Enhancement of Precise Underwater Object Localization

Underwater communication applications extensively use localization services for object identification. Because of their significant impact on ocean exploration and monitoring, underwater wireless sensor networks (UWSN) are becoming increasingly popular, and acoustic communications have largely overtaken radio frequency broadcasts as the dominant means of communication. The two localization methods that are most frequently employed are those that estimate the angle of arrival and the time difference of arrival. The military and civilian sectors rely heavily on UWSN for object identification in the underwater environment. As a result, there is a need in UWSN for an accurate localization technique that accounts for dynamic nature of the underwater environment. Time and position data are the two key parameters to accurately define the position of an object. Moreover, due to climate change there is now a need to constrain energy consumption by UWSN to limit carbon emission to meet net‐zero target by 2050. To meet these challenges, we have developed an efficient localization algorithm for determining an object position based on the angle and distance of arrival of beacon signals. We have considered the factors like sensor nodes not being in time sync with each other and the fact that the speed of sound varies in water. Our simulation results show that the proposed approach can achieve great localization accuracy while accounting for temporal synchronization inaccuracies. When compared to existing localization approaches, the mean estimation error (MEE) (MEE) and energy consumption figures, the proposed approach outperforms them. The MEEs is shown to vary between 84.2154 and 93.8275 m for four trials, 61.2256 and 92.7956 m for eight trials, and 42.6584 and 119.5228 m for 12 trials. Comparatively, the distance‐based measurements show higher accuracy than the angle‐based measurements.


10.1029/2023RS007782
2 of 29 for developing network design and routing algorithms.Because nodes move around while submerged, these protocols must regularly save location updates.This circumstance results in a very high data load and significant energy consumption.Similarly, to Terrestrial Wireless Sensor Networks (TWSN), the sensor nodes require batteries for operation; however, replacing or recharging such batteries in a marine setting presents several challenges.Therefore, the upkeep of sensor node availability and the extension of the network's lifetime presents a formidable challenge to any UWSN approach.Underwater localization is a problematic issue because of the harsh conditions of the ocean, such as its restricted bandwidth, long propagation delay, spreading, and so on.Figure 1 depicts the UWSN according to their system architecture.
Underwater WSNs are the foundation for various applications that manage observed data.This sensor node can be in various forms, including static, mobile, and hybrid nodes, all of which send data via a wireless network.While Global Positioning System (GPS) and RF IDentification (RFID) are today the most often used technologies for terrestrial localization, WSNs and several other technologies are paving the way for the future.However, RF transmissions are severely attenuated underwater, and underwater sensor networks can only use RF signals ranging from 30 to 300 Hz.As a result, either a powerful signal or a large antenna is necessary.Some characteristics of underwater sensor networks set them apart from their terrestrial counterparts.The physical parameters under which underwater acoustic channels operate are often considered to impose severe bandwidth constraints.Similarly, optical signals are attenuated and dispersed significantly in aquatic environments (J.Luo et al., 2018;Sathish, Hamdi, Chinthaginjala, Pau, et al., 2023;Sathish, Hamdi, Chinthaginjala Venkata, Alibakhshikenari, et al., 2023;Sathish, Ravikumar, et al., 2023;J. Yang et al., 2015).As a result, neither of these techniques is appropriate for application in submerged environments.Sound waves are, in any event, the utmost auspicious means of communication for UWSN.Lower acoustic frequencies (10 Hz-1 MHz) have a large wavelength but a narrow bandwidth.
Management and network protocols are intrinsically linked to the network's overall architecture.Underwater localization is essential since it serves as the foundation for all other possible capabilities, such as monitoring data and mobility of nodes (Erol-Kantarci et al., 2011).When developing localization algorithms, it is essential to consider the desired quality features.These are rapid coverage, extensive coverage, high accuracy, minimal communication costs, and scalability.These elements add complications to the algorithm, which must be circumvented if we are to achieve success.In addition to localization and temporal synchronization, the problems mentioned above need unique network and transit design methodologies for UWSN.Earlier studies, such as (Ullah et al., 2017;Zarar et al., 2016), and so on, have covered some of these topics in greater detail.In the context of WSNs, pin-pointing the specific locality of each sensor node in a UWSN is referred to as "localization."Several localization techniques for TWSNs have been proposed.In contrast, UWSNs have access to a limited number of localization approaches.The distinctive qualities of UWSN distinguish it from TWSN in fundamental ways.
Additionally, UWSNs have come a long way during the previous decade.Early warning systems for earthquakes and tsunamis, underwater martial surveillance, ocean research, celestial navigation, biological applications, and pollution control are just a few fields that can benefit UWSNs (Ullah et al., 2017).However, localization in an underwater environment poses a unique set of challenges due to factors such as the depth-dependent speediness of sound and the motion of sensor nodes due to activities like shipping and water current.Additional challenges are given by an underwater setting, such as the deployment of nodes, fluctuations in signal intensity, time synchronization, variations in sound speed, and acoustic wave characteristics, to name a few.Problems with energy efficiency, localization, and routing protocols are just a few examples of the many that still need to be addressed in the UWSN.Once a sensor node is localized, the observed data can be understood.Many localization mechanisms have been designed for WSNs, but they cannot be used in UWSNs without significant modification.
In the field of UWSNs, there has recently been a surge in the amount of interest in using distributed antenna systems to connect to wireless communication networks.In a WSN, individual antennas are dispersed and connected by UWANs, an external connection that connects sensor nodes via radio (Carroll et al., 2014;Sathish, Hamdi, Chinthaginjala, Pau, et al., 2023;Sathish, Hamdi, Chinthaginjala Venkata, Alibakhshikenari, et al., 2023;Sathish, Ravikumar, et al., 2023).Two or more of the internal sensor components of a sub-merged or acoustically isolated by cluster and cluster head sensors followed by sink and base station, as shown in Figure 2.
A variety of commercially available underwater navigation systems perform their own self-localization based on readings of direction and speed.When put through their paces in a laboratory context, some of these algorithms, 10.1029/2023RS007782 3 of 29 on the other hand, demonstrate a navigational function that is dependable across relatively short distances.In contrast, the cumulative errors in these systems often cause a decline in their performance over time, resulting in a loss of precision.As a result, network localization algorithms must use both range approaches and submerged acoustic emissions as essential components.It is within the realm of possibility for sensor nodes to independently estimate their depth, possibly through the utilization of pressure probes.In order for these methods of localization to be effective, it is necessary to acquire distance readings from a minimum of three anchor nodes or other reference nodes that are already known (Diamant & Lampe, 2013).Because of the high attenuation of acoustic signals when traveling through water, the topology of the positioning network will probably be impeded.
Information gathered by sensor nodes in a two-dimensional underwater sensor network is gathered at anchor nodes placed at various depths around the ocean.The anchored nodes and the submerged sinks can communicate via acoustic linkages.The sinks collect data from the sensor nodes and send it to the offshore base station through the surface station.As a result, we can now purchase sinks outfitted with horizontal and vertical transceivers.While the vertical transceiver communicates with the base station, the horizontal transceiver communicates with  the sensor nodes to collect data and send commands.Because of the greater depth at which a vertical transceiver may operate can cover a large area (Awan et al., 2019).The acoustic transceiver-equipped surface links can control parallel communication between many sinks at different depths.After that, long-range RF transmitters will establish a link between the surface and offshore sinks.
Localization algorithms are often classified into two types: Range-based algorithms and Range-free algorithms (Zandi et al., 2013).Sensor nodes in a range-based algorithm use angle or distance information to localize themselves and anchor sensor nodes.This information can be determined using time of arrival (ToA), time difference of arrival (TDoA), angle of arrival (AOA), and received signal strength indicator (RSSI).Furthermore, range-free localization makes use of connectivity information to find sensor nodes.
The primary goal of data mining in WSNs is to precisely and swiftly extract application-oriented patterns from a continuous stream of quickly changing data that originates from a sensor network.This goal can be accomplished through the use of specialized software.Because it is impossible to save all of the data under these circumstances, the data must be processed as quickly as possible (Mahmood et al., 2013;Sathish, Hamdi, Chinthaginjala Venkata, Alibakhshikenari, et al., 2023;Sathish, Ravikumar, et al., 2023).Processing high-velocity data at a higher rate is therefore required for data mining.The management of static data makes use of data mining techniques that were developed in the past.Both the multi-step and the multi-scan methods should be utilized in order to analyze static data sets.The data that WSNs produce cannot be mined efficiently using traditional data mining techniques because of its high dimensionality, massive volume, and distributed nature.
Underwater communication and positioning are indeed areas of ongoing research and development due to the challenges posed by the dynamic underwater environment and increasing interference.While it's important to recognize the significance of accurate and precise underwater positioning, it is also crucial to ensure that research in this field incorporates innovative approaches.
We can enhance its innovativeness and contribute to the advancement of underwater communication and positioning research by considering the following aspects: 1. Novel Techniques: Investigate and propose new techniques or methodologies that can overcome the existing limitations in underwater positioning.This could involve incorporating advancements in signal processing, machine learning, or sensor technologies specifically tailored for underwater environments.2. Multi-Sensor Integration: Explore the fusion of multiple sensors or data sources, such as acoustic, optical, or inertial sensors, to improve the accuracy and reliability of underwater positioning systems.Developing innovative algorithms that combine information from different sensors can lead to more robust positioning solutions.

Cooperative Localization: Investigate cooperative localization techniques that leverage collaboration among
underwater nodes or vehicles to enhance positioning accuracy.This could involve designing distributed algorithms or communication protocols that enable cooperative positioning using information exchanged among networked underwater devices.4. Autonomous Underwater Vehicles (AUVs): Focus on the integration of positioning capabilities into AUVs, allowing them to navigate autonomously and accurately in complex underwater environments.Consider exploring advanced algorithms for AUV localization and path planning, taking into account factors such as underwater terrain mapping and obstacle avoidance.5. Energy-Efficient Solutions: Address the energy constraints typically encountered in underwater communication and positioning systems.Innovative techniques for optimizing power consumption, such as low-power communication protocols, energy harvesting, or energy-efficient signal processing algorithms, can contribute to longer operational lifetimes and improved system performance.6. Underwater Network Architectures: Investigate novel network architectures or communication protocols that can enhance the reliability and efficiency of underwater positioning systems.For instance, exploring the use of underwater sensor networks, underwater acoustic networks, or hybrid communication approaches can offer new perspectives on underwater positioning.
The severe physical characteristics of the undersea environment characterize UWSN and contribute to the network's limited bandwidth.Underwater environments bring a distinct set of challenges for the localization process.These difficulties result from the significant delay in transmission induced by the variable speed of sound.In this article we have proposed two effective localization methods for UWSNs: measurements based on distance and angles.The sensor nodes are first determined underwater using the proposed approaches.When it comes to the localization and detection of targets in the underwater environment, the measurement of mean estimation errors (MEEs) is second to the localization of nodes as the most crucial step.The two fundamental aspects that make up localization are the localization of sensor nodes and the measurement of MEEs while localization is in progress.The simulation findings make it abundantly evident that proposed localization algorithms can significantly cut down on the MEEs, resulting in decreased communication costs and a high level of accuracy.
The contributions of this manuscript are: 1.The design and implement the optimization of precise and efficient object localization for underwater wireless sensors network.2. Analyzes of the object localization as a function of the number of underwater wireless sensor nodes.3. Trade-off analyzes between distance-based localization and angle-based localization algorithms in the UWSN environment.4. Recommendation of an appropriate localization algorithm based on the targeted performance metric for UWSN.
The remainder of the manuscript is organized as follows.The related studies that are discussed in Section 2 include the associated work, the context, the data, the information, the UWSN communication technologies, and the underwater localization methods.In Section 3, an explanation is given for each of the localization strategies that have been suggested.In addition, the proposed design and simulation parameters are discussed in Section IV.Simulation results are assessed in Section 5.In Section 6, a conclusion is drawn on the proposed results.

Related Studies
This section explains the idea of underwater localization.Then, we will look at some more popular methods for locating underwater things.Park et al. (2019) modified the well-known ALOHA (Medium Access Convention) model in 2019 to enhance channel utilization.The new model features enhanced ALOHA-Q (UW-ALOHA-Q).Unusual activity, a reduction in the number of openings per outline, and a unified arbitrary conspiracy are suggested as ways to improve UW-ALOHA-Q (Ravikumar & Bagadi, 2019).The suggested methodology comprehensively improves utilization regarding the number of openings per outline while providing yet another arbitrary back-off mechanism to achieve impact-free planning.For subsea systems with a range of 1,000 m, UW-ALOHA-Q boosted channel usability by up to 24.6 times (Sathish, Ravikumar, Rajesh, & Pau, 2022;Sathish, Ravikumar, Srinivasulu, & Gupta, 2022).Erol et al. (2008) described that most oceanographic applications rely on localizing sensor nodes along long or short baselines (LBL or SBL).In both instances, sensor positions are deduced from auditory interactions between sensors and a network of receivers placed in predetermined places (Rx).The region of operations includes subsurface moorings and the seafloor, which are home to acoustic antennas for the LBL system.In contrast, SBL involves a spacecraft passing behind sensor nodes and using a short-range emitter source.Additionally, a vessel is used as part of a commercially available SBL localization system to locate underwater machinery.Prior to deployment, both algorithms needed substantial preparation and financial expenditures.Cheng et al. (2007) gave two types of underwater acoustic localization: range-based and range-free.The range-based approach first uses TDOA, TOA, AOA, and RSSI to calculate distances or angles to selected anchor sensor nodes, as shown in Figure 3.They then translated the ranges into several coordinate systems using multilateration and triangulation techniques.As an alternative, the range-free method forecasts the positions of sensor nodes in the network based on the locations of neighboring anchor sensor nodes.Radar, sonar, and wireless communication devices depend on accurate distance assessment of targets.The minimum variance method, conventional beam forming, the weighted subspace fitting algorithm, and the Estimation of Signal Parameters via Rotational Invariance Techniques (ESPRIT) algorithm are just a few of the distance of arrival (DOA) estimation algorithms that have been developed in the past.Biao et al. (2015) provided a DOA estimate method for underwater acoustic targets and the micro underwater localization platform.In order to do this, the authors looked into several formulations for the acoustic target localization with sensor array problem within the context of sparse signal representation.Both narrow-band and wide-band environments are compatible with the strategy.The position of a signal at a dumb node is determined by its DOA.One can determine the signal's direction by calculating the receiver's propagation delay with the reference angle, which can be worked out with the help of a direct reference (Ravikumar & Bagadi, 2017).Using this method, the AOA for a dumb node's location is found using at least three beacon nodes.To find the dumb node, it is necessary to know where at least three beacon nodes and the three AOA are.When directional antennas are used, it is possible to figure out the AOA.Directional antennas can be put on beacon sensor nodes if they are used.A directional antenna at the top of a rotating sensor node sends beacon signals in all directions (Sathish, Ravikumar, Rajesh, & Pau, 2022, Sathish, Ravikumar, Srinivasulu, & Gupta, 2022).Rahman et al. (2018) proposed that the fundamental goal of a localization strategy is to find the location of sensor nodes in a network of sensors (nodes that already know where they are) relative to or precisely concerning a small number of anchor nodes.There are two ways to accomplish this.Furthermore, the article presents a system that uses less energy and can identify and collect data on moving objects.The localization algorithm can be classified into two groups based on the approaches used to establish the location of anything: range-based and range-free.Han et al. (2015) given range-based localization methods, the position of a sensor node can be computed by measuring the angle or distance between the node and its neighbors.Range-free algorithms, on the other hand, assume that the distance or angle information gathered by neighboring sensor nodes cannot be used for positioning due to hardware limits and costs, which spreads anchor sensor nodes over all networks and uses long-range acoustic channels to communicate with buoys on the water's surface, is widely regarded as one of the best attempts at a localization method in UWSN.Isik and Akan (2009) shared that Ordinary localization and anchor node localization constitute the majority of the Localization method, which can be further subdivided into its component pieces.The messages transmitted to the ordinary sensor node originate at the anchor sensor node.The anchor communicates with the surface buoys using the anchor sensor node.Following that, an ordinary node will identify its location by calculating its distance from surface buoys in the same manner as an anchor node.As a result, it is not required because a normal sensor node can establish its location.Furthermore, the researchers assume that many stationary sensor nodes underwater have the same bearing (Bagadi et al., 2022).Some sensor nodes can run the range algorithm by transmitting messages only in one way and synchronizing their clocks, both challenging operations in UWSNs.
H. Zhang et al. (2016) reported that due to the underwater environment's features for signal propagation, UWSNs face a particular set of obstacles in developing wireless communication and network protocols.In a mobile sink design, a mobile sink moves across the network to disseminate non-information without first waiting for it to be sent by the sensors, hence avoiding multi-hop transmissions.Some networks use a method known as area partitioning to decrease the travel time between the sink sensor node and the sink and to create clusters that boost output.We suggest a transmission strategy based on superposition coding to increase the throughput of down-link command messages to sensor nodes.Emokpae et al. ( 2014) discovered that because signals transmitted by the GPS cannot penetrate water, it will be necessary to find an alternative way to locate sensor nodes.Most of these techniques needed either the alignment of two approaches or range measurements between the talking sensor nodes, such as TDoA, ToA, AoA, and RSSI.Recent years have seen a rise in the focus placed on locating sensor nodes deep within the water.The vast majority of the localization systems that have been discussed aim to establish a reference sensor node before proceeding (Emokpae et al., 2014).However, this method has a significant limitation because it requires many reference sensor nodes in a distributed network.Without these reference sensor nodes, localization is difficult, if not impossible.The high cost of electricity, transport, and other infrastructure requirements makes it unfeasible to install many reference sensor nodes in the vast majority of underwater fields.This situation is because these demands must be met.The UWSN, taken into consideration by Hu et al. (2019) comprises several sensor nodes dispersed throughout the network's physical space.In order to keep the cost of the network to a minimum, sensor nodes are developed with constrained processing capabilities and simplified computational complexities.
Because marine environments are in a permanent state of flux, the sensor nodes are in a constant state of motion, following the flow of water and reacting to activity in the marine environment.
G. Yang et al. (2018) proposed that as a consequence of these difficulties, localization needs to be done as quickly if possible; otherwise, the estimated positions will remain the same even as the sensor nodes move from one location to another.Therefore, it is essential to organize a localization process that is both quick and economical with energy in a sensor network that has limited resources.The continual motion causes specific sensor nodes to have a greater chance of moving outside of the functioning field of the network, which exacerbates recycling and sustainability issues.The brininess, temperature, and depth of the water all have an effect, in addition to the elements estimated on the rate at which the waves below move.He et al. (2018) presented two techniques for underwater target localization in the study mentioned above: nonlinear weighted least squares-based underwater target localization and space-altering generalized expectation maximization-based underwater target localization (SAGE-UTL).Submarine target localization using nonlinear weighted least squares (UTL) is also known as UTL using a state-action-event model.Based on the information collected by a swarm of dispersed star receivers, these algorithms can pinpoint the location of a target with great accuracy.The network is hypothesized to perform the functions of both a primary receiver and several additional conventional receivers.A sound speed profile (SSP) with an iso-gradient and a network anchored to the water's depth is assumed.As temperature and salinity tend to fluctuate throughout the ocean, the iso gradient SSP theory makes sense for the environment under investigation.
Additionally, Yin et al. (2016), Hao et al. (2017), and B. Zhang et al. (2018) have researched the TDoA localization algorithms and the ToA localization algorithms.The unknown source location and hybrid estimations are initially connected to evaluate a solution with a closed form.The best sensor node association is then determined.The solution is then assessed.According to all of the Cramer-Rao Lower Bound (CRLB) is the lowest bound of any unbiased estimator and can be used to transmit details about the accuracy of localization.Even when there is just a small amount of inaccuracy, the MEE matrix can be derived.However, its actual value can only be realized in the context of practical application.First, a localization technique for closed structures must be studied before using the error covariance matrix that this strategy generates to estimate the CRLB.By recasting the issue as an optimization problem to identify the ideal node association, they could convert an unsolvable issue into a convex one.They were able to solve the issue as a result successfully.
Mridula and Ameer (2018) provided a localization approach for UWSNs that considers the problems in sensor node localization caused by ambiguity in the anchor location.When the anchor is submerged, it moves a lot.This circumstance is because water currents harm the network's environment.It is easier to carry out rigorous localization when clarity is inside an anchor node.The undersea environment's ray-bending quality must be considered for accurate location readings.This situation is because the speed of sound is considerably lowered under the surface.Using ray theory, one may determine the path that sound rays take when immersed in water.
Because the positions of the anchors are inherently imprecise, it is necessary to use Maximum likelihood to determine the precise location of the required sensor node.It is compared to several methods, each of which provides precise data on the exact location of the anchor node.If the anchor nodes are unclear, CRLB is calculated to help estimate the target's location.The UWSN is a collection of sensors that work together to monitor activity in marine habitats.To achieve these objectives, sensor nodes organize themselves into self-contained networks capable of characterizing a marine ecosystem.Because they do not require cable to be put beneath the water's surface and do not interfere with marine operations, USNs are designed to be easy and affordable to outfit.This circumstance is one of their primary goals.Because of their one-of-a-kind qualities, UWSNs necessitate a fresh approach to a wide range of localization-related difficulties.

Localization
Because of the lack of essential infrastructure, underwater networks have more difficulty performing localization tasks than their terrestrial equivalents.Propagation delays, in particular, can be highly significant when bandwidth is limited.The limited capability of building modems capable of simultaneous signal transmission and reception is another constraint that must be considered while designing and implementing UWSNs.A well-prepared transmission can prevent data loss due to the near-far effect.To keep network management overhead minimal, the amount of information sent between nodes must be limited by the node discovery mechanism.Another area of speculation in UWSNs is the connectivity of the sensor nodes.Several factors exacerbate the connection process, including noise, relative node orientation, fading, and propagation losses.This connectivity is influenced by several elements, including sensor node relative motion, sensor node and link failure, sensor node installation, and a range of other issues.Even if there is no direct link between standard sensor nodes and anchor sensor nodes, networks can be built to facilitate range measurement.Depending on the network architecture, a few additional localization methods can be utilized, such as the Euclidean, DV-hop, and DV-distance.
The Euclidean distance yields some promising results when dealing with anisotropic topologies.When doing a more complex calculation, higher overhead and communication costs are incurred.A sensor node can only localize itself if its position can be determined uniquely.The sensor node cannot pinpoint its precise location if it lacks it.Even if a node cannot localize itself, many alternative locations may still be measured (H.Luo et al., 2008).This circumstance is because potential locations are more precise than actual locations.Only a small number of sensor nodes have the potential to be precisely located.The great majority of approaches to localization include the sensor node being localized by doing a partial localization with the assistance of a collection of reference sensors.Specific sensor nodes, known as reference or sink nodes, must get their location information before the sensor node must be localized.This activity will commence at the beacon sensor node as its point of departure.
It is preferable to use as little energy as possible whenever possible.It is also critical to consider the localization algorithm's level of precision.A method called UDB (underwater directional beacon) is provided in reference (Bian et al., 2009) for underwater localization.

Measurement Based on Distance
When operating in an underwater environment, sensor data is frequently interpreted based on the location of a sensor node.Following a target, keeping an eye on the environment, or reporting an event are all examples of this.As previously stated, finding something on land is more accessible than finding something underwater.This is because RF waves do not decrease as underwater as on land.GPS cannot be used underwater as a result of this.There were numerous approaches to localization in the various localization schemes (Ullah, Chen, et al., 2019, Ullah, Liu, et al., 2019).These methods consider various factors, including the device's capabilities, the rate at which the signal spreads, and the quantity of energy available, to name a few.Most systems for determining where something is considered a sensor node's location in the network field.The nodes whose placements are known are the anchor sensor nodes.Most localization techniques employ these nodes.In Poursheikhali and Zamiri-Jafarian (2015), there is a plan for locating a target based on predicting the TDoA in a non-uniform underwater field.TDoA, which stands for "target depth of approach," is the strategy's concept.Because the underwater environment is not uniform, waves follow a curved path.As a result, locating the TDoA is far more complex than locating the terrestrial position.This method, which employs the methodology, considers TDoA-based localization in an algorithmic manner that varies over time.The approach is getting closer to the CRLB and has the potential to move beyond the line-of-sight (LoS) TDoA.This situation is accomplished by considering where an asynchronous target is located and how precise that location is.Kouzoundjian et al. (2017) offer a method for calculating the TDoA between different underwater beacon signals.The algorithms for this system rely on distance measurements.The suggested approach does not require beacons and receivers to be set simultaneously for propagation to end in underwater conditions.As a result, the TDoA estimate depends on the location of the beacon sensor node.The solution is demonstrated to be a 10.1029/2023RS007782 9 of 29 series of hyperbolic equations, with the theoretical location of the node being where these hyperbolas intersect.On the other hand, one popular method for determining the TDoA is to examine how strongly the signals cross-correlate.The underwater field generates a lot of phase and amplitude distortion in the waves that are picked up because the waves bounce back and forth in the water and cause reverberation.Another method for determining the TDoA is to examine the central section of the received signals for a succession of equal zero-crossing intervals that may be used to determine when they began and how much time has passed since they began.This method entails examining the primary portion of the received signals.Valente and Alves (2016) approach is implemented as a programmable system-on-chip coupled to an embedded ARM CPU and equipped with a custom-designed digital signal processor.The strategy was tested in both a closed environment (a tank) and an open environment (a field).
Using the relative antenna, the beacon may compute the distance between itself and a stationary or mobile node.
For this reason, a Doppler speed measurement is used; however, the precision of the result depends on the position of both the mobile device and the beacon.The following are assumed to exist if N is the number of participating antenna nodes like r i , s i , t i , where n = 1, 2, ,….N (Ullah, Chen, et al., 2019;Ullah, Liu, et al., 2019): The Zero Mean Additive White Gaussian noise for the active nodes is where V n (i) is and l n (i) is In case of two sensor nodes Equation 4 will be (5)

Measurement Based on Angle
Recent research has shown that angle-based metrics are an effective method for underwater localization, and that this method is feasible.
The method described in Choi et al. (2018) provides an accurate approximation of the AoA of an audio source.Two hydrophones are mounted on a marine vehicle traveling across the water, and the directional angles of the source are measured.Utilizing the properties of acoustic waves that occur in the ocean, specific equipment can send out signals sporadically or continually.The foundation of this strategy is based on the presumption that a particular acoustic source consistently produces the same signal.An initial probability is calculated by utilizing the state transition model in the first step (Jia-Tong et al., 2018).In the second step, an algorithm known as a generalized cross-correlation is used for the already collected acoustic data to derive directional information.A comparison of the likelihood with the entropy of the current correlation is performed in the very last stage.However, the system that is being proposed needs to go into research the physical properties of a wide variety of acoustic sources depending on their frequency ranges.This situation is because such research is yet to be feasible.These measurable qualities centered on precisely measuring the directional angle of the acoustic sources to make use of the information already available regarding the frequency band (Wilding et al., 2018).
In addition, a wide variety of AoA localization schemes are utilized in Choi and Choi (2015) and Huang and Zheng (2016).We provide a technique for real-time AUV localization based on bearing estimation alone and use the depth of a beacon already known in advance.The system is based on the Extended Kalman Filter (EKF) and uses a State-Space model.This goal is done to account for the mobility of the AUV in two degrees of freedom.In a similar vein, a technique for identifying and removing acoustic target signals from a variety of underwater sources by making use of frequency bands is required.A Bayesian technique is used to derive the data on the directions, while an EKF model calculates the angles associated with those directions.In addition, a localization technique that can be used in underwater Ad-hoc networks is given.This strategy uses AoA to calculate the distance between anchors and sensor nodes in two-dimensional and three-dimensional space.Once a sensor node has received distance estimates from at least three or four anchor nodes, it will be possible to calculate the sensor node's location.
To approximate the distances and angles between nodes P and Q, which are initially located at coordinates l 1 , m 1 and l 2 , m 2 , respectively (Ullah, Chen, et al., 2019;Ullah, Liu, et al., 2019).
Checking out the two nodes, P and Q: and The distance between the sensor nodes P and Q is The angle between nodes P and Q is and the angle θ is (11)

Proposed Localization Algorithm
To enhance the precise underwater object localization using TDoA and AoA, we need to consider introducing the following innovations: 1. Hybrid Localization Algorithm: Develop a hybrid localization algorithm that combines the AoA and TDoA measurements to improve the accuracy and precision of underwater object localization.The algorithm should leverage the strengths of both measurements to mitigate the limitations of each technique.This can involve using a weighted fusion approach or a Bayesian framework to integrate the angle and distance information effectively.It is important to ensure that the proposed enhancements are aligned with the objective of improving precision, and thoroughly validate the algorithm's performance to establish its superiority over existing methods.A basic localization algorithm for an underwater WSN can be based on trilateration, which involves estimating the position of a sensor node by measuring the distances to multiple anchor nodes with known positions.Here's a simplified version of the algorithm with its mathematical equations: 1. Initialization: 1-1.Assign initial positions to anchor nodes.1-2.Initialize the sensor node positions as unknown.

Distance Measurement:
2-1.Sensor nodes measure the distances (d) to multiple anchor nodes using techniques such as ToA, TDoA, or RSSI.

Trilateration:
3-1.Select a set of anchor nodes (at least three) with known positions and corresponding distance measurements.3-2.Use trilateration to estimate the position of the sensor node based on the distances and anchor node positions.3-3.The position (x, y, z) of the sensor node can be calculated using the following equations: 3-4.Solve the system of equations to find the coordinates (x, y, z) of the sensor node.4. Iterative Refinement: 4-1.Repeat steps 2 and 3 with different sets of anchor nodes to improve the localization accuracy.4-2.Use more sophisticated algorithms like least squares estimation or maximum likelihood estimation to refine the position estimates.

Localization Update:
5-1.Periodically update the positions of the anchor nodes based on their actual movements or changes in the underwater environment.5-2.Re-estimate the sensor node positions using the updated anchor node positions and distance measurements.
It's important to note that the actual implementation of the algorithm may involve additional steps and considerations, such as error handling, filtering techniques, and robustness to deal with issues like measurement noise, multipath propagation, and localization outliers.The equations provided above represent a basic framework for trilateration-based localization in an underwater WSN and an illustration diagram as shown in Figures 4a and 4b shows the localization process flow chart.
The localization process in an UWSN involves determining the positions of sensor nodes in an underwater environment.Here's an explanation of the steps in a typical localization process flowchart for UWSN: 1. Start: The localization process begins.2. Node Deployment: Deploy the sensor nodes in the underwater area of interest.These nodes may have limited or no knowledge of their own positions.It's worth noting that the specific techniques, algorithms, and parameters used in each step may vary depending on the localization method chosen, the characteristics of the UWSN, and the environmental conditions.The flowchart above provides a general framework for the localization process in UWSNs, highlighting the key steps involved in estimating node positions in an underwater environment as shown in Figure 4b.

Proposed Hybrid Localization Algorithms of TDOA and AOA
Hybrid algorithms that combine TDoA and AoA measurements can provide more accurate and robust localization in UWSN.Here are the mathematical equations for a common hybrid algorithm known as TDoA/AoA fusion: 1. TDoA Equations: The TDoA equations relate the time differences of arrival between anchor nodes and the distances between them.Let's consider three anchor nodes A, B, and C, and a sensor node S. The TDoA equations can be written as: Here, TDoA_AB, TDoA_AC, and TDoA_BC are the measured time differences of arrival between the anchor nodes, Distance_AB, Distance_AC, and Distance_BC represent the distances between the anchor nodes, Speed_of_Sound is the speed of sound in water, and Measurement_Error_AB, Measurement_Error_AC, and Measurement_Error_BC account for any measurement inaccuracies or noise.2. AoA Equations: The AoA equations relate the angles of arrival from anchor nodes to the sensor node's position.Let's consider the angles of arrival from anchor nodes A, B, and C to the sensor node S. The AoA equations can be formulated as:   Here, E represents the overall error, and w1 to w6 are the weight factors assigned to balance the influence of TDoA and AoA measurements.The weights can be adjusted based on the expected accuracy and reliability of the measurements.
The goal is to minimize the error MEE by finding the optimal values for (x_S, y_S), representing the estimated position of the sensor node.It's worth noting that the specific implementation of the fusion equation may vary depending on the localization algorithm and optimization technique used.Additionally, considerations such as environmental factors, measurement errors, noise mitigation, and calibration techniques should be taken into account to achieve accurate localization in UWSN.9. End

Proposed Design and Simulation Parameters
We shall now look at the techniques offered for underwater localization, which are first and foremost expected to achieve underwater target localization.After finding the target location, the MEE must be estimated.It takes advantage of previously defined distance and angle data.It is critical to first estimate the location of a sensor node before attempting to estimate the MEE in target localization.The simulation attributes of the proposed design are considered in Table 1.

Measurement Based on Distance
Assessing the network field over a region of 100 m by 100 m is the first step in putting into practice the distance-based localization strategy presented here.An area measuring 100 m by 100 m is open for exploration by underwater sensor nodes.In the first scenario, we test the method in a relatively tiny region that is only 100 m squared.This situation allows us to establish how big of an impact the distance has on the accuracy of the localization.We contact the four anchor nodes at the four cardinal points of the localization network to establish where anything is situated concerning other things.In this particular instance, there are just 10 mobile nodes that roam the network field that is 100 m × 100 m.For MEE monitoring, a sensor node in an irregular position is chosen.After the position of a sensor node has been produced randomly, numerous trails are used.However, just a subset of those trails is first studied in this situation.In this particular instance, the results of four trials are analyzed, MEEs are computed, and the same is extended for eight and 12 trials.Because the beacon sensor nodes are connected to a reference antenna, it is possible to calculate the distance between a mobile sensor node and a beacon node.

Measurement Based on Angle
In this part, we discuss the methods utilized to implement the proposed angle-based measuring methodology.With distance-based measurement in mind, we start by deciding on a 100 m × 100 m rectangle as the network field within which the mobile nodes can operate.Each of the four corners of the network field contains an anchor node, while the field as a whole contains 10 mobile nodes.There may likely be some variation in the positioning of the mobile nodes.Once the nodes' random positions have been estimated, the Euclidean distance may be calculated.Once the derivatives have been calculated, then the MEEs can be calculated.In this section, we can only use 10 sensors over 100 m × 100 m.We will also cover the effects of coverage and sensor density on the precision of localization in the following sections.Since the MEEs tend to fluctuate between the selected iterations, we use an angle-based measurement method in the first scenario.Skip occasionally across, but more often between, these four, eight, and 12 versions.This situation allows us to determine the angle between sensor nodes and calculate MEEs.The variability of MEEs is mainly attributable to the ever-changing nature of marine habitats, including ocean currents and shipping activity.Even though the proposed method increases the difficulty of underwater localization, it outperforms previous localization strategies in terms of accuracy.The new innovation in this scenario is the measurement-based angle localization strategy for underwater sensor nodes.Here's an explanation of the key elements and steps involved: 1. Network Field and Anchor Nodes: The experiment is conducted within a 100 m × 100 m rectangular network field.Each of the four corners of the field is equipped with an anchor node.These anchor nodes serve as reference points for localization.2. Mobile Nodes: The network field contains 10 mobile nodes that move within the area.These nodes contribute to the localization process by measuring angles between themselves and other nodes.

Simulation Results and Discussions
The efficiency of the proposed distance and angle-based measurements was validated by research conducted underwater, providing strong evidence for their use.Two fundamental methods were utilized to accomplish the primary goals of underwater localization and MEE estimation, respectively.Both tactics are an improvement over the methods that have come before them because, first, they precisely localize the sensor nodes, and then, second, they calculate the MEEs.

Measurement Based on Distance
Measuring distance is utilized in localizing a network by determining the distance between the sensor and anchor nodes.According to this strategy, the length of the boundary between each node in the network is set at 80 m, which results in the network being in the shape of a square.Sensing nodes are not permanently installed in any one location; as a result, mobile sensor nodes are free to move around wherever they like inside this zone.There are a total of 10 wandering nodes, along with four stationary nodes in the network.Each of the sensor nodes in the network can communicate with one of the four anchor nodes, which are positioned at each of the network's four corners.The schemes have an error ratio in the calculation of distance that is 0.1 m, which equates to an accuracy in the calculation of distance that is 90%.The precision of one m, approximately 0.1, is a good illustration of this concept.Before measuring the actual distances that separate the sensor nodes, it is first necessary to use a calculation to identify a non-uniform distribution of the sensor nodes.After the location of the sensor nodes has been determined, the procedure is analyzed through several iterations, and MEEs are acquired.Many trials of this process are carried out here; four, eight, and 12 trials are considered.The MEEs tend to move back and forth between the ranges of 2.1218 and 2.6501 m for four trials, 2.0604 and 3.1748 m for Eight trials, and 0.0669 and 0.2074 m for 12 trials, as can be seen in Figure 5 and the results of the trials that are presented in Table 2 for four trials, Figure 6 and the results of the trials that are presented in Table 3 for Eight trials and Figure 7 and the results of the trials that are presented in Table 4 for 12 trials.

Measurement Based on Angle
Measurement based on angles yields results comparable to those derived from measuring distances in terms of the range and the number of sensor nodes.
The network is dispersed 100 m by 100 m, and each of its 10 mobile nodes and four anchor nodes has been selected with care.The cardinal points are home to each of the four anchor nodes that make up the network.After selecting a random pair of nodes, P and Q, as the starting point, the next step is to compute their respective locations and angles.The MEEs can be computed once the nodes have been found in the network.This angular measurement has used four, eight, and 12 trials.The MEEs tend to move back and forth between the ranges of 84.2154 and 93.8275 m for four trials, 61.2256 and 92.7956 m for Eight trials, and 42.6584 and 119.5228 m for 12 trials, as can be seen in Figure 8 and the results of the trials that are presented in Table 5 for four trials, Figure 9 and the results of the trials that are presented in Table 6 for Eight trials and Figure 10 and the results of the trials that are presented in

Conclusion
The approaches of localization that are distance-based and angle-based are both covered in this article.After the locations of the subsea nodes have been determined, the MEEs are calculated.To perform distance-based measurements, a total network field of 100 m × 100 m in which mobile sensor nodes are permitted to roam has been established.There are 10 wandering nodes in the network, with the anchor nodes situated in the four corners of the network.When taking a reading of the MEE, the position of a sensor node is picked at random.After the random placements of the sensor nodes have been picked, different trials are applied; however, in the initial scenario, only a tiny subset of those trials are considered.The MEEs are computed after assessing six distinct combinations of the number of trials.The MEEs tend to move back and forth between the ranges of 2.1218 and 2.6501 m for four trials, 2.0604 and 3.1748 m for Eight trials, and 0.0669 and 0.2074 m for 12 trials, as can be seen in Figure 5 and the results of the trials that are presented in Table 2 for four trials, Figure 6 and the results of the trials that are presented in Table 3 for Eight trials and Figure 7 and the results of the trials that are presented in Table 4 for 12 trials.The network size for angle-based measurement is also 100 m × 100 m, which provides the mobile sensor nodes significant space to move.In each of the four corners of the square field, there is a total of 10 sensor nodes and four anchor nodes that have been placed.After angle estimations between sensor nodes have been determined, the MEEs can be computed.The MEEs can be computed once the nodes have been found in the network.This angular measurement has used four, eight, and 12 trials.The MEEs tend to move back and forth between the ranges of 84.2154 and 93.8275 m for four trials, 61.2256 and 92.7956 m for Eight trials, and 42.6584 and 119.5228 m for 12 trials, as can be seen in Figure 8 and the results of the trials that are presented in Table 5 for four trials, Figure 9 and the results of the trials that are presented in Table 6 for Eight trials and Figure 10 and the results of the trials that are presented in Table 7 for 12 trials.As seen in Tables 8-10, the measurements based on distance tend to produce more accurate findings than those based on the angle.

Figure 1 .
Figure 1.Underwater wireless sensor network system architecture.

Figure 2 .
Figure 2. Internal structure of underwater wireless sensor networks system architecture.
2. Advanced Signal Processing Techniques: Incorporate advanced signal processing techniques to enhance localization accuracy.This can include adaptive beamforming, array processing, or super-resolution algorithms to improve the quality of the received signals and reduce the effects of multipath propagation and interference.By processing the received signals more effectively, the localization accuracy can be significantly enhanced.

Figure 6 .
Figure 6.Measurement based on distance mean estimation errors for eight trials.

Figure 7 .
Figure 7. Measurement based on distance mean estimation errors for 12 trials.

Figure 8 .
Figure 8. Measurement based on angle mean estimation errors for four trials.

Figure 9 .
Figure 9. Measurement based on angle mean estimation errors for eight trials.

Figure 10 .
Figure 10.Measurement based on angle mean estimation errors for 12 trials.

Table 2
Measurement Based on Distance Mean Estimation Errors for Four Trials

Table 3
Measurement Based on Distance Mean Estimation Errors for Eight Trials

Table 5
Measurement Based on Angle Mean Estimation Errors for Four Trials

Table 8
Table 7 for 12 trials.Comparatively, the distance-based measurement is more accurate and time-efficient than the proposed angle-based measurement.When put next to the angular measurement, this is quite striking.The MEEs values obtained from distance measurements are smaller than those obtained from angle measurements.Compared to distance measurements, angular measurements are more challenging to take underwater due to the existence of impediments created by water currents.Depending on the measurement angle, MEEs can range from 42.6584 to 119.5228 m, whereas MEEs, based on distance, can swing from 0.0669 to 3.1748 m.The outcomes comparison of the data sets is provided in Table for Four trials, Table 9 for eight trials, and Table 10 for 12 trials.Measurement Based on Distance and Angle Mean Estimation Errors for Four Trials

Table 9
Measurement Based on Distance and Angle Mean Estimation Errors for Eight Trials

Table 10 Measurement
Based on Distance and Angle Mean Estimation Errors for 12 Trials Dr. Mohammad Alibakhshikenari acknowledges support from the CONEX-Plus programme funded by Universidad Carlos III de Madrid and the European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No. 801538.Additionally, the authors sincerely appreciate funding from Researchers Supporting Project number (RSP2023R58), King Saud University, Riyadh, Saudi Arabia.Besides above, the Article Processing Charge (APC) was afforded by Universidad Carlos III de Madrid (Agreement CRUE-Madroño 2023).