including UAC and underwater optical communica-
in decision making. Compared with sensor nodes of the IoT, tion (UOC). It even can be built by a set of UAC modules working
underwater sensor nodes of the UIoT need to have the ability on different bands. Because they have some complementary fea-
of self-target recognition for initially validating the value of the tures such as low-frequency acoustic communications achieve
data in complex and changeable underwater environment. For kbit/s transmission rates over ranges of a few kilometers, and
example, a turbid underwater environment can cause a large high-frequency acoustics tops tens of kbit/s over up to a few
number of worthless image data to be generated. It is needed hundred meters.
to reduce energy consumption of worthless data transmission
because the cost of transmitting data in underwater is huge and C. Networking Layer
underwater sensor nodes are often given more computing power.
In the UIoT, the networking layer is used to construct an
The underwater self-target recognition of underwater sensor
efficient topology for transferring the data produced by the
nodes can be realized by many machine learning algorithms,
sensing layer to the fusion layer through underwater multimodal
such as neural network [38], support vector machine [39], deep
channels and radio communication above the water. The UIoT
learning [40], etc. In the UIoT, the sensors are deployed in
network model of above water part is similar to that of the IoT.
monitoring areas, and a self-organizing topology, which is re-
Here, we focus on the underwater network part of the networking
quired by the dynamic underwater environment, is constructed.
layer of the UIoT. The underwater network model of the UIoT
A typical UWSN system includes underwater sensor nodes,
needs to be presented that provide an efficient and reliable data
surface sink nodes, and surface management nodes. The sensing
transmission capacity for nodes, because of the poor commu-
data from underwater sensor nodes are transmitted to surface
nication channels, unfixed node location, and limited energy
sink nodes by multihop links quickly and reliably. Then, the
supply [51], [52]. In the underwater network part, underwater ve-
sink node processes or transmits the data to the monitoring
hicles and AUVs, as two super nodes of the underwater network,
center. Users manage the sensor network and issue monitoring
can significantly improve the underwater network performance,
task commands through the surface management nodes. Fur-
in terms of robustness and connectivity [53], [54]. Meanwhile,
thermore, underwater sensor nodes are expensive at present, so
in order to improve the efficiency of data transmission in un-
their maintenance and upgrading need further attention with the
derwater, the UMCS networking strategies using a variety of
increasing worldwide attention given to the ocean.
underwater communication technologies and underwater nodes
are becoming a research hotspot. However, the demand for
ocean big data transmission in underwater is still not fully met.
B. Communication Layer Therefore, the future UIoT will need a high data transmission
capacity to forward the big data to fusion layer. Machine learning
The main function of the UIoT communication layer is to
is widely used in many fields because of its excellent analytical
allocate the corresponding underwater communication technol-
ability [55], [56]. It is a potential tool to further improve network
ogy to be used for data transmission based on business re-
performance because a variety of machine learning methods
quirements, task categories, and data importance. Because radio
can learn network state and environment parameters adaptively,
communication, which is usually used in the IoT, cannot be
such as reinforcement learning [57], broad learning [58], deep
adapted to the underwater environment, the underwater part of
learning [59], decision tree [60], and self-organizing map [61].
the UIoT has many other communication modes with different
Hence, a self-learning and self-organizing routing algorithm
levels of performance, such as UAC [41], [42], optical wireless
based on learning the changes in the underwater environment
communication [43], [44], and magnetic induction communi-
should be adopted to further optimize the underwater network
cation [45]–[47]. However, on the comprehensive performance
model of the UIoT.
of transmission efficiency and reliability, any single underwater
wireless communication technology cannot be compared with
that of radio communication. The communication distance of the D. Fusion Layer
UAC is long, but its bandwidth is narrow in contrast to that of The fusion layer of the UIoT is mainly composed of cloud
wireless optical communication. In the underwater environment, server and edge server. With the combination of fog computing
the transmission stability of the magnetic induction communi- and cloud computing technology, the UIoT can quickly and
cation is better than that of the wireless optical communication accurately handle large amounts of heterogeneous data from
and UAC. In addition, the communication rate of the UAC is different underwater sensors and sensing devices. Meanwhile,
proportional to the frequency, whereas the communication dis- the fusion layer will receive and process data from other layers
tance is inversely proportional to the frequency [48]. Therefore, to optimize network resources and improve network operation
in order to further promote the development of the UIoT, these efficiency. The cloud servers of the fusion layer have a powerful
underwater communication technologies need be combined to analytical computing capacity and can also make decisions
form many underwater multimodal communication systems related to human ocean activities and provide the performance
(UMCSs) with different characteristics for various UIoT ap- level of underwater equipment based on analytical results [62],
plications. An UMCS is defined as the system encompasses [63]. The architecture of cloud of the IoT usually is divided into
any set of nonmutually interfering underwater communication four layers: datacenter, infrastructure, platform, and application.
technologies, which may have various advantages from different In addition, cloud services can be grouped in the following three
communication technologies [49], [50]. For an UMCS, it can be main categories: software as a service, platform as a service, and
Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.
QIU et al.: UNDERWATER INTERNET OF THINGS IN SMART OCEAN: SYSTEM ARCHITECTURE AND OPEN ISSUES 4301

networks dynamically in the UIoT. Edge servers of the UIoT

are usually deployed on the water surface or even underwater
in order to get perceptual data and make decisions quickly.
With the increasing level of heterogeneity and quantity of ocean
data, smarter decision making using effective fog computing and
cloud computing is required in the fusion layer.

E. Application Layer
The application layer of the UIoT uses the ocean data pro-
cessed by the fusion layer to provide convenient and intelligent
services to users according to their requirements. Various intelli-
gent functions based on machine learning are adopted to enable
application managers to better address user needs and to improve
user comfort in using the UIoT applications, for example, smart
Fig. 3. Fusion layer service architecture of the UIoT.
language recognition, smart image recognition, and smart pref-
erence recommendation, etc. These machine learning techniques
includes hidden Markov model, discriminative learning, struc-
infrastructure as a service [64]. The service and architectural
tured sequence learning, Bayesian learning, and deep learning.
models of cloud in the UIoT are similar to those of cloud in
Moreover, an important responsibility of the application layer is
the IoT. The fusion layer service architecture of the UIoT is
to protect user privacy security. The cloud servers can remotely
shown in Fig. 3, it consists of five modules: infrastructure service
control underwater intelligent nodes and robots based on the
layer module (ISL), platform service layer module (PSL), cloud
analytical data results [69], [70]. Currently, the main applications
service layer module (CSL), fog service layer module (FSL), and
of the UIoT are as follows: environmental monitoring, resource
application service layer module (ASL). The ISL is mainly the
exploration, disaster preparations, military application, sea res-
management and service of cloud computing and fog computing
cue, and offshore entertainment. The categories of the existing
for underwater intelligent devices in the sensing layer of the
UIoT applications are shown in Table I. In addition, with the
UIoT, which can reduce the cost of underwater devices while
development of the smart ocean, potential applications of the
meet the requirement of application tasks. The resource of the
UIoT exist that should be seriously considered, such as smart
underwater network and nonunderwater network (communica-
monitoring of fish stocks and feeding environments in the deep
tion layer, networking layer, and fusion layer) of the UIoT are
range (the deep range monitoring network) and a design system
dynamically optimized and scheduled by cloud computing and
for seabed tourist routes (underwater tourism network). UIoT
fog computing in the PSL. The ASL is the management of cloud
applications are utilized daily to make maritime activities safer
computing and fog computing for the ocean application software
and more convenient.
in application layer of the UIoT. It can provide users with high-
quality services and protect users’ privacy. The main functions of
the CSL and FSL in the fusion layer of the UIoT are to support IV. CHALLENGES AND OPEN ISSUES
the management, security, and service of other modules. The It is not an easy task to realize the vision of the UIoT, because
fusion layer service architecture has some similarities with the of the many challenges that must be addressed. In the following
system architecture of the UIoT. The CSL module is connected paragraphs, first the main theories and key methodologies for
with the FSL module, PSL module, and ASL module. The FSL building the future UIoT systems of the smart ocean are demon-
module is directly connected to all other modules. From this strated, as shown in Fig. 3. Then, we provide a brief discussion
architecture, we can see that the response time that CSL module of the main challenges for the underwater network part of the
manage the devices in the ISL module is longer than that of the UIoT. Moreover, we present the open issues of the UIoT based
FSL module. on these challenges.
In the underwater network of the UIoT, because of the band- In Fig. 4, the main theories applied for building the UIoT are
width and transmission speed limitations of the UAC, time- divided into four parts: oceanography, communication theory,
sensitive ocean sensing information cannot be transmitted to the network theory, and the theory of computation. In oceanography,
cloud quickly, even though the UAC is the only relatively mature significant ocean theories are included, which can influence the
underwater communication technology at present. Fog comput- design of the UIoT, such as ocean optics, ocean acoustics, marine
ing is a promising technology that can be utilized by the IoT to electromagnetics, and ocean dynamics. Communication theories
efficiently provide elastic computation, storage, and networking include underwater coding techniques, underwater MAC proto-
services for distributed resource constrained sensors. The ap- cols, and various underwater communication technologies. The
plication and architecture based on fog computing is a research main underwater network theories that should be considered
hotspot [65]–[68]. Hence, fog computing is a more appropriate when building the underwater network part of the UIoT that have
method for some emergency applications of the UIoT, because it been mentioned are underwater topology control, underwater
can respond quickly based on emergency event-aware strategies. routing protocols, underwater network design, and underwater
In addition, cooperative computation using edge servers can ef- network repair. The theory of computation includes algorithm
fectively optimize the structure and performance of underwater design and artificial intelligence that can process underwater
Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.
 4302 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 16, NO. 7, JULY 2020

TABLE I
CATEGORIES OF THE UIOT APPLICATIONS

Fig. 4. Main theories and methodologies of building the UIoT.

data intelligently. These theories can provide design idea and multipath effect, time-varying effect, narrow bandwidth, and
theoretical analysis for the establishment of the underwater part high bit error rate. Acoustic signals are easily affected by various
of the UIoT. According to our analysis of the characteristics of factors, such as transmission power, battery status, multipath
the underwater environment and the existing research studies on effect, ocean ambient noise (turbulence noise, shipping noise,
the underwater network, three major challenges and open issues wind-driven waves noise, and thermal noise), and underwater
that the UIoT currently presents are as follows. obstacles. Thus, an important issue related to acoustic signals is
their irregularity. Network topology and connectivity are directly
or indirectly affected by this signal irregularity. In addition,
A. Underwater Wireless Communication Technology because of the high cost of underwater nodes as compared
1) Vulnerable Underwater Channel Interferes With the Qual- with those of the IoT, the underwater network of the UIoT is
ity of Underwater Wireless Communication: Radio signals are sparser. As a result, the establishment and maintenance of the
absorbed in the underwater environment very quickly and to a communication between different UIoT nodes is more difficult.
serious extent. Underwater optical communication and magnetic The inadequacy of the UAC results in unsatisfactory accuracy,
induction communication cannot effect long-distance commu- reliability, and integrity of data transmission, which seriously
nication. Therefore, the UAC is currently the main underwater restricts the development of the UIoT.
communication method in the large-scale UIoT. However, the 2) Development of High-Speed and Long-Distance Under-
characteristics of the UAC are its slow transmission speed, water Wireless Communication Technology for Smart Ocean:
Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.
QIU et al.: UNDERWATER INTERNET OF THINGS IN SMART OCEAN: SYSTEM ARCHITECTURE AND OPEN ISSUES 4303

Underwater wireless communication technology, such as acous- 2) Exploration of Highly Reliable and Self-Adaptive Network-
tic, wireless optical, and magnetic induction communication, ing Strategy for Large-Scale Perception of the UIoT: In the
has been widely studied by many research institutes worldwide UIoT, a large number of distributed underwater sensing nodes
for many years. However, most emergency ocean applications that form the underwater network part are integrated through
currently rely on submarine cables for data transmission. This self-organizing routing protocols [97] and topology control
is because the long-distance transmission performance of un- strategies. These nodes collaborate on data communication to
derwater wireless communication is poorer than that of radio ensure that data can be sent to the sink node or the control center.
communication. For underwater data transmission of the UIoT Because of the complex underwater environment and inefficient
at present, the underwater multimodal communication based on and unreliable underwater communication technology, the reli-
multiple underwater wireless communication technologies has ability of underwater networks is poor. In particular, when the
become a research hotspot. scale of the underwater network is large, the existing networking
In the future UIoT, large-scale deep-sea monitoring networks strategies are not suitable. Machine learning-based networking
will be established in the oceans. For the development of the and routing strategies for large-scale terrestrial wireless sen-
UIoT to realize the smart ocean, the following three aspects will sor networks in IoT have been extensively studied, because
be the main research interests in underwater communication. of its strong adaptive and computational capabilities. Various
1) The underwater multimodal communication-based intel- underwater network performance of the UIoT is difficult to
ligent communication module needs to be designed for control and optimize by the traditional methods that are used
dynamic switching communication technology according for the IoT because the marine environment is more complex
to transmission tasks. and changeable than the terrestrial environment.
2) The high-speed long-distance underwater wireless com- In the future UIoT, the reliability and security of large-scale
munication technology based on the UMCS is urgently underwater networks should be concerned by researchers for
required in the communication layer. timely and complete transmission of massive perception data.
3) For the underwater network part of the UIoT, the in- The following three aspects should be main open issues for the
telligent medium access control protocol based on the underwater networking strategy.
machine learning is necessary to improve the transmission 1) The prospects of machine learning in terms of the under-
stability in complex underwater environment. In this case, water multimodal network routing design and AUV path
network designers may develop solutions that initially planning in the networking layer of the large-scale UIoT
may not operate as expected. are good.
2) It is essential to study the means of underwater topol-
ogy control, topology reconstruction, and malicious node
detection with self-learning function in the future UIoT.
B. Underwater Networking Strategy 3) An efficient underwater node location strategy is required
1) Node Mobility and Equipment Heterogeneity Impairs the by the underwater topology generation and network de-
Reliability of Underwater Networking: The mobility of the ocean ployment because of the irregular dynamic movement of
current is a unavoidable issue in the networking of underwater ocean currents.
sensors in the UIoT [92]. An interruption of the underwater
network of the UIoT is easily incurred, and therefore, the re-
C. Cooperative Computing Method
liability of the network topology is seriously challenged. In
addition, the accuracy of the underwater localization is also 1) Collection Efficiency of Massive Underwater Data De-
affected mainly by the movement of currents [93]. Although creases the Accuracy of Collaborative Computing: In the large-
there are some ocean current models [94]–[96] that have been scale UIoT, a main challenge is the prompt and effective
applied in UIoT simulations, these models cannot accurately processing of big data, particularly sensing data from different
represent the actual ocean environment, because the ocean cur- underwater networks transmitted by the UAC, by using coop-
rent is affected by many factors, such as wind and temperature. erative computing. Because of the large amount of data and
This greatly increases the difficulty of underwater networking. the narrow bandwidth of the UAC, the data collection intervals
Meanwhile, the UIoT is a complex system that consists of many are frequently too long. Thus, the real-time performance of
heterogeneous network elements. The underwater network part collaborative computing cannot be guaranteed. In addition, the
of the UIoT includes underwater acoustic networks, underwater underwater environment is particularly vulnerable to malicious
wireless optical communication networks, underwater magnetic attacks, because the UIoT sensor nodes are always deployed in
induction networks, and the dynamic wireless network based on inaccessible, unattended, and even hostile environments. This
AUVs. The features of these heterogeneous network elements renders them highly susceptible to various types of damage and
vary widely, and the heterogeneity of devices, technologies, and threats. That is, the accuracy of the collaborative computing is
services presents many serious challenges. The optimization of seriously challenged.
all the network elements of the UIoT can be coordinated for 2) Research on Distributed Smart Cooperative Computing
communication, and the maximization of the efficiency of each Method for Massive Perception Data of the UIoT: As a result of
underwater network is also an important future challenge for the construction of a smart ocean, the applications of the UIoT
researchers. will be more diversified. The interoperability between different

Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.
 4304 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 16, NO. 7, JULY 2020

layers and different underwater intelligent devices is required However, its design idea can be used in the underwater network
for the operation of these UIoT applications. However, heteroge- to speed up the establishment of the UIoT related standards.
neous devices make it difficult to interact with different devices. For example, IEEE 802.15.4 can support three bandwidths and
Therefore, intelligent middleware needs to be developed to assist can provide communication security services. The design of the
different devices to interact. Furthermore, the distributed smart UAC related standard can use it. In addition, it is also required to
cooperative computing technology needs to be designed rea- design intelligent gateways for ensuring normal data transmis-
sonably for different ocean applications according to the impor- sion between different underwater communication technologies.
tance of their data because of the limitations of the underwater Because the future UIoT is a complex system that consists
communication mode as compared with radio communication; of various heterogeneous networks and billions of underwater
for example, the sensing data for underwater disaster warning intelligent objects, it is difficult to guarantee the security and
systems need to be transmitted faster and processed earlier than privacy of users. Any layer of the UIoT receives malicious
other data. By introducing smart perception technology and fog attacks, which may lead to the destruction of the entire network
computing into cooperative computing, it is possible to reduce and the loss of a large number of valuable users’ data. Hence,
the time required to issue an underwater disaster warning, and how to design a perfect data security mechanism to prevent local
thus, reduce people’s response time to these disasters. The edge data theft or tampering for large-scale heterogeneous networks
server can also analyze the network state and underwater envi- in the UIoT should be one of the future research focuses.
ronment to further optimize the underwater network topology
and communication performance. In addition, the utilization of
E. Realization of the Smart Ocean
super nodes for distributed collaborative computing can greatly
reduce the amount of data transmissions, thus improving the The smart ocean is widely applied in ocean activities closely
network energy efficiency in underwater networks. related to humans, such as smart ocean pollution monitor-
To sum up we know, the following four aspects will be main ing, smart deep range monitoring, smart underwater naviga-
concerns for researching on the distributed smart cooperative tion, smart underwater resource exploration, smart underwater
computing method of the UIoT. tourism, smart disaster warning, and smart underwater intrusion
1) Lightweight target recognition algorithms based on ma- detection. It can help people understand the ocean more compre-
chine learning for underwater nodes need to be designed hensively, utilize the ocean’s potentials more fully, and govern
in the sensing layer to improve the effectiveness of per- the ocean more conveniently so that it serves humankind better.
ception data and reduce underwater network transmission The smart ocean requires that the UIoT transmit, calculate,
load. process, and protect various valuable data in the ocean. That
2) It is necessary to study the intelligent middleware in is, the realization of the smart ocean depends on the above five
the communication layer for assisting different hetero- research points of the UIoT: underwater wireless communica-
geneous devices to interact in the large-scale underwater tion, underwater networking, cooperative computing, security,
network with a variety of underwater devices. and standardization. The acceleration of the research of these
3) How to reasonably deploy surface edge servers and un- aspects for the UIoT is of great significance to the development
derwater super nodes in the UIoT and how to reasonably of smart oceans.
distribute load in surface edge servers are two important
issues for the fusion layer to realize high-performance
V. CONCLUSION
distributed cooperative computing.
4) A high level of security, encryption, and authentication The future UIoT will be based on oceanography, underwa-
services should be provided for the data transmission ter sensor technology, underwater multimodal communication
and cooperative computing process because underwater technology, hybrid networks, cooperative computing, fog com-
networks may be exposed to international waters. puting, and cloud computing to construct the smart ocean. In
this article, we examined the existing survey papers on the UIoT
and presented our proposed five-layer UIoT system architecture,
D. Security and Standardization of the UIoT
where the sensing layer focuses on monitoring sensing data from
In addition to the challenges mentioned previously, the UIoT smart sensor devices in the ocean. We discussed the typical
faces the same security and standardization problems as the features of the underwater communication in the communica-
IoT. There is no common global standard to regulate the trans- tion layer. For the networking layer, the routing strategy for
mission and communication between massive heterogeneous underwater part of the UIoT was discussed. In the fusion layer,
underwater devices in the early stage of the construction of the different approaches for decision making based on cloud
the UIoT. Therefore, a common global standard of the UIoT computing and fog computing were examined. Then, we classi-
should be built to simplify transmission and interaction between fied the existing typical applications of the UIoT and proposed
underwater devices. The UIoT standard can learn from relevant some potential applications for the application layer. Finally, we
standards in the IoT to speed up design, such as 6LowPAN [98] highlighted the issues and challenges of the UIoT that must be
and the IEEE 802.15.4 Standard [99]. The main goal of these addressed in the future to build smart oceans, and envisioned the
IoT standards is low power consumption, similar to the UIoT. open issues in the UIoT research. We hope that this article will
They cannot be directly used in underwater networks of the facilitate researchers and developers being able to understand
UIoT because of their different communication technologies. the perspectives and challenges of developing the UIoT.
Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.
QIU et al.: UNDERWATER INTERNET OF THINGS IN SMART OCEAN: SYSTEM ARCHITECTURE AND OPEN ISSUES 4305

REFERENCES [23] L, Liu, N. Zhang, and Y, Liu, “Topology control models and solutions
for signal irregularity in mobile underwater wireless sensor networks,” J.
[1] A. Al-Fuqaha, M. Guizani, M. Mohammadi, M. Aledhari, and M. Ayyash, Netw. Comput. Appl., vol. 51, pp. 68–90, 2015.
“Internet of things: A survey on enabling technologies, protocols, and [24] R. Jurdak, C. Lopes, and P. Baldi, “Battery lifetime estimation and op-
applications,” IEEE Commun. Surveys Tut., vol. 17, no. 4, pp. 2347–2376, timization for underwater sensor networks,” IEEE Sensor Netw. Operat.,
Fourthquarter 2015. vol. 2006, pp. 397–420, Jun. 2004.
[2] T. Qiu, N. Chen, K. Li, M. Atiquzzaman, and W. Zhao, “How can hetero- [25] M. Domingo, “An overview of the internet of underwater things,” J. Netw.
geneous Internet of Things Build our future: A survey,” IEEE Commun. Comput. Appl., vol. 35, no. 6, pp. 1879–1890, 2012.
Surveys Tut., vol. 20, no. 3, pp. 2011–2027, Thirdquarter 2018. [26] M. Berlian et al., “Design and implementation of smart environment
[3] C. Lu, “Context-aware service provisioning via agentized and recon- monitoring and analytics in real-time system framework based on internet
figurable multimodel cooperation for real-life IoT-enabled smart home of underwater things and big data,” in Proc. Int. Electron. Symp., Denpasar,
systems,” IEEE Trans. Syst., Man, and Cybern., Syst., to be published, Indonesia, Sep. 2016, pp. 403–408.
doi: 10.1109/TSMC.2018.2831711. [27] Z. Zhao et al., “EBTCOR: An energy-balanced 3D topology control
[4] T. Qiu, J. Liu, W. Si, and D. O. Wu, “Robustness optimization scheme with algorithm based on optimally rigid graph in UWSNs,” Ad Hoc Sensor
multi-population co-evolution for scale-free wireless sensor networks,” Wireless Netw., vol. 42, no. 3, pp. 295–315, 2018.
IEEE/ACM Trans. Netw., vol. 27, no. 3, pp. 1028–1042, Jun. 2019. [28] C. Kao, Y. Lin, G. Wu, and C. Huang, “A comprehensive study on
[5] Z. Zhou, C. Gao, C. Xu, Y. Zhang, S. Mumtaz, and J. Rodriguez, “Social the internet of underwater things: Applications, challenges, and channel
big-data-based content dissemination in internet of vehicles,” IEEE Trans. models,” Sensors, vol. 17, no. 7, pp. 1–20, 2017.
Ind. Informat., vol. 14, no. 2, pp. 768–777, Feb. 2018. [29] C. Liou, C. Kao, C. Chang, Y. Lin, and C. Huang, “Internet of underwater
[6] T. Qiu, X. Wang, C. Chen, M. Atiquzzaman, and L. Liu, “TMED: A things: Challenges and routing protocols,” in Proc. IEEE Int. Conf. Appl.
spider web-like transmission mechanism for emergency data in vehicular Syst. Invention, Chiba, Japan, Apr. 2018, pp. 1171–1174.
ad hoc networks,” IEEE Trans. Veh. Technol., vol. 67, no. 9, pp. 8682–8694, [30] M. Gopinath et al., “A secure cloud-based solution for real-time moni-
Sep. 2018. toring and management of Internet of underwater things (IOUT),” Neural
[7] D. Saxena and V. Raychoudhury, “Design and verification of an NDN- Comput. Appl., vol. 31, no. 1, pp. 293–308, 2018.
based safety-critical application: A case study with smart healthcare,” [31] K. Urunov, S. Shin, J. Namgung, and S. Park, “High-level architectural
IEEE Trans. Syst., Man, Cybern., Syst., vol. 49, no. 5, pp. 991–1005, design of management system for the Internet of Underwater Things,” in
May 2019. Proc. 10th Int. Conf. Ubiquitous Future Netw., Prague, Czech Republic,
[8] H. Nguyen, F. Mirza, M. Naeem, and M. Nguyen, “A review on IoT Jul. 2018, pp. 326–331.
healthcare monitoring applications and a vision for transforming sensor [32] S. Oteafy and H. Hassanein, “Resilient IoT architectures over dynamic
data into real-time clinical feedback,” in Proc. IEEE 21st Int. Conf. sensor networks with adaptive components,” IEEE Internet Things J.,
Comput. Supported Cooperative Work Des., Wellington, New Zealand, vol. 4, no. 2, pp. 474–483, Apr. 2017.
Apr. 2017, pp. 257–262. [33] I. Yaqoob et al., “Internet of Things architecture: Recent advances, tax-
[9] M. Wollschlaeger, T. Sauter, and J. Jasperneite, “The future of industrial onomy, requirements, and open challenges,” IEEE Wireless Commun.,
communication: Automation networks in the era of the Internet of Things vol. 24, no. 3, pp. 10–16, Jun. 2017.
and Industry 4.0,” IEEE Ind. Electron. Mag., vol. 11, no. 1, pp. 17–27, [34] P. Kumar and U. Gandhi, “A novel three-tier Internet of Things architecture
Mar. 2017. with machine learning algorithm for early detection of heart diseases,”
[10] T. Qiu, H. Wang, K. Li, H. Ning, A. Sangaiah, and B. Chen, “SIGMM: A Comput. Elect. Eng., vol. 65, pp. 222–235, 2018.
novel machine learning algorithm for spammer identification in industrial [35] P. Mell and T. Grance, “The NIST definition of cloud computing,” National
mobile cloud computing,” IEEE Trans. Ind. Informat., vol. 15, no. 4, Institute of Standards and Technology, Gaithersburg, MD, USA, 2011.
pp. 2349–2359, Apr. 2019. [36] F. Bonomi, R. Milito, J. Zhu, and S. Addepalli, “Fog computing and its
[11] L. Yang, S. Yang, and L. Plotnick, “How the internet of things technology role in the internet of things,” in Proc. 1st Edition MCC Workshop Mobile
enhances emergency response operations,” Technological Forecasting Soc. Cloud Comput., Helsinki, Finland, Aug. 2012, pp. 13–16.
Change, vol. 80, no. 9, pp. 1854–1867, 2013. [37] P. Langley and H. A. Simon, “Applications of machine learning and rule
[12] T. Qiu, R. Qiao, and D. Wu, “EABS: An event-aware backpressure induction,” Commun. ACM, vol. 38, no. 11, pp. 54–64, 1995.
scheduling scheme for emergency Internet of Things,” IEEE Trans. Mobile [38] S. Kamal, A. Mujeeb, and M. H. Supriya, “Novel class detection of
Comput., vol. 17, no. 1, pp. 72–84, Jan. 2018. underwater targets using self-organizing neural networks,” in Proc. IEEE
[13] R. Naylor and M. Burke, “Aquaculture and ocean resources: Raising tigers Underwater Technol., Chennai, India, Feb. 2015, pp. 1–5.
of the sea,” Annu. Rev. Environ. Resour., vol. 30, pp. 185–218, 2005. [39] H. Yang, A. Gan, H. Chen, Y. Pan, J. Tang, and J. Li, “Underwater acoustic
[14] G. Besio, L. Mentaschi, and A. Mazzino, “Wave energy resource assess- target recognition using SVM ensemble via weighted sample and feature
ment in the Mediterranean Sea on the basis of a 35-year hindcast,” Energy, selection,” in Proc. 13th Int. Bhurban Conf. Appl. Sci. Technol., Melbourne,
vol. 94, pp. 50–63, 2016. Australia, Dec. 2017, pp. 522–527.
[15] R. Roslynna and D. Brunei, “A review of tidal current energy resource [40] G. Hu, K. Wang, Y. Peng, M. Qiu, J. Shi, and L. Liu, “Deep learning
assessment: Current status and trend,” in Proc. 5th Int. Conf. Renewable methods for underwater target feature extraction and recognition,” Com-
Energy, Gener. Appl., Al Ain, United Arab Emirates, Feb. 2018, pp. 34–40. put. Intell. Neurosci., vol. 2018, pp. 1–10, 2018.
[16] M. Xu and L. Liu, “Sender-receiver role-based energy-aware scheduling [41] I. Akyildiz, D. Pompili, and T. Melodia, “Challenges for efficient com-
for Internet of Underwater Things,” IEEE Trans. Emerg. Topics Comput., munication in underwater acoustic sensor networks,” ACM SIGBED Rev.,
vol. 7, no. 2, pp. 324–336, Apr./Jun. 2019. vol. 1, no. 2, pp. 3–8, 2004.
[17] M. Heesemann, T. Insua, M. Scherwath, K. Juniper, and K. Moran, “Ocean [42] S. Sendra, J. Lloret, J. Jimenez, and L. Parra, “Underwater acoustic
networks Canada: From geohazards research laboratories to smart ocean modems,” IEEE Sensors J., vol. 16, no. 11, pp. 4063–4071, Jun. 2016.
systems,” Oceanography, vol. 27, no. 2, pp. 151–153, 2014. [43] M. Jamali, J. Salehi, and F. Akhoundi, “Performance studies of underwater
[18] L. Liu, R. Wang, and F Xiao, “Topology control algorithm for underwater wireless optical communication systems with spatial diversity: MIMO
wireless sensor networks using GPS-free mobile sensor nodes,” J. Netw. scheme,” IEEE Trans. Commun., vol. 65, no. 3, pp. 1176–1192, Mar.
Comput. Appl., vol. 35, no. 6, pp. 1953–1963, 2012. 2017.
[19] W. Kim, H. Moon, Y. Yoon, “Adaptive triangular deployment of un- [44] Z. Zeng, S. Fu, H. Zhang, Y. Dong, and J. Cheng, “A survey of underwater
derwater wireless acoustic sensor network considering the underwater optical wireless communications,” IEEE Commun. Surveys Tut., vol. 19,
environment,” J. Sensors, vol. 2019, pp. 1–11, 2019. no. 1, pp. 204–238, Firstquarter 2017.
[20] F. Senel, K. Akkaya, M. Erol-Kantarci, and T. Yilmaz, “Self-deployment [45] M. Domingo, “Magnetic induction for underwater wireless communica-
of mobile underwater acoustic sensor networks for maximized coverage tion networks,” IEEE Trans. Antennas Propag., vol. 60, no. 6, pp. 2929–
and guaranteed connectivity,” Ad Hoc Netw., vol. 34, pp. 170–183, 2015. 2939, Jun. 2012.
[21] Z. Zhou, B. Yao, R. Xing, L. Shu, and S. Bu, “E-CARP: An energy efficient [46] I. Akyildiz, p. Wang, and Z. Sun, “Realizing underwater communication
routing protocol for UWSNs in the internet of underwater things,” IEEE through magnetic induction,” IEEE Commun. Mag., vol. 53, no. 11, pp. 42–
Sensors J., vol. 16, no. 11, pp. 4072–4082, Jun. 2016. 48, Nov. 2015.
[22] A. Amar, A. Gilad, and S. Milica, “Low complexity residual doppler shift [47] B. Gulbahar and O. Akan, “A communication theoretical modeling and
estimation for underwater acoustic multicarrier communication,” IEEE analysis of underwater magneto-inductive wireless channels,” IEEE Trans.
Trans. Signal Process., vol. 65, no. 8, pp. 2063–2076, Apr. 2017. Wireless Commun., vol. 11, no. 9, pp. 3326–3334, Sep. 2012.

Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.
 4306 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 16, NO. 7, JULY 2020

[48] M. Stojanovic, “On the relationship between capacity and distance in an [70] P. Pandey, D. Pompili, and J. Yi, “Dynamic collaboration between net-
underwater acoustic communication channel,” ACM SIGMOBILE Mobile worked robots and clouds in resource-constrained environments,” IEEE
Comput. Commun. Rev., vol. 11, no. 4, pp. 34–43, 2007. Trans. Autom. Sci. Eng., vol. 12, no. 2, pp. 471–480, Apr. 2015.
[49] P. Gjanci, C. Petrioli, S. Basagni S, and C. Phillips, “Path finding for [71] L. Vieira, M. Vieira, J. Nacif, and A. Vieira, “Autonomous wire-
maximum value of information in multi-modal underwater wireless sensor less lake monitoring,” Comput. Sci. Eng., vol. 20, no. 1, pp. 66–75,
networks,” IEEE Trans. Mobile Comput., vol. 17, no. 2, pp. 404–418, 2018.
Feb. 2018. [72] C. Trasvina-Moreno, R. Blasco, A. Marco, R. Casas, and A. Trasvina-
[50] R. Diamant, P. Casari, F. Campagnaro, O. Kebkal, V. Kebkal, and M. Zorzi, Castro, “Unmanned aerial vehicle based wireless sensor network for
“Fair and throughput-optimal routing in multimodal underwater net- marine-coastal environment monitoring,” Sensors, vol. 17, no. 3, pp. 1–22,
works,” IEEE Trans. Wireless Commun., vol. 17, no. 3, pp. 1738–1754, 2017.
Mar. 2018. [73] G. Cario, A. Casavola, P. Gjanci, M. Lupia, C. Petrioli, and D. Spaccini,
[51] S. Jiang, “On reliable data transfer in underwater acoustic networks: A “Long lasting underwater wireless sensors network for water quality
survey from networking perspective,” IEEE Commun. Surveys Tut., vol. 20, monitoring in fish farms,” in Proc. Oceans, Aberdeen, U.K., Jun. 2017,
no. 2, pp. 1036–1055, 2Q 2018. pp. 1–6.
[52] S. Jiang, “On securing underwater acoustic networks: A survey,” [74] A. Abdillah et al., “Design and development of low cost coral monitor-
IEEE Commun. Surveys Tut., vol. 21, no. 1, pp. 729–752, 1Q 2019, ing system for shallow water based on internet of underwater things,”
doi: 10.1109/COMST.2018.2864127. J. Telecommun., Electron. Comput. Eng., vol. 9, no. 2–5, pp. 97–101,
[53] Z. Gong, C. Li, and F. Jiang, “AUV-aided joint localization and time 2017.
synchronization for underwater acoustic sensor networks,” IEEE Signal [75] J. Lloret, M. Garcia, S. Sendra, and G. Lloret, “An underwater wireless
Process. Lett., vol. 25, no. 4, pp. 477–481, Apr. 2018. group-based sensor network for marine fish farms sustainability monitor-
[54] J. Yan, X. Yang, X. Luo, and C. Chen, “Energy-efficient data collection ing,” Telecommun. Syst., vol. 60, no. 1, pp. 67–84, 2015.
over AUV-assisted underwater acoustic sensor network,” IEEE Syst. J., [76] G. Acar and A. Adams, “ACMENet: An underwater acoustic sensor
vol. 12, no. 4, pp. 3519–3530, Dec. 2018. network protocol for real-time environmental monitoring in coastal areas,”
[55] T. Zhang, C. L. P. Chen, L. Chen, X. Xu, and B. Hu, “Design of highly Proc. Inst. Elect. Eng.—Radar, Sonar Navigat., vol. 153, no. 4, pp. 365–
nonlinear substitution boxes based on I-Ching operators,” IEEE Trans. 380, Aug. 2006.
Cybern., vol. 48, no. 12, pp. 3349–3358, Dec. 2018. [77] S. Srinivas, P. Ranjitha, R. Ramya, and G. Narendra, “Investigation of
[56] C. L. P. Chen, T. Zhang, L. Chen, and S. C. Tam, “I-Ching divination evo- oceanic environment using large-scale UWSN and UANETs,” in Proc.
lutionary algorithm and its convergence analysis,” IEEE Trans. Cybern., 8th Int. Conf. Wireless Commun., Netw. Mobile Comput., Shanghai, China,
vol. 47, no. 1, pp. 2–13, Jan. 2017. Sep. 2012, pp. 1–5.
[57] S. Basagni, V. Di Valerio, P. Gjanci, and C. Petrioli, “Finding MARLIN: [78] N. Mohamed, I. Jawhar, J. Al-Jaroodi, and L. Zhang, “Sensor network ar-
Exploiting multi-modal communications for reliable and low-latency un- chitectures for monitoring underwater pipelines,” Sensors, vol. 11, no. 11,
derwater networking,” in Proc. IEEE Conf. Comput. Commun., Atlanta, pp. 10738–10764, 2011.
GA, USA, May 2017, pp. 1–9. [79] H. Saeed, S. Ali, S. Rashid, S. Qaisar, and E. Felemban, “Reliable mon-
[58] C. L. P. Chen and Z. Liu, “Broad learning system: An effective and efficient itoring of oil and gas pipelines using wireless sensor network (WSN):
incremental learning system without the need for deep architecture,” REMONG,” in Proc. 9th Int. Conf. System Syst. Eng., Adelade, Australia,
IEEE Trans. Neural Netw. Learn. Syst., vol. 29, no. 1, pp. 10–24, Jan. Jun. 2014, pp. 230–235.
2018. [80] R. Marin-Perez, J. Garcia-Pintado, and A. Gomez, “A realtime measure-
[59] Y. Su, R. Fan, X. Fu, and Z. Jin, “DQELR: An adaptive deep Q-network- ment system for long-life flood monitoring and warning applications,”
based energy-and latency-aware routing protocol design for underwa- Sensors, vol. 12, no. 4, pp. 4213–4236, 2012.
ter acoustic sensor networks,” IEEE Access, vol. 7, pp. 9091–9104, [81] P. Kumar, P. Kumar, P. Priyadarshini, and Srija, “Underwater acoustic
2019. sensor network for early warning generation,” in Proc. Oceans, Hampton
[60] G. Ahmed, N. M. Khan, Z. Khalid, and R. Ramer, “Cluster head selection Roads, VA, USA, Oct. 2012, pp. 1–6.
using decision trees for wireless sensor networks,” in Proc. Int. Conf. Intell. [82] O. Nweke and F. Henry, “Wireless sensor networks based pipeline vandal-
Sensors, Sensor Netw. Inform. Process., Sydney, Australia, Dec. 2008, isation and oil spillage monitoring and detection: Main benefits for Nigeria
pp. 173–178. oil and gas sectors,” SIJ Trans. Comput. Sci. Eng. Its Appl., vol. 3, no. 1,
[61] J. Barbancho, C. Leon, F. Molina, and A. Barbancho, “A new QoS routing pp. 1–6, 2015.
algorithm based on self-organizing maps for wireless sensor networks,” [83] K. Casey, A. Lim, and G. Dozier, “A sensor network architecture for
Telecommun. Syst., vol. 36, no. 1–3, pp. 73–3, 2007. tsunami detection and response,” Int. J. Distrib. Sensor Netw., vol. 4, no. 1,
[62] H. Luo, K. Wu, R. Ruby, Y. Liang, Z. Guo, and L. Ni, “Software-defined pp. 28–43, 2008.
architectures and technologies for underwater wireless sensor networks: [84] K. Denos, M. Ravaut, A. Fagette, and H. Lim, “Deep learning applied to
A survey,” IEEE Commun. Surveys Tut., vol. 20, no. 4, pp. 2855–2888, underwater mine warfare,” in Proc. Oceans, Aberdeen, U.K., Jun. 2017,
Fourthquarter 2018. pp. 1–7.
[63] F. Akyildiz, P. Wang, and C. Lin, “SoftWater: Software-defined networking [85] C. Barngrover, R. Kastner, and S. Belongie, “Semisynthetic versus real-
for next-generation underwater communication systems,” Ad Hoc Netw., world sonar training data for the classification of mine-like objects,” IEEE
vol. 46, pp. 1–11, 2016. J. Ocean. Eng., vol. 40, no. 1, pp. 48–56, Jan. 2015.
[64] A. Botta, W. De Donato, V. Persico, and A. Pescapé, “Integration of cloud [86] R. Manjula and S. Manvi, “Coverage optimization based sensor de-
computing and internet of things: A survey,” Future Gener. Comput. Syst., ployment by using PSO for anti-submarine detection in UWASNs,”
vol. 56, pp. 684–700, 2016. in Proc. 12th Symp. Ocean Electron., Kochi, India, Oct. 2013,
[65] K. Suto, H. Nishiyama, N. Kato, and C. Huang, “An energy-efficient pp. 15–22.
and delay-aware wireless computing system for industrial wireless sensor [87] J. Yan, X. Li, X. Luo, and X. Guan, “Virtual-lattice based intrusion
networks,” IEEE Access, vol. 3, pp. 1026–1035, 2015. detection algorithm over actuator-assisted underwater wireless sensor
[66] R. Yang, Q. Xu, M. Au, Z. Yu, H. Wang, and L. Zhou, “Position based networks,” Sensors, vol. 17, no. 5, pp. 1–17, 2017.
cryptography with location privacy: A step for fog computing,” Future [88] M. Bayat, N. Crasta, A. Aguiar, and A. Pascoal, “Range-based underwater
Gener. Comput. Syst., vol. 78, pp. 799–806, 2018. vehicle localization in the presence of unknown ocean currents: Theory and
[67] T. Lu, S. Chang, and W. Li, “Fog computing enabling geographic routing experiments,” IEEE Trans. Control Syst. Technol., vol. 24, no. 1, pp. 122–
for urban area vehicular network,” Peer-to-Peer Netw. Appl., vol. 11, no. 4, 139, Jan. 2016.
pp. 749–755, 2018. [89] J. Yan, X. Zhang, X. Luo, Y. Wang, C. Chen, and X. Guan, “Asyn-
[68] J. Lin, W. Yu, N. Zhang, X. Yang, H. Zhang, and W. Zhao, “A survey chronous localization with mobility prediction for underwater acoustic
on internet of things: Architecture, enabling technologies, security and sensor networks,” IEEE Trans. Veh. Technol., vol. 67, no. 3, pp. 2543–2556,
privacy, and applications,” IEEE Internet Things J., vol. 4, no. 5, pp. 1125– Mar. 2018.
1142, Oct. 2017. [90] J. Marshall, “Magnetic field swimmer positioning,” IEEE Sensors J.,
[69] M. Yuriyama and K. Takayuki Kushida, “Sensor-cloud infrastructure- vol. 15, no. 1, pp. 172–179, Jan. 2015.
physical sensor management with virtualized sensors on cloud comput- [91] F. Magalhaes, G. Vannozzi, G. Gatta, and S. Fantozzi, “Wearable inertial
ing,” in Proc. 13th Int. Conf. Netw.-Based Inf. Syst., Takayama, Japan, sensors in swimming motion analysis: A systematic review,” J. Sports Sci.,
Sep. 2010, pp. 1–8. vol. 33, no. 7, pp. 732–745, 2015.

Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.
QIU et al.: UNDERWATER INTERNET OF THINGS IN SMART OCEAN: SYSTEM ARCHITECTURE AND OPEN ISSUES 4307

[92] I. Akyildiz, D. Pompili, and T. Melodia, “Underwater acoustic sensor Tong Zhang (S’12–M’16) received the B.S. de-
networks: Research challenges,” Ad Hoc Netw., vol. 3, no. 3, pp. 257–279, gree in software engineering from Sun Yat-sen
2005. University, Guangzhou, China, in 2009, and the
[93] C. Liu, Z. Zhao, W. Qu, T. Qiu, and A. Sangaiah, “A distributed node M.S. degree in applied mathematics and Ph.D.
deployment algorithm for underwater wireless sensor networks based on degree in software engineering from the Univer-
virtual forces,” J. Syst. Archit., vol. 97, pp. 9–19, 2019. sity of Macau, Macau, China, in 2011 and 2016,
[94] D. Pompili, T. Melodia, and I. Akyildiz, “Three-dimensional and two- respectively.
dimensional deployment analysis for underwater acoustic sensor net- He is currently an Assistant Professor with
works,” Ad Hoc Netw., vol. 7, no. 4, pp. 778–790, 2009. the School of Electronics and Information, South
[95] Y. Ren, W. Seah, and P. Teal, “Performance of pressure routing in drifting China University of Technology, Guangzhou. His
3 D underwater sensor networks for deep water monitoring,” in Proc. 7th research interests include affective computing,
ACM Int. Conf. Underwater Netw. Syst., Los Angeles, CA, USA, Nov. evolutionary computation, neural network, and other machine learning
2012, pp. 1–8. techniques and their applications.
[96] A. Caruso, F. Paparella, L. Vieira, M. Erol, and M. Gerla, “The meander- Dr. Zhang has been working in publication matters for many IEEE
ing current mobility model and its impact on underwater mobile sensor conferences.
networks,” in Proc. IEEE 27th Conf. Comput. Commun., Phoenix, AZ, Chen Chen (M’09–SM’18) received the B.Eng.,
USA, Apr. 2008, pp. 221–225. M.Sc., and Ph.D. degrees in electrical engineer-
[97] Z. Zhao, W. Qu, C. Liu, T. Qiu, and X. Guang, “A novel self-organizing ing and computer science (EECS) from Xid-
routing algorithm for underwater internet of things,” in Proc. IEEE 23nd ian University, Xi’an, China, in 2000, 2006, and
Int. Conf. Comput. Supported Cooperative Work Des., Porto, Portugal, 2008, respectively.
May 2019, pp. 470–475. He is currently an Associate Professor with
[98] Z. Shelby and C. Bormann, 6LoWPAN: The Wireless Embedded Internet, the Department of EECS, Xidian University. He
vol. 43. New York, NY, USA: Wiley, 2011. is also the Director with the Xi’an Key Labora-
[99] IEEE Standard for Local and Metropolitan Area Networks–part 15.4: tory of Mobile Edge Computing and Security,
Low-Rate Wireless Personal Area Networks (LR-WPANs), IEEE Std. 802. the Director with the Intelligent Transportation
15. 4-2011, 2011. Research Laboratory, Xidian University, and the
Visiting Professor with the Key Laboratory of Embedded System and
Service Computing, Tongji University, Shanghai, China. He was a Visit-
ing Professor with the EECS Department, University of Tennessee and
with the Computer Science Department, University of California. He has
Tie Qiu (M’12–SM’16) received the Ph.D. de- authored/coauthored two books and more than 100 scientific papers in
gree in computer science from the Dalian Uni- international journals and conference proceedings. He has contributed
versity of Technology, Dalian, China, in 2012. to the development of ten copyrighted software systems and invented
He is currently a Full Professor with the 80 patents.
School of Computer Science and Technology, Dr. Chen serves as a General Chair, a Program Chair (PC) Chair,
Tianjin University, Tianjin, China. Prior to this a Workshop Chair, or a Technical Program Committee Member of a
position, he was an Assistant Professor in 2008 number of conferences. He is a Senior Member of the China Computer
and an Associate Professor in 2013 with the Federation and a Member of the ACM and the Chinese Institute of
School of Software, Dalian University of Tech- Electronics.
nology. He was a Visiting Professor with Elec-
C. L. Philip Chen (S’88–M’88–SM’94–F’07) re-
trical and Computer Engineering, Iowa State
ceived the M.S. degree from the University of
University, Ames, IA, USA (2014–2015). He has authored/coauthored
Michigan, Ann Arbor, Michigan, U.S.A., in 1985,
nine books, and more than 100 scientific papers in international journals
and Ph.D. degree from Purdue University, West
and conference proceedings, such as the IEEE/ACM TRANSACTIONS ON
Lafayette, Indiana, U.S.A., in 1988, both in elec-
NETWORKING, IEEE TRANSACTIONS ON MOBILE COMPUTING, IEEE TRANS-
trical engineering.
ACTIONS ON KNOWLEDGE AND DATA ENGINEERING, IEEE TRANSACTIONS ON
He is a Chair Professor with the Department
VEHICULAR TECHNOLOGY, IEEE COMMUNICATIONS SURVEYS AND TUTORI-
of Computer and Information Science, Faculty of
ALS, IEEE COMMUNICATIONS MAGAZINE, etc. There are 11 papers listed
Science and Technology, University of Macau,
as ESI highly cited papers. He has contributed to the development of
Macau, China. Being a Program Evaluator of the
three copyrighted software systems and invented 14 patents.
Accreditation Board of Engineering and Tech-
Dr. Qiu serves as an Associate Editor for the IEEE TRANSACTIONS ON
nology Education, Baltimore, MD, USA, for computer engineering, elec-
SYSTEMS, MAN, AND CYBERNETICS: SYSTEMS, an Area Editor of Ad Hoc
trical engineering, and software engineering programs, he successfully
Networks (Elsevier), an Associate Editor for the IEEE ACCESS JOURNAL
architects the University of Macau’s Engineering and Computer Sci-
and the Computers and Electrical Engineering (Elsevier), and a Guest
ence programs receiving accreditations from Washington/Seoul Accord
Editor for the Future Generation Computer Systems. He serves as a
through the Hong Kong Institute of Engineers (HKIE), of which is consid-
General Chair, a Program Chair, a Workshop Chair, a Publicity Chair,
ered as his utmost contribution in engineering/computer science educa-
a Publication Chair, or a TPC Member of a number of international
tion for Macau as the former Dean of the Faculty. His current research
conferences. He is a Senior Member of the China Computer Federation
interests include systems, cybernetics, and computational intelligence.
and a Senior Member of the ACM.
Dr. Chen is a Fellow of the American Association for the Advancement
of Science (AAAS), the International Association for Pattern Recognition
(IAPR), the Chinese Association of Automation (CAA), and the HKIE,
and a Member of the Academia Europaea, the European Academy
Zhao Zhao received the bachelor’s degree in of Sciences and Arts, and the International Academy of Systems and
electrical engineering and automation from the Cybernetics Science. He was the recipient of the IEEE Norbert Wiener
Hebei University of Technology City College, Award in 2018 for his contribution in systems and cybernetics, and
Tianjin, China, in 2015 and the master’s de- machine learnings. He was also the recipient of the 2016 Outstanding
gree in computer science and technology from Electrical and Computer Engineers Award from his Purdue University.
the Hebei University of Engineering, Handan, He was the IEEE Systems, Man, and Cybernetics Society President
China, in 2018. He is currently working to- from 2012 to 2013, and currently, he is the Editor-in-Chief for the the
ward the Ph.D. degree in software engineering IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS: SYSTEMS,
with the School of Software, College of Intelli- and an Associate Editor for the IEEE TRANSACTIONS ON FUZZY SYSTEMS
gence and Computing, Tianjin University, Tian- and the IEEE TRANSACTIONS ON CYBERNETICS. He was the Chair of TC
jin. 9.1 Economic and Business Systems of the International Federation of
His research interests include Internet of Things and the topology Automatic Control from 2015 to 2017 and is currently a Vice President
control of underwater wireless sensor networks. of the CAA.

Authorized licensed use limited to: Thapar Institute of Engineering & Technology. Downloaded on March 31,2024 at 09:38:01 UTC from IEEE Xplore. Restrictions apply.