PaperlayerView exact revision on arXiv

Source-linked AI summary

Underwater Optical Wireless Communications, Networking, and Localization: A Survey

Nasir Saeed, Abdulkadir Celik, Tareq Y. Al-Naffouri, Mohamed-Slim Alouini

arXiv:1803.02442v1cs.NI

TL;DR

Underwater optical wireless communication offers high data rates and low latency, but absorption, scattering, turbulence, and limited range complicate underwater networking. The paper surveys UOWN architectures and layers, including propagation, link and routing methods, transport reliability, applications, and localization. It concludes by identifying open implementation and research challenges, including misalignment, transmission range, bandwidth, energy consumption, and transceiver compactness.

  • Problem

    UOWCs face severe aquatic channel impairments and limited communication ranges, while reliable networking and localization require solutions tailored to underwater optical conditions.

  • Method

    The paper provides a comprehensive layer-by-layer survey covering UOWN architectures, physical and link layers, relaying and routing, transport protocols, applications, and localization.

  • Results

    The survey synthesizes UOWN techniques and applications across network layers and identifies open research challenges and future directions.

  • Takeaways & Limitations

    Relay-assisted UOWC can expand coverage, extend communication range, improve energy efficiency, provide cooperative diversity, and enhance end-to-end performance.

  • Takeaways & Limitations

    UOWN implementation remains limited, with further work needed on transceivers addressing misalignment, transmission range, bandwidth, energy consumption, and compactness.

Abstract

from arXiv · show

Underwater wireless communications can be carried out through acoustic, radio frequency (RF), and optical waves. Compared to its bandwidth limited acoustic and RF counterparts, underwater optical wireless communications (UOWCs) can support higher data rates at low latency levels. However, severe aquatic channel conditions (e.g., absorption, scattering, turbulence, etc.) pose great challenges for UOWCs and significantly reduce the attainable communication ranges, which necessitates efficient networking and localization solutions. Therefore, we provide a comprehensive survey on the challenges, advances, and prospects of underwater optical wireless networks (UOWNs) from a layer by layer perspective which includes: 1) Potential network architectures; 2) Physical layer issues including propagation characteristics, channel modeling, and modulation techniques 3) Data link layer problems covering link configurations, link budgets, performance metrics, and multiple access schemes; 4) Network layer topics containing relaying techniques and potential routing algorithms; 5) Transport layer subjects such as connectivity, reliability, flow and congestion control; 6) Application layer goals and state-of-the-art UOWN applications, and 7) Localization and its impacts on UOWN layers. Finally, we outline the open research challenges and point out the future directions for underwater optical wireless communications, networking, and localization research.

I. INTRODUCTION

UOWCs offer high-speed, low-latency underwater communication but face severe channel and range limitations. This survey addresses UOWN architectures, protocol layers, applications, and localization in an integrated perspective.

  • Motivation: UOWCs provide higher data rates and lower latency than acoustic and RF systems, but their communication range is limited by underwater channel conditions.Optical propagation is affected by absorption, scattering, and turbulence, among other aquatic impairments.
  • Motivation: Blue and green wavelengths experience lower attenuation in water than other optical wavelengths, motivating their use in underwater optical links.The reported favorable wavelength range is 450–550 nm.
  • Motivation: Networking solutions are needed to mitigate limited optical communication ranges, while localization supports networking protocols and location-dependent sensing applications.Existing terrestrial and underwater acoustic localization methods are not directly applicable to underwater optical channels.
  • Survey scope: The survey reviews UOWNs layer by layer, covering architectures, physical and data link layers, network and transport protocols, applications, and localization.It also identifies open research challenges and future research directions.
  • Potential architectures: UOWN architectures can be organized by spatial coverage, node mobility, and channel, with ad hoc and infrastructure-based modes among the potential designs.Infrastructure-based networks may use optical access points or optical base stations to coordinate nearby nodes.
  • Potential architectures: A generic infrastructure architecture can connect sensor nodes, underwater optical base stations, surface stations, and terrestrial RF networks through horizontal and vertical optical links.The architecture also accommodates communication with submarines and autonomous underwater vehicles.

III. PHYSICAL LAYER: ESSENTIALS OF UOWCS

UOWCs offer high data rates, low latency, and low power or size advantages, but their short range and sensitivity to aquatic impairments create major physical-layer challenges. This section compares underwater carrier waves and reviews optical propagation, channel properties, and alignment requirements.

  • A. Waves Under the Sea: A Tour of the Underwater Communications: Acoustic systems reach tens of kilometers, whereas RF systems are restricted to shallow water and extremely low frequencies with limited data rates.These constraints motivate optical systems for higher-speed underwater communication.
  • 3) Optical Waves:: UOWC supports data rates on the order of Gbps over tens of meters with very low delay, but its communication range is short.Optical propagation is nearly at the speed of light, approximately 2.25 × 10^8 m/s.
  • 3) Optical Waves:: Optical links suffer from misalignment, absorption, scattering, and multipath fading, which can cause short-term disconnection, performance degradation, and reduced range.Misalignment can result from sea-surface motion, depth variations, deep currents, and oceanic turbulence; blue or green wavelengths mitigate but do not eliminate attenuation.
  • B. Underwater Propagation Characteristics of Optical Waves: Inherent optical properties include absorption, scattering, and attenuation coefficients, while apparent properties include radiance, irradiance, and reflectance determined by beam geometry.The extinction coefficient is the sum of absorption and scattering coefficients and depends strongly on water type and depth.
  • B. Underwater Propagation Characteristics of Optical Waves: Pure seawater has negligible scattering, while coastal and turbid harbor waters exhibit increasingly severe absorption and scattering from dissolved and suspended particles.Turbid harbor water has the highest concentration of suspended and dissolved particles and the most hostile optical conditions.

C. Underwater Optical Wireless Channel Modeling

The survey organizes underwater optical channel modeling around attenuation, scattering, radiative transfer, and photon-based simulation, while noting practical tradeoffs among model fidelity, flexibility, and computational cost.

  • Modeling overview: The section covers Beer-Lambert, volume-scattering, radiative-transfer, and Monte-Carlo approaches as complementary underwater optical channel models.These methods address attenuation and scattering with different levels of physical detail and computational burden.
  • Attenuation models: Beer-Lambert modeling expresses received power through transmission power, extinction coefficient, and transceiver distance.It assumes perfect pointing and treats scattered photons as lost, limiting its representation of multipath propagation.
  • Scattering models: Volume scattering functions describe scattered intensity per incident irradiance and water volume, while integration yields the scattering coefficient.The scattering phase function is obtained by normalizing the volume scattering function by that coefficient.
  • Radiative transfer: The radiative transfer equation describes light-beam energy conservation while incorporating spatial propagation, radiance, source radiance, and scattering.Exact analytical solutions are difficult and generally require assumptions or simplifications.
  • Monte-Carlo methods: Monte-Carlo simulation models underwater propagation probabilistically by emitting and tracking large numbers of photons.It offers accurate results, easy programming, and flexibility, but suffers from time complexity, efficiency, and statistical errors.

2) Oceanic Turbulence Modeling:

Oceanic turbulence modeling adapts optical-turbulence concepts to underwater channels, analyzes depth and scintillation effects, and considers optical and aperture-based mitigation methods.

  • Motivation: Underwater channel studies have focused mainly on absorption and scattering, while oceanic turbulence has received comparatively less attention.Several studies therefore employ traditional free-space optical turbulence models because atmospheric and oceanic turbulence share physical features.
  • Modeling studies: Turbulence research examines underwater imaging, refractive-index fluctuation spectra, Gaussian beam propagation, and turbulence-dependent system behavior.These studies extend characterization beyond absorption and scattering effects.
  • Mitigation: Aperture averaging in weak oceanic turbulence can improve system performance by reducing the scintillation index.Adaptive optics have also been proposed to mitigate turbulence effects in underwater optical wireless communication and imaging.
  • Pointing and misalignment: Misalignment is modeled with a beam spread function that relates irradiance distributions, propagation distance, aperture displacement, and the scattering phase function.The model has been used to evaluate BER under misalignment and across different water types.
  • Pointing and misalignment: Sea-surface movement, transmitter divergence and elevation angles, and point-to-point misalignment have been studied using statistical or Monte-Carlo approaches.Point-to-point misalignment modeling was verified through water-tank experiments.

D. Optical Wireless Modulation Techniques

The survey compares intensity and coherent optical modulation, emphasizing implementation tradeoffs involving complexity, power and spectral efficiency, synchronization, noise, and channel variation.

  • Modulation categories: Optical modulation schemes fall into intensity modulation and coherent modulation, implemented with direct or external modulators.The categories differ in whether message information uses light intensity alone or both amplitude and phase.
  • Modulator implementations: Direct modulation offers low complexity and price but is limited in range and data rate by chirping, whereas external modulation maintains a continuous light beam.External devices modulate or gate the source intensity or phase.
  • Intensity modulation: IM/DD is prominent because it is simple and low cost, requiring no phase information at the receiver.The receiver uses a direct detector after intensity modulation.
  • Intensity-modulation schemes: OOK uses light presence and absence for binary symbols, but channel variations can severely degrade performance.Dynamic thresholding can improve OOK performance under such variations.
  • Intensity-modulation schemes: PPM improves power and spectral efficiency at the cost of transceiver complexity and tight synchronization requirements, which make jitter harmful.PWM improves spectral efficiency and immunity to inter-symbol interference, while DPIM supports asynchronous variable-length symbols but suffers error propagation.
  • Coherent modulation: Coherent modulation can provide higher receiver sensitivity, spectral efficiency, and background-noise resistance, but requires greater cost and complexity.Homodyne and heterodyne detection use a local oscillator to convert the optical carrier to baseband or an intermediate frequency.
  • Noise sources: Receiver noise includes dark current, transmitter, shot, thermal, and background noise, with background noise depending on water type, depth, and wavelength.Solar interference dominates in the euphotic zone, while deeper regions also experience blue-green bioluminescence; total background noise combines solar and blackbody components.

IV. DATA LINK LAYER: LINK CONFIGURATIONS AND MULTIPLE ACCESS SCHEMES

The data-link discussion connects underwater optical link design with cross-layer networking, emphasizing the coverage–range tradeoff, three link configurations, performance evaluation, and multiple access options.

  • Data-link functions: The data link layer transfers data between neighboring entities, detects or corrects physical-layer errors, and arbitrates shared network resources.Its arbitration role prevents frame collisions and supports collision detection and recovery.
  • Beam configuration: Wide-beam sources provide broader nearby coverage, whereas narrow-beam sources reach farther within tighter sectors, creating a divergence-angle tradeoff.In water, extinction dominates nearly collimated beams, while geometric spreading dominates broad-divergence sources.
  • Multiple access: The section addresses TDMA, FDMA, CDMA, NOMA, WDMA, and SDMA as potential multiple-access schemes for UOWNs.The comparison is framed as a cross-layer extension of link configuration and performance analysis.
  • Beam configuration: Narrow-beam transmission provides higher received power and longer range, with 90% of photons arriving within 2 ns versus 10 ns for wide-beam transmission.Narrow beams also reduce time spread through a higher proportion of ballistic photons.
  • Beam configuration: A potential narrow-beam system achieves 1 Gbps at 515 nm over about 132 m, while the comparable wide-beam setup achieves 3.5 kbps.The narrow-beam case uses 100 mW transmission, 2 cm aperture size, 87 dB attenuation loss, and 16-ary PPM with half-rate FEC.
  • Link configurations: The survey considers LoS, reflective NLoS, and retro-reflective links as the principal UOWN configurations.LoS is straightforward but mobile transceivers may require sophisticated pointing, acquisition, and tracking to remain bore-sighted.
  • Link budgets: LoS link budgets combine transmission power, transceiver efficiencies, telescope gain, concentrator gain, and path loss.Laser divergence is generally a few milliradians or less, whereas typical LEDs can have divergence angles below 140 milliradians; narrower receiver fields of view increase concentrator gain.

2) NLoS Links:

NLoS links use reflected or diffused light to reach obscured receivers and support point-to-multipoint transmission, but reflection and scattering substantially reduce link performance.

  • Diffused light reflected from the sea surface or a mirror can reach obscured receivers and facilitate point-to-multipoint transmission.
  • NLoS geometry models reflected propagation using the sea-surface annular area, transmission angle, critical angle, and telescope gain.The critical angle is the incidence angle above which total internal reflection occurs.
  • 100 MHz LoS bandwidth falls to 20 MHz for NLoS links even in clear water.Reflection losses at the sea surface and diffusion of the reflected beam account for the reduction, while multi-scattering is an additional concern.
  • Retro-reflective communication modulates a continuous source beam and reflects it back using a simple, low-power reflector.The reflector is characterized by aperture area, divergence angle, trajectory angle, and retroreflector efficiency.
  • Photon arrival rate for an LoS link varies with receiver position and orientation relative to a source fixed at the origin.

C. Error and Data Rate Performance

The survey evaluates UOWC link configurations through photon-counting IM/DD OOK and derives BER and achievable data-rate relationships from photon arrivals and noise.

  • IM/DD OOK with SiPM photon-counting detectors models photon arrivals as Poisson distributed within each symbol slot.
  • Photon arrival rates depend on detector efficiency, data rate, Planck’s constant, and the speed of light.
  • For large photon counts, Poisson arrivals are approximated by a Gaussian distribution to derive the bit error rate.The BER expression incorporates binary-one and binary-zero arrival rates, background illumination noise, and dark-count noise.
  • Achievable data rate is obtained from the BER relationship for a given target BER.
  • 3.8 × 10^-4 is the recommended ITU-T FEC-BER threshold for hard-decision forward error correction.Errors below this threshold can be identified and corrected by HD-FEC.

D. Multiple Access Schemes

UOWN multiple access schemes address intracell interference through electrical or optical multiplexing, with each approach trading bandwidth, energy, spectral, spatial, or implementation complexity.

  • Intracell multiple access interference remains a primary challenge, motivating resource allocation, multicarrier transmission, and multiple access protocols.Intercell interference is expected to remain low because of severe aquatic channel impairments and can be reduced through OBS deployment.
  • Multiple access schemes comprise electrical TDMA, FDMA, CDMA, and NOMA, plus optical WDMA and SDMA.
  • TDMA assigns non-overlapping synchronized time slots, while its low propagation-delay penalty is less severe in UOWCs than in acoustic systems.
  • FDMA provides high spectral efficiency and ISI robustness but loses energy efficiency as the number of subcarriers increases.
  • Optical OFDM enforces real and unipolar signals using Hermitian symmetry and either DCO-OFDM or ACO-OFDM positivity schemes.Hermitian symmetry sacrifices half the bandwidth.
  • OCDM distinguishes channels with optical orthogonal codes across time, wavelength, or combined two-dimensional domains, while OCDMA supports asynchronous resource sharing.
  • NOMA superposes users at distinct power levels and uses successive interference cancellation, but no study targeting NOMA for UOWCs was identified.
  • WDMA dedicates wavelengths and tunable optical filters to users, whereas SDMA exploits spatial distribution and beam directivity for parallel transmission.A cited VLC study reports SDMA throughput 10 times higher than conventional TDMA.

V. NETWORK LAYER: RELAYING TECHNIQUES AND ROUTING PROTOCOLS

Relay-assisted UOWC extends limited optical coverage through serial or parallel relaying, while routing must account for aquatic propagation, beam control, interference, mobility, and localization.

  • Relaying techniques: Relay-assisted UOWC can expand coverage and range, improve energy efficiency and cooperative diversity, and enhance end-to-end performance.
  • Serial relaying: Serial relaying uses nodes along a routing path and is especially useful for extending range and cell coverage.
  • Serial relaying: Narrow-beam multihop transmission concentrates received power but requires directional beams and rapid PAT to handle turbulence and platform motion.
  • Serial relaying: Robust adaptive divergence and power control are essential for narrow-beam multihop transmission.
  • Parallel relaying and relay selection: Parallel relaying lets neighboring nodes jointly forward traffic as a virtual antenna array, but relay selection must account for divergence, received power, and communication range.
  • Parallel relaying and relay selection: Relay node k can become a bottleneck when carrying two streams because multiple access interference prevents simultaneous service without an efficient access scheme.
  • Relaying methods: DF decodes and re-encodes each hop to limit background-noise propagation, whereas AF amplifies signals and BDF forwards detected bits without error correction.DF can add power consumption and encoding/decoding delay; all-optical AF reduces electronic requirements.
  • Routing protocols: Routing protocols surveyed include location-based, source-based, hop-by-hop, cross-layer, clustered, and reinforcement-learning-based approaches.Location-based routing forwards data using node, target-area, and neighbor location information.

2) Source based routing:

The survey contrasts source-based, hop-by-hop, cross-layer, cluster-based, and learning-based routing approaches for UOWNs. It also frames connectivity as a central transport-layer concern and analyzes how sector-graph parameters affect connectedness.

  • 2) Source based routing:: Source-based routing selects a source-to-sink path, with minimum-delay selection reported to outperform traditional protocols in delay, energy consumption, and packet delivery ratio.Source-based routing can reduce routing energy consumption, while nodes along the selected route forward data toward the sink.
  • Hop-by-hop routing: Hop-by-hop routing lets relay nodes select the next hop, improving flexibility and scalability but not necessarily producing an optimal route.Channel-aware, beamwidth-aware, and adaptive-depth variants incorporate channel, depth, direction, or distance information.
  • Cross-layer and cluster-based routing: Cross-layer routing combines information across layers to address scheduling, routing policy, and power control while considering delay, distance, channel conditions, and buffer size.The survey also identifies cluster-based routing as especially suitable for infrastructure UOWNs, using cluster heads as gateways.
  • Learning-based routing: Learning-based routing uses Q-learning or machine learning to adapt to topology changes, account for node energy, and improve network lifetime.A layered protocol is also described for hybrid acoustic-optical networks, where upper-layer cluster heads supervise lower-layer routing.
  • Connectivity analysis: Connectivity analysis models multihop UOWNs as random sector directed graphs with directional, range-limited links characterized by orientation, scanning angle, range, and coordinates.Experiments vary node count, connectivity degree, transmission range, and beam angle to evaluate connected-network probability.
  • Connectivity analysis: Increasing beam scanning angle, node count, and transmission range increases the probability of a connected UOWN.The evaluation considers 100- and 500-node deployments and cases where each node connects to at least one or two others.

B. Reliability

UOWN reliability is constrained by packet losses from both hostile channel conditions and congestion, while application-layer protocols remain largely unexplored. Reliable transport therefore requires mechanisms that account for limited connectivity, shadow zones, and channel impairments.

  • B. Reliability: Packet losses in UOWNs can result from hostile underwater channel impairments and network congestion.This creates a need to distinguish channel-related losses from congestion-related losses during transport.
  • B. Reliability: ACK/NACK-based error verification may require full connectivity because optical nodes may be unable to return acknowledgments through temporary shadow zones.The passage links this challenge to high bit error rates and limited alternate paths.
  • B. Reliability: TCP is poorly matched to UOWCs when obstruction, pointing, misalignment, or channel impairments cause losses that TCP may interpret as congestion.UDP may suit very simple transmissions, while hop-by-hop reliability does not necessarily guarantee end-to-end reliability.
  • C. Congestion and Flow Control: Congestion control prevents oversubscription, whereas flow control regulates sender transmission rates to prevent receiver buffer overruns.Window-based TCP mechanisms can be unsuitable when round-trip times have high delay and variance, motivating UOWN-specific controls.
  • C. Congestion and Flow Control: Lower-layer information should be leveraged to predict and handle shadow zones because connectivity is the main delimiter of potential UOWN transport protocols.This requirement connects transport-layer design to network connectivity conditions.
  • Application layer: A potential UOWN application layer would query the network, advertise events, assign tasks, and manage lower-layer hardware and software.The survey characterizes application-layer protocol research as completely unexplored.

3) Navigation:

UOWN navigation and localization support applications including exploration, surveillance, monitoring, and disaster prevention, but underwater mobility, harsh channels, and absent GPS create substantial estimation challenges. The survey therefore emphasizes localization techniques specifically developed for UOWNs.

  • 3) Navigation:: Irregular, dark underwater habitats make navigation challenging and require assistance from underwater sensing and localization systems.Surface navigation technologies cannot be directly used underwater because the transmission medium differs.
  • 3) Navigation:: Existing terrestrial and acoustic localization techniques are not directly applicable to UOWNs because the optical channel introduces new challenges.The survey reports ToA- and RSS-based distributed and centralized localization schemes for UOWNs.
  • Applications: UOWN applications span environmental monitoring, underwater exploration, habitat monitoring, surveillance, military operations, and disaster monitoring.Examples include water-quality measurement, cable and pipeline monitoring, marine-life observation, mine detection, and tsunami or flood warning.
  • 3) Navigation:: UOWN localization is needed for resource exploration, surveillance, environmental monitoring, and disaster prevention.Localization also supports data tagging, object detection, and tracking underwater nodes.
  • Localization challenges: Underwater localization faces challenges from node deployment, node mobility, harsh channel variation, and synchronization without GPS signals.These factors affect anchor placement, position stability, range measurements, and ToA accuracy.

1) Acoustic Ranging:

The survey reviews acoustic localization ranging methods and contrasts them with optical UOWN positioning. It covers time-, difference-, strength-, and angle-based approaches, together with baseline systems and optical anchor-based estimation.

  • 1) Acoustic Ranging:: Acoustic localization algorithms are classified into time-based, received-signal-strength-based, and angle-based ranging methods.Acoustic channels exhibit attenuation and spreading losses that influence ranging performance.
  • ToA-based acoustic ranging: ToA ranging estimates underwater distance from signal arrival time and commonly uses at least three anchors for trilateration.Its accuracy is affected by acoustic dispersion, multipath fading, and Doppler spread.
  • TDoA-based acoustic ranging: TDoA ranging avoids requiring synchronization between sensor and anchor nodes by using range differences from multiple anchors.The survey describes a silent-positioning system based on range differences from four anchors.
  • RSS-based acoustic ranging: RSS ranging is less suitable for underwater acoustics because attenuation and spreading losses complicate distance estimation, although it has been used at certain depths.The cited work applies RSS to underwater source localization using propagation modeling.
  • Optical ranging: Optical UOWN localization uses ToA and RSS methods, including linear least squares from multiple optical anchors and RSS-based distance estimation for a given data rate.Optical propagation causes angular, temporal, and spatial widening and attenuation.
  • Baseline systems: Underwater acoustic monitoring systems use long-baseline and short-baseline approaches, with triangulation or trilateration applied after ranging.These systems support applications including AUV tracking, global network views, and intruder detection.

C. Localization Techniques for UOWNs

UOWN localization schemes address the inapplicability of terrestrial and acoustic methods underwater, using distributed or centralized architectures with ToA or RSS ranging. These approaches support node positioning but remain constrained by channel impairments, connectivity, and channel-model accuracy.

  • Terrestrial and acoustic localization methods cannot be directly applied to UOWNs, motivating novel underwater optical localization schemes.
  • Distributed schemes let each optical sensor node localize itself through communication with multiple anchor nodes, whereas centralized schemes report locations from a buoy or sink.
  • Distributed Localization Schemes: ToA localization estimates distances from at least three optical base stations and uses a linear least-squares solution for two-dimensional positioning.
  • Distributed Localization Schemes: RSS localization also requires signals from at least three optical base stations, offers low cost, but depends strongly on precise channel modeling.
  • Centralized Localization Schemes: Centralized multihop schemes use RSS-based neighborhood distances and distance-completion strategies to estimate the global network view under limited transmission ranges.
  • Performance Comparison: A comparison simulation used 50 randomly deployed optical sensor nodes in a 50 m × 50 m area and four corner anchors for distributed localization evaluation.

IX. FUTURE PERSPECTIVES OF UOWNS

Future UOWN research must address channel modeling, networking, localization, transceiver implementation, energy sustainability, and IoUT integration. The survey also identifies cross-layer optimization and practical testing as important directions toward deployable systems.

  • 1) UOWC Channel Modeling: Accurate and simple UOWC channel models are needed because analytical models simplify photon propagation, while computational models may be too complex for large networks.
  • 2) Networking Issues: UOWN research remains concentrated at the physical layer, while limited optical range and directivity create connectivity challenges requiring adequate protocols and architectures.
  • 2) Networking Issues: Routing algorithms are needed to extend communication range and coverage, but existing UOWN routing research remains insufficient beyond a centralized multihop scenario.
  • 3) Cross Layer Design Issues: Cross-layer optimization could adapt divergence angle, transmission power, communication range, and routing paths to channel conditions and application QoS requirements.
  • 4) Localization Issues: Distributed localization is important for online monitoring, yet severe UOWC channel conditions make accurate distributed schemes challenging.
  • 5) UOWN Implementation Issues: Practical UOWN implementation requires transceivers addressing link misalignment, limited range, low bandwidth, energy consumption, and compactness, alongside real-environment testing.
  • 6) Energy Harvesting for UOWNs: Energy harvesting is promising for extending network lifetime, but most existing techniques target outdoor environments and are not applicable underwater.
  • 7) Internet of Underwater Things: Multi-hop UOWNs are identified as a potential technology for implementing IoUTs, whose routing, scheduling, and data analytics research remains in its infancy.
Loading 1803.02442v1…