A uw-Cellular Network: Design, Implementation and Experiments

Abstract

The most significant increase of current task is in the desire for operational flexibility and agility in large-scale underwater network application scenarios in recent years. In order to address the challenging problems in Underwater Wireless Sensor Networks (UWSNs), we propose a large-scale UWSN based on the cellular network architecture called Underwater Cellular (uw-Cellular) network. It is designed especially for application scenarios where a large number of both fixed and mobile network nodes exist in a wide area to monitor the underwater environment. We also design protocols in each network layer in order to ensure reasonable channel sharing, data forwarding path selection and data reliability. The purpose of the simulation study we implement is to evaluate the performance of the CLA routing strategy compared to the VBF and the MFLOOD protocols in the uw-Cellular network. We also deploy a twenty-node uw-Cellular network in the real-world environment as the field case study. The experimental results showed that the Data Rate between any nodes could reach above 500 bps, and the network Average Throughput was no less than 550 bps under various load situations.

1. Introduction

With the increasing requirements in a broad range of applications such as underwater environment monitoring and protection, disaster early warning, homeland defense, underwater vehicle navigation and underwater entertainment activities [1,2,3,4], the concept of the Internet of Underwater Things (IoUT) has continued to gain in popularity since the 2010s [5]. It is a global network of smart, interconnected underwater devices that enable unprecedented monitoring and surveillance of vast areas of our oceans and seas.

There are numerous means to enable intercommunication between underwater devices. Submarine fiber optic cables provide stable high-speed communication capabilities, but come at a huge cost. Electromagnetic (EM) waves can also be used for underwater wireless communication. However, on one hand, due to the severe attenuation of EM waves in seawater, the communication distance of high frequency EM waves is extremely limited. On the other hand, Very Low Frequency (VLF) and Super Low Frequency (VLF) communications have a significant disadvantage in terms of communication data volume. Underwater Optical and Magnetic Induction (MI) communication also has a wide range of applications in some specific scenarios. But they can only provide high communication rates at extremely limited distances. Many efforts have proven that underwater acoustic communication is the most suitable method for medium or long-range transmission in the marine environment [6,7]. This is beneficial for the deployment and utilization of large-scale underwater wireless networks. The Underwater Wireless Sensor Networks (UWSNs), defined as networks composed of multiple types of underwater network nodes that communicate with each other using acoustic signals for some targeted tasks, are the key enabling technology of large-scale IoUT.

However, the terrestrial networking protocols would not work well in the underwater environment. And the limitations of underwater acoustic communication, such as long propagation delays, limited bandwidth, signal fading, asymmetric links, high error probability and extended multipath propagation [8,9], are always the key factors affecting the performance and application development of UWSNs. In addition, compared to the past, the most significant increase of current task is in the desire for operational flexibility and agility in large-scale underwater network application scenarios. This means that more network nodes, both fixed and mobile, are needed to cover a wide area. Remote Operated Vehicles (ROVs), Autonomous Underwater Vehicles (AUVs), and other smart unmanned underwater mobile devices meet this urgent need and have been adapted to a wide range of application scenarios. However, the movement of network nodes makes it hard to maintain a reliable communication link from the source node to the destination node. The frequent changes of the network topology and the Doppler effect caused by the movement of nodes make communication and networking more difficult. Most of the research questions, however, have focused on systems with fixed sensors and devices which are deployed in a particular location without any form of mobility. Due to these issues, how to spontaneously establish a network and quickly find a relatively reliable path to forward information is one of the key problems of the large-scale UWSNs including both static and dynamic network nodes.

In this paper, we propose a large-scale UWSN based on the cellular network architecture called uw-Cellular network. It is designed specifically for application scenarios where a large number of both fixed and mobile network nodes exist in a wide area to monitor the underwater environment. In this network, several fixed network nodes act as the Gateway node and the Base Station (BS) nodes, which together form the cellular network. And other nodes, as terminal nodes of the network, are distributed within the coverage area of the BS nodes. In particular, it is difficult to establish or maintain a stable and reliable communication link between each mobile terminal nodes and the BS nodes. Meanwhile, the uw-Cellular network faces both the problems of communication channel conflict avoidance and reliable data transmission. In order to solve these problems, we proposed a MAC protocol for the hierarchy network called SDDA for effective and efficient channel resource allocation. In addition, we also proposed a method called Communication Link Awareness (CLA) to evaluate the quality of the communication link status between two nodes. The main idea of this method is to transmitting information successfully with better efficiency in terms of delay based on the information available to each node. Strategies to guarantee data transmission reliability are also applied to uw-Cellular networks. Moreover, achieve the target functions, we also developed several application protocols including processing sensor data and remote commands.

In order to verify the performance of the uw-Cellular network with the introduction of the CLA method, we implement the CLA method at the routing layer. We compare it with the VBF and MFLOOD protocols on some network metrics in the simulation environments. And the simulation test results prove the CLA method can achieve better performance in a network with frequent node movement. As a field case study, we deployed a twenty-node uw-Cellular network in the real underwater environment. The network has a single fixed Gateway node in order to connect terrestrial and underwater networks. And there are three fixed nodes act as the BS nodes. Ten fixed and six mobile terminal network nodes are deployed within the access point coverage area. During the sea trial, this large-scale network accumulated more than 240 h of trouble-free operation. The experimental results show that the uw-Cellular network with both fixed and mobile network nodes can be used in the practical application scenario with good robustness and performance.

The main contributions of this paper are summarized as follows.

• We design the uw-Cellular network, with the goal of realizing a large-scale monitoring hydroacoustic network, which can be long-term deployed in the real underwater environments. The network is composed of three parts, the Gateway node, the BS nodes and the terminal nodes. The Gateway node has the responsibility to connect the networks above and below the surface. The BS nodes provide access and forwarding services for the terminal nodes. According to different practical applications, various types of terminal nodes can be added to the network for a variety of different tasks.

• For a hydroacoustic network that requires long-term deployment, the communication link is highly time-varying because of the changes in the underwater sound field and network topology. We design protocols in each network layer in order to ensure reasonable channel sharing, data forwarding path selection and data reliability. The protocols used to process sensor data at the application layer are also implemented in order to face the need of practical scenarios.

• We implement and deploy a twenty-node uw-Cellular network in the real underwater environment. The results of experiments in real environment show that the large-scale monitoring network can be used in the practical application scenario with good robustness and performance.

The rest of this paper is organized as follow. The related previous works is presented in Section 2. The design of uw-Cellular network is presented in Section 3. Section 4 introduces the experiments in simulation environment and verifies the performance of the protocol. At the same time, we give a detailed description about the filed environment in Section 5. Finally, we summarize the contributions of our work and discuss the directions of future work in Section 6.

2. Related Work

Research on hydroacoustic networks started in the 1990s. The concept of Autonomous Ocean Sampling Network (AOSN) [10] was proposed in 1993. AOSN is an approach toward four-dimensional ocean sampling in the coastal frontal zone. It is done with several AUVs as well as with distributed acoustic and point sensors. And the AOSN-II experiment was conducted in 2003 [11]. During this a month-long experiment, up to 10 or more AUVs were deployed simultaneously to complete the depth, temperature, salinity, particulates, chlorophyll and light intensity data collection. The Seaweb program was proposed and developed in 2000s [12,13,14,15]. In the trials conducted in 2001, a submarine maneuvered around a 14-node Seaweb network accompanied by two RACOM buoys. And a chain-like network experiment conducted in 2004 with 40 fixed nodes. Persistent Littoral Undersea Surveillance Network (PLUSNet) [16] has been established by the US navy since 2005. It is a semi-autonomous controlled network of fixed bottom and mobile sensors for keeping a constant eye on littoral zones. Its prototype system was delivered in 2015 and includes five underwater gliders and six REMUS AUVs. The Acoustic Communication network for Monitoring of Environment in coastal areas (ACMEnet) [17] project, supported by the European Union, is a long-term hydroacoustic communication network for real-time observation of the coastal environment. ACMEnet uses the MFSK/TDMA master-slave network protocol. Three fixed nodes form the minimum ACMENet system. Smart and Networking Underwater Robots in Cooperation Meshes (SWARMs) [18] project is sponsored by Horizon 2020. The main goal of this project is to reduce the operational cost and increase the safety of tasks assigned to unmanned underwater vehicles (AUVs/ROVs) in the operations.The project validated networked communication between the land-based control center and multiple underwater robots in multiple trials. All the above mentioned works have made significant progress in hydroacoustic networks, but they still need further research and validation in terms of network scale, heterogeneous node networking, and networking flexibility.

In recent years, significant advances have been made in the development of UANs. As the demands increasing in communication between various remote instruments within a network environment, a lot of efforts have been made to establish large-scale UWSNs for different scenarios. A full duplex UAN [19] was proposed in 2022. In this study, they proposed a spatial reuse scheduling algorithm that exploits full duplex communication. In [20], the AUVs’ synchronization is used to achieve multi-hop communication for delivering data to a sink over large-scale coverage areas. However, the synchronization mechanism is a greater challenge for the underwater networks which is energy-constrained. Ref. [21] also introduce a data collection method especially for the AUV aided UWSNs. It helps to solve two critical problems, the energy constraint of underwater devices and large demand for data collection. In [22], a probabilistic Space Division Multiple Access (SDMA) method for short/medium distances (less than 2 km) is presented. This method is useful for achieving a reliable coordination among the AUVs in the acoustic communication channel. DARP [23] is a depth adaptive routing protocol which considering the speed of acoustic signal varies with water depth. But its application scenario is more oriented towards 3D underwater networks consisting of fixed nodes. In [24], two approaches for scheduling large-scale underwater networks are proposed. The centralized scheduling approach yields the most efficient schedules and the clustered-based approach provides significant benefits in terms of communication and computation. In addition, a lot of efforts have been made in the field of localization schemes for large-scale UWSNs [25,26].

There is also a lot of work focused on the testbeds of the UANs in a real environment. The contributions of  [27] include the initial steps towards an indoor testbed, an outdoor testbed, and an open-access software suite. Ref. [28] proposes a new low-cost distributed networked localization and time synchronization framework for underwater acoustic sensor networks. This scheme is implemented and tested in a testbed based on Teledyne Benthos Telesonar SM-975 underwater modems. Ref. [29] is UAN testbed which can provide the reference to ocean going autonomous vehicles, static sensor and communication nodes, and underlying software tool chain. Ref. [30] aims to provide a community testbed for underwater wireless sensor networks.

Comparing to the terrestrial wireless networks, there are more difficulties in selecting and managing the optimal path for transmitting data from the source node and to the destination node in the large-scale UWSN application scenarios. The issues routing protocol design should address includes the movement of nodes, high and inefficient propagation delay, energy constraints, location estimation and 3D deployment [31,32,33,34]. Depending on the application scenarios, the underwater routing protocol may have different core issues to address and can be divided into three categories: energy-based protocols, data-based protocols, and geographic information-based protocols [35].

Other than that, combining the above information to make routing decisions is another type of routing protocol. Our approach in this article falls into this category. In [36], an asymmetric link-based reverse routing (AREP) is proposed. In AREP, each node maintains a neighbor table in which items are used to analyze the link state. This method determines whether there is a symmetrical link or an asymmetrical link between two nodes that are neighbors of each other by determining whether they can communicate with each other successfully. LAFR [37] also proposed a link detection mechanism. In order to get the link state information (symmetrical link or asymmetric link), nodes need to send Link Detection messages periodically. However, due to the nature of the hydroacoustic channel, it is not enough to estimate the communication link state by whether the communication was successful or not. In addition, link detection packets usually contain no communication information, which not only wastes energy but also occupies the channel, potentially preventing successful transmission of the data packet.

Compared with other works, the main contribution of this paper focuses on the realization of a large-scale UAN system that can operate stably in a real environment by solving the data forwarding path planning problem and the large number of node channels sharing problem in a mobile topology. The design methodology and experimental results of this paper will be described in detail as follow.

3. Design of uw-Cellular Network

3.1. Application Scenarios

The growing interest in a large number of applications, such as offshore oil and natural gas pipelines security, coastal surveillance and port security, has created the urgent need for underwater wireless monitoring network. As shown in Figure 1, the underwater wireless monitoring network is kind of UWSN which focused on the underwater target monitoring. Usually, the monitoring network need cover a wide underwater area from hundreds to thousands square kilometers. In order to cover such wide area and obtain various types of monitoring data, a large number of heterogeneous network nodes with different functional need to be deployed in the target area. Fixed nodes, such as surface and subsurface buoys, are deployed at different depths in underwater to acquire data from different locations. Mobile nodes are required to navigate over a specified or the entire area to enable flexible monitoring and data acquisition tasks independently or collaboratively as a cluster. This kind of network needs to provide the adaptive interaction links for different kinds of data transmission. For example, early warning data and command data need to be provided with a stable and reliable link and have a high demand for real-time performance. Instead, we can tolerate long latency and even packet loss for general monitoring data.

In response to the above issues, we design and implement a hierarchy based UWSN called uw-Cellular Network. It can not only cover a larger area but also accommodate more different types of nodes. In this application scenario, network nodes, whether they are fixed or mobile, can join the network either as clusters or as independent individuals for different tasks. It is worth pointing out that as mobile nodes join the network, frequent changes in the network topology and low degradation in communication success rate have put the information interaction within the network to a greater test. The uw-Cellular network provides reliable data transfer and conflict avoidance strategies based on the characteristics of mobile in nodes in the network. Simultaneously, in this network, users are able to interact with the network system via controlling the behavior, checking the status and setting the parameters of the designated network nodes.

3.2. Network Architecture

In this paper, design a large-scale network based on the cellular network architecture called uw-Cellular network. The uw-Cellular network is a combination of several star topology networks of different levels which provides scalability and robustness. This network architecture is friendly to those scenarios that consist of multiple mobile network nodes. In this network, more terminal nodes can join the network easily without needing to change other configurations. In addition, failure of a single terminal node does not affect the entire network.

Figure 2 shows the topology of uw-Cellular network. There are three types of nodes in this network, the Gateway node, the Base Station (BS) nodes, and the Terminal nodes. The Gateway node have the responsibility to interconnect the network nodes above and under the water surface by means of satellite or radio communications. Several static nodes in the network act as the BS nodes of the uw-Cellular. Each BS node can cover an area, which called cell, by hydroacoustic. All cells together provide hydroacoustic coverage over larger geographical areas. In particular, the BS nodes can be connected using wired or wireless methods depending on the different application scenarios. Each BS node acts as an access point and provides access and forwarding services for all Terminal nodes within its coverage area. The Terminal nodes, which can be both fixed and mobile, access to the network by establishing a hydro acoustic connection with the BS nodes. On the one hand, when the terminal nodes are within the communication range of a particular BS node, they can interoperate directly with the shore-based center via the backbone of the network. On the other hand, if the terminal node does not have direct access to the BS nodes, usually refers to the mobile node such as AUV, it will find a more reliable and less expensive path to interact with the shore-based center based on dynamic routing strategies. Similarly, when the shore-based center needs to send instructions, it needs to find the designated node or cluster via the cellular network. In addition, the AUVs need to interact directly with each other in order to perform collaborative tasks efficiently within the dynamically changing network topology. The uw-Cellular network give subscribers advanced features over alternative solutions, including increased capacity, small battery power usage and a larger geographical coverage area.

3.3. Protocol Design

The design of network protocols needs to take full account of the network application scenarios that as mentioned above. In this sub-section, we give the design principles for the network protocols of each layer.

3.3.1. Physical Layer Protocol Design

Multicarrier modulation in the form of orthogonal frequency division multiplexing (OFDM) has received a great deal of attention due to its promise for high data rate underwater acoustic communications. In this paper, we implement OFDM modulation and demodulation for underwater acoustic communication in the physical layer.

The data frame of the OFDM modem we implemented consists of a trigger, a Preamble and several Data Blocks. The trigger is a Hyperbolic Frequency Modulated (HFM) signal, which is widely used with hydroacoustic signal detection. The Preamble is used for packet detection and synchronization. The data payload contains up to sixteen ZP-OFDM Data Blocks. The data frame structure is designed as shown in the Figure 3.

3.3.2. DataLink Layer Protocol Design

In this work, an on-demand scheduling based on reservation MAC protocol for hierarchical networks is designed. Channel resources can be reasonably scheduled according to the size of the data volume to be sent by each node. The protocol utilizes spatio-temporal characteristics to improve channel utilization while avoiding data conflicts. The protocol also incorporates an automatic retransmission mechanism to ensure the reliability of the data. The basic principle of protocol design is shown in the Figure 4.

First, the base-station node sends a trigger (TR) packet to notify the terminal nodes within its transmission range to reserve the channel with RTS packets, and then waits for these RTS packets. Second, each terminal node which receives the TR packet as well as has DATA packets to send fills the RTS packet with the size of these packets and then sends the RTS packet. Next, the terminal node whose RTS packet is received by base-station node will be assigned some time slots for sending DATA packets. The CTS packet sent by the base-station node requests each terminal nodes who reserved the channel successfully to send DATA packets at the specified time. The base-station node will wait for the DATA packets until the time expired. Afterwards, the base-station node sends ACK packet as feedback to notify each node whether the data packets were received successfully or not. At last, if base-station node has DATA or CTRL packets to send, the packets would be sent following the ACK packet. And then the base-station sends TR packet to start a new round of data transmission.

In uw-Cellular networks, which are designed as hierarchical network, the lower-level network BS nodes are not only responsible for collecting data from the terminal nodes within their coverage area, but also for transmitting data to the designated higher-level network BS nodes. Under the condition that all nodes can communicate with each other using only hydroacoustic, the low-level BS node needs to act as a terminal node of the high-level BS node and obey the scheduling of it. In the event that in some specific applications where the BS nodes of different levels are interconnected by cable, the scheduling between the lower-level and higher-level base station nodes does not need to be considered.

3.3.3. Routing Layer Protocol Design

We design a method called CLA to aware the communication link status by analyzing the characteristics of the underwater acoustic channel. And ultimately select a relatively reliable route to deliver the information and help improve the success rate of communication between two nodes.

Related Parameters: As mentioned earlier, underwater acoustic communication is affected by several important characteristics, and we will use the following characteristics to evaluate the state of the underwater acoustic communication channel.

• SNR (Signal-to-Noise Ratio): In this work, we consider the SNR together with the communication interval. In a given time period, greater weighting comes from more communications, higher $SNR$ and shorter communication intervals. We set the parameter m to the maximum number of historical values used to calculate the average, and parameter a to the number of successful communications between two nodes in time T which is the maximum length of time that can be retraced before current communication requests. $t_i$ is used to represent the $i^{th}$ interval between two adjacent communications. The weight of the SNR can be calculated as

  (1)$SN{R}_w=\left\{\begin{array}{ccc}\hfill m\times \sum _{i=1}^{m-1}\frac{SN{R}_i}{t_i}& \hfill & \hfill a\ge m\\ \hfill a\times \sum _{i=1}^{a-1}\frac{SN{R}_i}{t_i}& \hfill & \hfill a<m\end{array}\right.$

• Distance: The propagation attenuation and absorption attenuation, which is a function of communication range d and signal frequency f, can be expressed as [38]

  (2)$A(d,f)=d^n\times {\left({10}^{\frac{\alpha \left(f\right)}{10}}\right)}^d$

  In a given application scenario, the frequency of the communication signal is fixed and the attenuation is only related to the communication distance. As the communication distance increases, the attenuation becomes greater and the possibility of successful communication decreases. Therefore, in this work, the weight of distance can be described as

  (3)$D_w=\frac{1}{t_{last}}\times \frac{p}{d_{last}}$

  where $d_{last}$ is the distance of last communication, p is the success rate of communication at that distance which can be obtained from experimental data in a real environment. $t_{last}$ is the interval between the current communication request and the last communication. The shorter the communication interval or communication distance brings the greater benefit.

• Energy: The key to extending the life time of the entire network and fulfilling the intended tasks lies in balancing the energy consumption of all nodes. In consequence, the residual energy of a candidate forwarding node is an important factor in determining its suitability. This means that the node with more power remaining is more likely to be selected as a forwarding node. In particular, when the power level of a node is below a preset threshold, it is no longer selected as a forwarding node in order to guarantee the smooth operation of the other functions of the node. We set that $p_{current}$ is the current power of the candidate node, $p_{total}$ is the battery full charge capacity and $p_{THR}$ is the preset threshold. The weight of power can be described as

  (4)$P_w=\left\{\begin{array}{ccc}\hfill 100\%\times \frac{p_{current}}{p_{total}}& \hfill & \hfill p_{current}\ge p_{THR}\\ \hfill 0& \hfill & \hfill p_{current}<p_{THR}\end{array}\right.$

• Number of hops: The number of times a packet is forwarded between the source and destination nodes, which is the number of hops, has a significant impact on the entire underwater network system. On the one hand, more forwarding times means a higher probability of packet loss. On the other hand, the fewer times a packet is forwarded between the source and destination nodes, the less system energy is consumed. We need to consider the number of hops for each possible path from the source to the destination node. And the fewer the number of hops, the more weight will be given to.

Estimation Function: In the real environment, the underwater acoustic channel is time-varying and the UWSNs can be considered as the uncertain system. The prediction of whether two nodes will be able to communicate successfully requires a combination of parameters. In this paper, we use Estimated Score ($\gamma$) as the result of the communication link awareness. A high score for a communication link means that the link has a high transmission success rate, while the opposite means that the communication link is in a poor state and does not easily communicate successfully. As we mentioned above, the $\gamma$ can be influenced by the $SN{R}_w$, distance between nodes $D_w$, energy $P_w$ and number of forwarding times n. So, the $\gamma$ is defined as follows:

(5)$\gamma =\frac{1}{n^2}\times \sum _{j=1}^{n}SN{R}_{wj}\times D_{wj}\times P_{wj}(n=1,2,3...).$

Forwarding Strategy: The processing strategy regards as a process of calculating the value of $\gamma$ for all possible paths to the destination node and find the maximum value. In this method, all the parameters needed to calculate the $\gamma$ must be stored in an Information Table. And they are contained in the valid data packets that interacting between network nodes without needing send extra packets periodically. What’s more, in order to maintain and update the information promptly, every network node needs to process the packets which received in the physical layer whether or not it is a destination node and extracts the required information. And the idea of cross-layer information interaction is applied to obtain these information before calculating. Algorithm 1 and Algorithm 2 describe the different processing of sending and receiving in the forwarding strategy, respectively.

3.3.4. Transport Layer Protocol Design

In the transport layer, we provide two mechanisms to guarantee the reliable data transmission. In some cases, retransmission mechanism is a basic approach to ensure the reliability of uw-Cellular. In other cases where higher reliability is required, reliable coding mechanisms can be used as a supplement to ensure data reliability.

We designed a protocol named Reliable Segmentation Transport Protocol (RSTP) for uw-Cellular network. When the transport layer gets a jumbo message from application layer that beyond the current ability to transmit, the RSTP breaks the message into a group of smaller segments. The receiving end must wait for all segments which can be reassembled into the original jumbo messages and pass them to the application layer with no error. Each segment has several additional fields to make it possible to provide selective retransmission service when packet loss occurs. Reliable network coding is performed for each data segment to increase the probability of successful data parsing at the receiving end when the communication channel quality is poor.

Algorithm 1: Sending side processing flow.

Input: The pre-defined minimum Estimated Score ${\gamma }_{min}$. The destination address of the  packet $I{D}_{dst}$. Send Queue $q_s$.Output: The highest Estimated Score ${\gamma }_{max}$. Send out the packet or put the packet in the $q_s$.  1:. Get a packet from upper layer.  2:. Obtain the $I{D}_{dst}$ from the packet.  3:. Calculate the $\gamma$ of the possible path, and record ${\gamma }_{max}$.  4:. if$$ > ${\gamma }_{min}$ then  5:.    Update data header information.  6:.    Sending packet to lower layer.  7:. else  8:.    if The packet type is Send Anyway then  9:.      Select the path represented by ${\gamma }_{max}$, and mark ${\gamma }_{max}$ need to be recalculated.  10:.    else  11:.      Putting packet into the $q_s$.  12:.      Send a request to send packet in a flood manner.  13:.      Waiting for feedback packet.  14:.      if Timeout then  15:.         Perform step 9.  16:.      else  17:.         Perform step 3.  18:.      end if  19:.    end if  20:. end if

Algorithm 2: Receiving side processing flow.

Input: The highest Estimated Score ${\gamma }_{max}$. The source address of the packet $I{D}_{src}$. The    destination address of the packet $I{D}_{dst}$. The next hop address of the packet $I{D}_{nxt}$. The    local address $I{D}_{loc}$.

Output: Receive the packet or forward the packet.

1:. Get a packet from lower layer.  2:. Obtain the $I{D}_{src}$, $I{D}_{nxt}$, $I{D}_{dst}$ from the packet.  3:. Obtain the information from physical layer and packet header.  4:. if$I{D}_{dst}$ == $I{D}_{loc}$ then  5:.    Update routing table information.  6:.    Sending packet to upper layer.  7:. else  8:.    if $I{D}_{nxt}$ == $I{D}_{loc}$ then  9:.      Update routing table information.  10:.      if The packet is marked. then  11:.         Enter the step 2 of Sending side processing flow  12:.      else  13:.         No need to recalculate the ${\gamma }_{max}$ and send the packet directly to the lower layer  14:.      end if  15:.    else  16:.      Update routing table information only.  17:.      Send a request to send packet in a flood manner.  18:.    end if  19:. end if

3.3.5. Application Layer Protocol Design

The protocols in this layer provide the rules and formats that govern how data is treated. Specific application protocol defines the processes and methodology to deal with the specific type of data including formatting, packing, sending and receiving.

According to the application requirements of the monitoring network, we provide dedicated protocols to format and parsed the sensor data, such as CTD, Current Meter and navigation data from AUV. In addition to that, we also proposed protocols to deal with the underwater optical and acoustic images. Protocols for processing instruction data have been also designed including checking, setting and control.

4. Simulation Study

The purpose of the simulation study we implemented is to evaluate the performance of the CLA strategy compared to the VBF and the MFLOOD protocols in the uw-Cellular network. The Aqua-Net Mate [39], a real-time virtual channel/modem simulator for SeaLinx [40], is used to complete the simulation study. The Aqua-Net Mate supports real-time schedule and provides user-friendly interfaces for parameter configuration which are more conducive to the simulation of mobile node networking.

4.1. Simulation Environment

In order to make the simulation results more realistic, we have conducted Point-to-Point communication tests in a real field environment and used the results of this test as a parameter for the simulation system. In this test, we used the OFDM-based Underwater Acoustic Modem (UAM) with the communication frequency of 21–27 kHz. The physical layer technologies are described in the previous Section 3.3.1. The maximum transmitting power consumption of this UAM is 70 W. The communication range of the UAM can achieve up to 5 km in the smooth underwater channel conditions.

As shown in Figure 5, we deployed six sets of buoys equipped with acoustic communicators at a distance of 1 km apart in the designated sea area. Among them, A1–A5 are the receiving nodes. A0 is the transmitting node and sends a 400B packet every 60 s. The whole experiment lasted for more than 48 h and A0 sent more than 3000 packets. The experimental results, which are shown in Table 1, verify that increasing the communication distance leads to a decrease in the communication success rate and the communication $SNR$. We use the results of this experiment to simulate the communication results.

In the simulation, there is only one destination node in the test area which act as the BS node. The coverage area of the base station is a circular area with a radius of 5 km. We set up three different network topologies to simulate the number of 1 to 3 hops of forwarding required between the source and destination nodes under ideal conditions. In each topology, there is only one source node, one destination node and several relay nodes. The value of the nodes’ minimum speed is 0.5 m/s and maximum speed is 2.5 m/s. Due to the mobility of the nodes, the distance between nodes varies periodically between 3 km and 5 km. The network topologies are shown in Figure 6. To increase the credibility of the results, each set of tests lasted for 24 h without intervention. Due to the packet size limitation, each node maintains at most two hops of forwarding information.

4.2. Performance Metrics

To prove the effectiveness of the proposed approach, we use three metrics as follows.

• Packet Latency indicates the time it takes for data (a packet or bits) to travel from one end of the network to the other, and typically includes transmission delay and propagation delay. It is considered to be the most important concern for users.

• Packet Delivery Fraction is defined as the ratio of the number of packets successfully received by the destination to the number of packets generated by the source. This metric reflects the effectiveness of the routing strategy which proposed in this paper.

• Effective Receive-to-Send Ratio (ERSR) means the ratio of the number of received packets and average number of sent packets per hop in the network. It reflects the relationship between the number of packets received and the total amount sent in the network. A larger ratio means fewer invalid transmissions and less total system energy consumption. Conversely, the higher the total system energy consumption.

4.3. Test Results of the Packet Latency

The purpose of the simulations presented in this section is to investigate the impact of data generation frequency and number of forwarding on data latency. All packets sent in this experiment were generated in specific length and generation interval. The packet length for this test was set to 359 Bytes which is the length of packet reported by mobile nodes. The packet generation interval is constant in each run of this test. The minimum generation interval was set as 5 s, and generation interval of each test is increased by 2 s until reached 13 s.

Figure 7 illustrate the effects of the data generation frequency on the data latency. According to the settings of CLA, every network node can obtain up to two hops of forwarding information. That means if the destination node is within two hops of the sending node, the relay node can even forward it without re-calculation. When there are more than 2 hops, there will always be some relay nodes that can forward the packets directly based on the results of the previous calculation. In contrast, VBF requires a new search for the next hop node based on information such as the radius of pipeline before each forwarding, which exacerbated the data latency in the case of node movement. MFLOOD does not require additional calculations to find the next hop. However, a backoff time is added before forwarding for conflict reduction reasons, which can significantly increase the source-to-destination data latency. Thus, the data latency of CLA are better than the VBF and the MFLOOD.

4.4. Test Results of the Packet Delivery Fraction

The purpose of the simulations presented in this section is to investigate the effects of different number of forwarding on the packet delivery fraction of these protocols. Figure 8 presents the relationship between the packet delivery fraction and forwarding hops. In this test, we do not use additional methods to guarantee the reliable transmission of data. It is clear that the metric is decreased as the number of hops grows. The packet delivery fraction of the CLA is obviously better than the VBF and MFLOOD. It is due to the fact that CLA selects the relatively most reliable path using the forwarding information which maintained locally. This forwarding strategy improves the efficiency of forwarding packets. However, VBF and MFLOOD have no ability to choose a relatively reliable path or next-hop node. That make it worse to the probability of packet loss in mobile scenarios.

4.5. Test Results of the ERSR

The purpose of the simulations presented in this section is to investigate the effects of different number of forwarding on the ERSR of these protocols. The experimental set-up is the same as the previous one. Figure 9 illustrates the results of the ERSR versus the number of hops. As shown in the figure, the metric of CLA is better than that of VBF and much better than that of MFLOOD. The reason is that CLA always select the path most likely to succeed in transmission. Smaller probability of packet loss over the entire transmission link. ERSR of CLA is much closer to 1.00 than other methods. Especially, since the MFLOOD forwards packets in a blind approach. Many unrelated network nodes are involved in forwarding packets, and most of these packets eventually fail to reach the destination node and are dropped. Thus, there are a lot of packets are sent in the network, and the ERSR of MFLOOD is the lowest of these three methods.

5. Experiments Study

As the experiment case study, we implement the network with twenty underwater network nodes and one onshore node in the real environment. The main task of the network is to enable monitoring of the target area by collecting data from underwater sensors. At the same time, users can control the network nodes by sending various instructions. Protocols in each layer are implemented to ensure that the uw-Cellular network can be applied to the targeted scenarios.

5.1. Network Components

In this network, there is a Network Management System which acts as a monitoring and control center on shore. Its main function is to store and visualize the reported underwater sensor data. It also provides user interfaces for sending various instructions to control the network nodes. One Surface Gateway Node and three Level One Base-Station (L1-BS) Nodes together form the backbone of the network in the underwater environment. They are deployed in real environments in the form of surface buoys. Especially, the Surface Gateway Node has a radio communication module which is an important relay station between management system and uw-Cellular network. Ten Fixed Terminal Network Nodes were deployed evenly within the coverage area of three L1-BS Nodes. We also deployed six Mobile Terminal Network Nodes in this uw-Cellular network. Three of these are shipboard mobile nodes and the other three are AUV nodes.

For our job, each underwater network node is equipped with an Underwater Acoustic Modem (UAM) we mentioned above. The protocols and SeaLinx network framework are implemented in the Embedded Operating Systems of the UAM.

5.2. Deployment Environment

We deployed the large-scale uw-Cellular network in the northern Bohai Strait, Liaoning, China. The field experiment lasted for about 2 months from April to May, 2021. Figure 10 illustrates the area of the network nodes deployed based on the GPS information collected during the experiment. The average water depth of the implement experiment is 40 m. The modems were deployed about 20 m below the sea surface. As the experiments were deployed in a shallow marine environment, severe multi-path effects and propagation losses of underwater acoustic transmission will occur in such environments. So, the distance between the surface gateway node and each L1-BS node does not exceed 3500 m. Similarly, all terminal nodes, except the AUVs, are deployed within 3500 m of the L1-BS nodes. The AUVs can move freely throughout the network coverage area. The entire network area covers over 100 km${}^2$. The weather was rated from moderate to rough during the experiment with wave heights between 1.25 and 2.5 m. The records of the field test are shown in Figure 11.

5.3. Performance Metrics & Experiment Results

There are some differences between evaluating the performance of the protocol and network system. Performance of a network pertains to the measure of service quality of the network as perceived by the user. There are different ways to measure the performance of a network, depending upon the nature and design of the network. In this paper, we use several metrics described below as the evaluation criteria to evaluate the performance of the uw-Cellular network.

5.3.1. Test for Network Packet Delivery Fraction & Data Rate

Network Packet Delivery Fraction indicates the ratio of packets successfully received to the total sent in the network. As the experiments were conducted in a real underwater environment, there are a variety of factors that can cause packet loss. The following two metrics were tested along with the Network Packet Delivery Fraction.

Data Rate represents the number of bits of data transmitted on the channel per second, also known as the data rate or bit rate. The unit of rate is bps (bits per second) or b/s. The main purpose of this group of experiments is to test the data rate (1) from the terminal node to the surface gateway nodes, (2) terminal nodes within the same subnet, and (3) terminal nodes in different subnets. The test results can be seen in the Table 2.

The results of this experiment verify the feasibility of the CLA protocol proposed in this paper in a real underwater environment. All network nodes, whether fixed or mobile, can perform path selection based on the communication link state evaluation strategy mentioned above. For example, we had deliberately deployed the AUV 01 outside the reliable communication range of the L1-BS 01. It selected the suitable relay node to forward data to node L1-BS 01 based on its real-time location. And finally need 3 hops and 5 hops to forward packets to the Surface Gateway Node and Terminal Node 22. However, due to the adverse effect of the Doppler and Multipath on signal demodulation, continuous and serious packet loss occurred when the waves become stronger or the mobile terminal node move faster.

It is worth noting that the performance of the communication between two fixed network nodes is obviously better than that between fixed and mobile nodes. That is because the statues of the communication links between fixed nodes is relatively stable and occasional transmission failures do not have a serious and lasting impact on these communication links. However, for mobile nodes, the potential for packet loss is greatly increased by adverse factors such as the Doppler and multipath effects, resulting in relatively low data rates. What’s more, the attitude of AUVs in the underwater also can have a negative impact on communication performance. Figure 12 illustrates the possible attitudes of AUVs in the underwater. The AUV itself, acting like a sound baffle, causes masking of the acoustic signal, which leads to a low $SNR$ of the received signal from the UAM.

5.3.2. Test for Throughput

Throughput represents the amount of data that passes through a network or interface per unit of time, including all uploaded and downloaded traffic. This set of experiments tested the network throughput under different network load conditions. In each group of tests, the Minimum of 1 and maximum of 3 terminal nodes in each level 1 subnet. The test results can be seen in the Table 3.

Figure 13 illustrates the results of each test group. Regardless of network load, throughput test results indicate significant changes, with the lowest throughput below 350 bps and the best throughput over 750 bps. It is due to the fact that the performance of UWSNs is significantly affected by the underwater communication environment. During the test, when the environment is stable, the network can reliably send and receive packets, allowing for high throughput. However, persistent and severe environmental changes resulted in a considerable fall in communication success rates and negative impact on the results of network performance test which is difficult to eliminate by optimizing network protocols.

During this field test, CTD data and AUV navigation data were collected as shown in Figure 14 and Figure 15. By adjusting the size of the sensor data and the frequency of data production, we can obtain underwater information in real time. At the same time, some network nodes are also equipped with underwater cameras to take and transmit the underwater pictures. As the transmission of the original image would consume limited bandwidth resources of underwater communication, in this case study, we have used image compression and reduction techniques to reduce resource consumption while retaining the basic information of the image. The following Figure 16 shows the underwater pictures.

In this case study, we deployed the large-scale uw-Cellular network in a real-world environment and the experimental results showed that the Data Rate between any nodes could reach above 500 bps, and the network Average Throughput was no less than 550 bps under various load situations. The mobile node can choose the relatively optimal path to transmit data according to the routing strategy proposed in this paper.

6. Conclusions

In this article, we proposed a large-scale UWSN with the main purpose of underwater monitoring called uw-Cellular network. It is a hydroacoustic network with a hierarchical architecture. Different types of network nodes, whether fixed or mobile, can be flexibly added to or removed from the network according to different needs. We also give the design principles of the protocols for the uw-Cellular network.

As a simulation study, we implement the CLA protocol and evaluate its performance in the Aqua-Net Mate. The extensive simulation results show that CLA outperforms than VBF and MFLOOD protocol in terms of the Packet Delivery Fraction, Packet Latency and Effective Receive-to-Send Ratio (ERSR). As an outfield case study, an uw-Cellular network with twenty underwater network nodes and one onshore node has been implemented. Not only CLA but also other layer protocols and functions have been implemented and applied to verify that the protocol we proposed is feasible for the scenarios we targeted. We conducted the experiments in the real environment, and the results of Network Packet Delivery Rate, Data Rate and Throughput show that the uw-Cellular network not only can run stably in the real environment, but also achieve good performance.

We will continue to optimize the large-scale uw-Cellular network architecture. Hundreds of network nodes, including more fixed and mobile network nodes, will be deployed in real underwater environments. Meanwhile, we are working to identify new needs from the real scenarios and trying to provide better support for the latest technique and application advances. uw-Cellular network will evolve as the underwater network research proceeds. And we consider it can be applied in underwater environments and can solve real problems.

Author Contributions

Conceptualization, J.Z.; methodology, J.Z. and X.P.; software, J.Z., X.P. and M.L.; validation, J.Z., X.P. and M.L.; formal analysis, J.Z., X.P. and M.L.; investigation, J.Z., X.P. and J.G.; resources, J.Z. and X.P.; data curation, J.Z., X.P., J.G., T.Z. and Y.G.; writing—original draft preparation, J.Z.; writing—review and editing, J.-H.C. and Z.P.; visualization, J.Z.; supervision, J.-H.C. and Z.P.; project administration, J.-H.C.; funding acquisition, J.-H.C. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The data presented in this paper are available after contacting the corresponding author.

Acknowledgments

The authors would like to thank the anonymous Reviewers for their careful reading and valuable comments.

Conflicts of Interest

The authors declare no conflict of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Figures and Tables

Generate image description

Figure 1. Underwater wireless monitoring network is used in port security scenarios.

Generate image description

Figure 2. Topology of uw-Cellular network.

Generate image description

Figure 3. OFDM data frame structure.

Generate image description

Figure 4. Scheduling diagram of SDDA.

Generate image description

Figure 5. Point-to-Point communication tests in a real field environment.

Generate image description

Figure 6. Three different network topologies. (a) One hop to Destination. (b) Two hop to Destination. (c) Three hop to Destination.

Generate image description

Figure 7. Test results of the Packet Latency. (a) Data Latency when need One hop. (b) Data Latency when need Two hop. (c) Data Latency when need Three hop.

Generate image description

Figure 8. Packet delivery fraction of different hops.

Generate image description

Figure 9. ERSR of different hops.

Generate image description

Figure 10. The area of uw-Cellular network deployed.

Generate image description

Figure 11. The records of the field test.

Generate image description

Figure 12. The possible attitudes of AUVs in the underwater. (a) the pitch of AUV is positive. (b) the pitch of AUV is negative.

Generate image description

Figure 13. The records of the throughput field test.

Generate image description

Figure 14. The CTD sensor data.

Generate image description

Figure 15. The AUV navigation data. (a) the AUV navigation data received by the Surface Gateway Node. (b) navigation trajectory plotted from the received data.

Generate image description

Figure 16. Comparison of the original image with the processed image. (a) the original image which take in the underwater environment. (b) the processed image which shown in the shore-based center.

Table 1

P2P communication test results in underwater environment.

Communication Distance	Communication Success Rate	Average SNR
0–1 km	98%	12
1–2 km	90%	10
2–3 km	80%	7
3–4 km	65%	5
4–5 km	45%	3.5
>5 km	10%	1.5

Table 2

Data rate test results for different data links.

Test Data Link	Source Node	Destination Node	Number of Hops	Amount of Data(bit)	Network PacketDelivery Fraction	Data Rate(bps)
Between terminal nodes andSurface Gateway Node	AUV 01	Surface Gateway Node	3	144,936	80.5%	574.1
	USV 21	Surface Gateway Node	2	57,096	92.9%	588.6
	Terminal Node 33	Surface Gateway Node	2	87,840	90.9%	625.4
Between terminal nodeswithin the same subnet	USV 11	Terminal Node 12	2	17,560	93.3%	675.4
	Terminal Node 23	Terminal Node 24	2	21,560	100.0%	718.7
Between terminal nodesin different subnets	AUV 01	Terminal Node 22	5	10,848	78.6%	542.4
	Terminal Node 23	Terminal Node 34	4	9336	92.7%	666.9

Table 3

Throughput test results under different network load conditions.

Network Load	Amount ofData (bit)	Data Latency(s)	Network PacketDelivery Fraction	Average Throughput(bps)
One Terminal Nodein Each Subnet	290,464	490	85.7%	592.783
Two Terminal Nodein Each Subnet	489,623	868	80.6%	564.081
Three Terminal Nodein Each Subnet	612,588	1111	71.7%	551.384

References

1. Cui, J.H.; Kong, J.; Gerla, M.; Zhou, S. Challenges: Building Scalable and Distributed Underwater Wireless Sensor Networks (UWSNs) for Aquatic Applications. Channels; 2005; 45, pp. 22-35.

2. Akyildiz, I.F.; Pompili, D.; Melodia, T. Underwater acoustic sensor networks: Research challenges. Ad Hoc Netw.; 2005; 3, pp. 257-279. [DOI: https://dx.doi.org/10.1016/j.adhoc.2005.01.004]

3. Sozer, E.M.; Stojanovic, M.; Proakis, J.G. Underwater acoustic networks. IEEE J. Ocean. Eng.; 2000; 25, pp. 72-83. [DOI: https://dx.doi.org/10.1109/48.820738]

4. Felemban, E.; Shaikh, F.K.; Qureshi, U.M.; Sheikh, A.A.; Qaisar, S.B. Underwater Sensor Network Applications: A Comprehensive Survey. Int. J. Distrib. Sens. Netw.; 2015; 11, 14. [DOI: https://dx.doi.org/10.1155/2015/896832]

5. Domingo, M.C. An overview of the Internet of underwater things. J. Netw. Comput. Appl.; 2012; 35, pp. 1879-1890. [DOI: https://dx.doi.org/10.1016/j.jnca.2012.07.012]

6. Ali, M.; Miraz, M.H. A Review of Underwater Acoustic, Electromagnetic and Optical Communications. Proceedings of the International Conference for Emerging Technologies in Computing; London, UK, 19–20 August 2020; Springer: Cham, Switzerland, 2020; pp. 86-97.

7. Ranjan, A.; Ranjan, A. Underwater wireless communication network. Adv. Electron. Electr. Eng.; 2013; 3, pp. 41-46.

8. Stojanovic, M.; Preisig, J. Underwater acoustic communication channels: Propagation models and statistical characterization. IEEE Commun. Mag.; 2009; 47, pp. 84-89. [DOI: https://dx.doi.org/10.1109/MCOM.2009.4752682]

9. Jiang, Z. Underwater acoustic networks–issues and solutions. Int. J. Intell. Control. Syst.; 2008; 13, pp. 152-161.

10. Curtin, T.B.; Bellingham, J.G.; Catipovic, J.; Webb, D. Autonomous oceanographic sampling networks. Oceanography; 1993; 6, pp. 86-94. [DOI: https://dx.doi.org/10.5670/oceanog.1993.03]

11. Ramp, S.; Davis, R.; Leonard, N.; Shulman, I.; Chao, Y.; Robinson, A.; Marsden, J.; Lermusiaux, P.; Fratantoni, D.; Paduan, J. et al. Preparing to predict: The Second Autonomous Ocean Sampling Network (AOSN-II) experiment in the Monterey Bay. Deep. Sea Res. Part II Top. Stud. Oceanogr.; 2009; 56, pp. 68-86. [DOI: https://dx.doi.org/10.1016/j.dsr2.2008.08.013]

12. Rice, J.A.; Creber, R.K.; Fletcher, C.L.; Baxley, P.A.; Rogers, K.E.; Davison, D.C. Seaweb Underwater Acoustic Nets. SSC San Diego Bienn. Rev.; 2001; pp. 234-250.

13. Rice, J.; Diego, S. Seaweb Acoustic Communication and Navigation Networks. Proceedings of the International Conference on Underwater Acoustic Measurements: Technologies and Results; Crete, Greece, 28 June–1 July 2005; Citeseer: Princeton, NJ, USA, 2005.

14. Rice, J.; Creber, B.; Fletcher, C.; Baxley, P.; Rogers, K.; McDonald, K.; Rees, D.; Wolf, M.; Merriam, S.; Mehio, R. et al. Evolution of Seaweb underwater acoustic networking. Proceedings of the OCEANS 2000 MTS/IEEE Conference and Exhibition. Conference Proceedings (Cat. No.00CH37158); Providence, RI, USA, 11–14 September 2000; Volume 3, pp. 2007-2017.

15. Rice, J.; Green, D. Underwater Acoustic Communications and Networks for the US Navy’s Seaweb Program. Proceedings of the 2008 Second International Conference on Sensor Technologies and Applications (Sensorcomm 2008); Cap Esterel, France, 25–31 August 2008; pp. 715-722.

16. Grund, M.; Freitag, L.; Preisig, J.; Ball, K. The PLUSNet Underwater Communications System: Acoustic Telemetry for Undersea Surveillance. Proceedings of the OCEANS 2006; Boston, MA, USA, 18–21 September 2006; pp. 1-5.

17. Açar, G.; Adams, A. ACMENet: An underwater acoustic sensor network protocol for real-time environmental monitoring in coastal areas. IEE Proc. Radar Sonar Navig.; 2006; 153, pp. 365-380. [DOI: https://dx.doi.org/10.1049/ip-rsn:20045060]

18. Alcaraz Real-Arce, D.; Morales, T.; Barrera, C.; Hernandez Brito, J.; Llinás, O. Smart and networking underwater robots in cooperation meshes: The swarms ECSEL: H2020 project. Proceedings of the MARTECH 2016 “Instrumentation Viewpoint”; Lisbon, Portugal, 4–6 July 2016.

19. Morozs, N.; Henson, B.; Shen, L.; Mitchell, P.; Zakharov, Y. Poster: Full Duplex Underwater Acoustic Networking. Proceedings of the 16th International Conference on Underwater Networks & Systems; Boston, MA, USA, 14–16 November 2022; Association for Computing Machinery: New York, NY, USA, 2022; Number 23 in WUWNet’22 pp. 1-2.

20. Khan, J.U.; Cho, H.S. Data-Gathering Scheme Using AUVs in Large-Scale Underwater Sensor Networks: A Multihop Approach. Sensors; 2016; 16, 1626. [DOI: https://dx.doi.org/10.3390/s16101626] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27706042]

21. Zhuo, X.; Liu, M.; Wei, Y.; Yu, G.; Qu, F.; Sun, R. AUV-Aided Energy-Efficient Data Collection in Underwater Acoustic Sensor Networks. IEEE Internet Things J.; 2020; 7, pp. 10010-10022. [DOI: https://dx.doi.org/10.1109/JIOT.2020.2988697]

22. Rahmati, M.; Pompili, D. Probabilistic Spatially-divided Multiple Access in Underwater Acoustic Sparse Networks. IEEE Trans. Mob. Comput. TMC; 2020; 19, pp. 405-418. [DOI: https://dx.doi.org/10.1109/TMC.2018.2877683]

23. Chen, Y.D.; Lien, C.Y.; Wang, C.H.; Shih, K.P. DARP: A depth adaptive routing protocol for large-scale underwater acoustic sensor networks. Proceedings of the 2012 Oceans—Yeosu; Yeosu, Republic of Korea, 21–24 May 2012; pp. 1-6.

24. van Kleunen, W.; Meratnia, N.; Havinga, P.J.M. MAC scheduling in large-scale underwater acoustic networks. Proceedings of the International Conference on Wireless Information Networks and Systems; Seville, Spain, 18–21 July 2011; pp. 27-34.

25. Luo, J.; Yang, Y.; Wang, Z.; Chen, Y.; Wu, M. A mobility-assisted localization algorithm for three-dimensional large-scale UWSNs. Sensors; 2020; 20, 4293. [DOI: https://dx.doi.org/10.3390/s20154293]

26. Toky, A.; Singh, R.P.; Das, S. Localization schemes for Underwater Acoustic Sensor Networks—A Review. Comput. Sci. Rev.; 2020; 37, 100241. [DOI: https://dx.doi.org/10.1016/j.cosrev.2020.100241]

27. Mayberry, S.; Wang, J.; Tao, Q.; Zhang, F.; Song, A.; Hong, X.; Dong, S.; Webb, C.; Dugaev, D.; Peng, Z. First Step Towards UNet: Open-Access Aquatic Testbeds and Robotic Ecosystem. Proceedings of the 15th International Conference on Underwater Networks & Systems; Shenzhen, China, 22–24 November 2021; Association for Computing Machinery: New York, NY, USA, 2022; Number 10 pp. 1-8.

28. Kulhandjian, H.; Melodia, T. A low-cost distributed networked localization and time synchronization framework for underwater acoustic testbeds. Proceedings of the 2014 Underwater Communications and Networking (UComms); Sestri Levante, Italy, 3–5 September 2014; pp. 1-5.

29. Martins, R.; de Sousa, J.B.; Caldas, R.; Petrioli, C.; Potter, J. SUNRISE project: Porto university testbed. Proceedings of the 2014 Underwater Communications and Networking (UComms); Sestri Levante, Italy, 3–5 September 2014; pp. 1-5.

30. Peng, Z.; Wei, L.; Wang, Z.; Want, L.; Zuba, M.; Cui, J.H.; Zhou, S.; Shi, Z.; O’Donnell, J. Ocean-TUNE UCONN testbed: A technology incubator for underwater communication and networking. Proceedings of the 2014 Underwater Communications and Networking (UComms); Sestri Levante, Italy, 3–5 September 2014; pp. 1-4.

31. Ayaz, M.; Baig, I.; Abdullah, A.; Faye, I. A survey on routing techniques in underwater wireless sensor networks. J. Netw. Comput. Appl.; 2011; 34, pp. 1908-1927. [DOI: https://dx.doi.org/10.1016/j.jnca.2011.06.009]

32. Batool, S.; Khan, M.A.; Saqib, N.A.; Rehman, S. A Survey of Routing Protocols for Underwater Wireless Sensor Networks. Proceedings of the Future Technologies Conference (FTC); Vancouver, BC, Canada, 29–30 November 2017; pp. 29-30.

33. Li, N.; Martínez, J.F.; Meneses Chaus, J.M.; Eckert, M. A survey on underwater acoustic sensor network routing protocols. Sensors; 2016; 16, 414. [DOI: https://dx.doi.org/10.3390/s16030414]

34. Islam, T.; Lee, Y.K. A Comprehensive Survey of Recent Routing Protocols for Underwater Acoustic Sensor Networks. Sensors; 2019; 19, 4256. [DOI: https://dx.doi.org/10.3390/s19194256]

35. Luo, J.; Chen, Y.; Wu, M.; Yang, Y. A survey of routing protocols for underwater wireless sensor networks. IEEE Commun. Surv. Tutorials; 2021; 23, pp. 137-160. [DOI: https://dx.doi.org/10.1109/COMST.2020.3048190]

36. Han, G.; Liu, L.; Bao, N.; Jiang, J.; Zhang, W.; Rodrigues, J.J. AREP: An asymmetric link-based reverse routing protocol for underwater acoustic sensor networks. J. Netw. Comput. Appl.; 2017; 92, pp. 51-58. [DOI: https://dx.doi.org/10.1016/j.jnca.2017.01.009]

37. Xiaohong, S.; Haiyan, W.; Yuzhi, Z.; Ruiqin, Z. Adaptive Technique for Underwater Acoustic Communication; IntechOpen: London, UK, 2012.

38. Xie, P.; Zhou, Z.; Peng, Z.; Yan, H.; Hu, T.; Cui, J.H.; Shi, Z.; Fei, Y.; Zhou, S. Aqua-Sim: An NS-2 based simulator for underwater sensor networks. Proceedings of the OCEANS 2009; Biloxi, MS, USA, 26–29 October 2009; pp. 1-7.

39. Zhu, Y.; Le, S.; Pu, L.; Lu, X.; Peng, Z.; Cui, J.H.; Zuba, M. Aqua-Net Mate: A real-time virtual channel/modem simulator for Aqua-Net. Proceedings of the 2013 MTS/IEEE OCEANS-Bergen; Bergen, Norway, 10–14 June 2013; pp. 1-6.

40. Le, S.N.; Peng, Z.; Cui, J.H.; Zhou, H.; Liaod, J. SeaLinx: A Multi-instance Protocol Stack Architecture for Underwater Networking. Proceedings of the Eighth ACM International Conference on Underwater Networks and Systems; Taiwan, China, 11–13 November 2013; Association for Computing Machinery: New York, NY, USA, 2013; Number 46 in WUWNet’13 5.