An Arctic underwater terrain matching method integrating template matching and DEM super-resolution

AUV path planning in complex underwater environment via an improved multi-strategy artificial lemming algorithm

Introduction The underwater environment triggers the autonomic nervous system (ANS) responses during self-contained underwater breathing apparatus (SCUBA) diving, which aims to conserve oxygen during submersion. The aim of the study is to evaluate cardiovascular responses in professional SCUBA divers by analysing heart rate dynamics across dive phases and comparing responses between novice and experienced divers in a real-world setting. Method Twenty certified divers performed standard dives to a minimum depth of 66 feet and remained there for 5 minutes as part of the protocol. The dive was divided into six phases: rest, pre-dive, descent, bottom, ascent, and post-dive. The heart rate across different dive phases and cardiovascular reflex indices, such as heart rate drop and the minimum heart rate, is calculated. Statistical analysis was performed to comprehend the differences across phases and changes in cardiovascular reflex indices between diving experience groups. Results The statistical analysis shows noteworthy differences in heart rate across dive phases. The notable variations were between the pre-dive and bottom phases (p<0.05) and between the descent and bottom phases(p<0.05). The study highlights the persistence of bradycardia even when depth remained constant, and a difference of 7.71 percentage points in the percentage drop in heart rate between experienced and novice divers. Discussion Experienced divers showed a pattern of lower heart-rate responses than novice divers across selected phases of the dive. The findings point to the value of further work on diver health monitoring, training, and cardiovascular adaptation in larger samples. Experienced divers should be regularly screened for cardiovascular disease to avoid any adverse events.

The ocean environment is a highly complex acoustic space where Sound Navigation and Ranging (SONAR) systems play a crucial role in detecting, classifying, and interpreting underwater signals for defense, geological exploration, and environmental monitoring. In recent years, over 70% of underwater monitoring data has been identified as acoustically rich but noisy, with more than 60% of manually analyzed samples showing inconsistency due to signal overlap and human bias. The need for automated and accurate sound classification and regression modeling arises in applications such as submarine detection, marine geological structure mapping, and industrial underwater noise assessment, where realtime and precise sound source identification is essential. Traditional manual classification methods suffer from high subjectivity, delayed analysis time, and inefficiency in handling large multivariate datasets with overlapping frequency domains. To overcome these limitations, this study introduces a Hybrid Hydro-Acoustics framework that combines regression and classification through an ensemble learning approach. The framework first preprocesses multivariate SONAR data—such as frequency, amplitude, and power spectral density—and then utilizes the ensemble mechanism to integrate decision boundaries and regression estimates from multiple learners. The existing algorithms, including Support Vector Classifier (SVC), Support Vector Regression (SVR) models and Gradient boosting (GB) CART as baseline learners for both regression and classification tasks, while enhancing prediction robustness using an Ensemble Extra Decision Tree (Ensemble EDT) strategy. The proposed Ensemble EDT for Classification effectively captures nonlinear separations in acoustic features, while Ensemble EDT for Regression provides improved prediction accuracy and stability for continuous parameters like sound intensity and frequency response. This hybrid ensemble framework demonstrates superior adaptability and generalization, enabling efficient modeling of complex underwater sound environments with higher accuracy, reduced overfitting, and enhanced computational efficiency.

Detecting earthquakes in seismological time series is a core task in observational seismology, supporting a range of applications from early warning systems to tectonic research. Typically, seismic sensors passively record data and send it to the cloud or edge for integration, storage, and analysis. While this cloud-based approach is effective in urban or well-connected areas, it is impractical in remote, underwater, or underground environments where network infrastructure is unreliable. In such settings, the sensors must operate independently for extended periods while coping with strict constraints on power, memory, and connectivity. To address these challenges, we present LightEQ, a system that combines an efficient data processing pipeline and a lightweight deep-learning model specifically designed for seismic event detection in such environments. LightEQ runs on ultra-low-power microcontrollers with just 100 kB of RAM, enabling real-time, on-device earthquake detection without the need for continuous streaming of raw data to a central location. We evaluate LightEQ against a traditional STA/LTA approach and state-of-the-art (SOTA) machine learning models, using the Stanford Earthquake Dataset. Unlike existing neural network (NN) models, which are too large for microcontrollers, LightEQ is over ten times smaller than most of the SOTA models. Our results demonstrate that communication is the most energy-intensive task in this setting, and that traditional model-driven filters like STA/LTA are inefficient due to their high false positive rate. In contrast, LightEQ improves detection accuracy with NN, providing a more energy-efficient solution by reducing the number of false positives before transmission. Compared to the STA/LTA method alone, LightEQ extends battery life by at least 3-fold by minimizing energy consumption associated with transmitting false positives to the cloud.

Accurate perception of hydrodynamic information is crucial for intelligent navigation and control of underwater robotics in challenging underwater environments. Current diaphragm-based differential pressure sensors are generally constrained by limited resolution for hydrodynamic perception. Here, we present a high-sensitivity calorimetric differential pressure sensor featured with precisely designed calorimetric components located on cantilever beams. The proposed sensor achieves an impressive underwater differential pressure resolution of 18.9 mPa and a repeatability standard deviation of 0.38%. By integrating a sensor array consisting of three such sensors into an underwater robotic model, the velocity and yaw angle were estimated simultaneously with average solution errors of 2.9 mm·s-1 and 0.94°, respectively. Underwater obstacles can be recognized with an accuracy of 97.5% by perceiving hydrodynamic variations in the Kármán vortex street due to its high sensitivity. Overall, the proposed sensor shows many potential applications in underwater flow sensing and the control of underwater robotics.

The rapid evolution of underwater communication technologies has become essential for applications such as marine exploration, environmental monitoring, and defense systems. Historically, underwater communication has relied primarily on acoustic methods due to their ability to travel long distances through water. However, with the increasing demand for high-speed and reliable data transmission, traditional systems face significant challenges. Acoustic communication, while effective over long ranges, suffers from low data rates, high latency, and susceptibility to noise and signal distortion. Similarly, radio frequency (RF) communication is highly inefficient underwater due to severe attenuation. These limitations highlight the need for an advanced communication approach capable of providing faster and more efficient data transfer in underwater environments. In this context, Light Fidelity (LiFi) technology emerges as a promising solution by utilizing visible light for data transmission. The proposed system integrates LiFi for high-speed, short-range underwater data communication and acoustic communication for long-range transmission, thereby combining the advantages of both technologies. The system employs LED-based transmitters and photodetector receivers for optical communication, along with acoustic modules for extended reach. This hybrid approach enhances data transmission efficiency, reduces latency, and improves reliability in dynamic underwater conditions. The significance of this work lies in its ability to overcome the inherent limitations of conventional systems, offering a scalable and efficient solution for nextgeneration underwater communication devices, with potential applications in oceanographic research, underwater robotics, and naval operations.

Replicating the skin's ability to sense touch, feel pain, and heal itself is key to developing the next generation of durable soft electronics. These capabilities become more critical in underwater environments, where divers and underwater machines face severe challenges such as limited dexterity, device damage, and restricted power availability. Here, we develop a self-healing magnetoelectric sensory system (SMES) that uniquely integrates self-powered tactile and proximity sensing with damage detection and autonomous recovery for amphibious operation. The SMES features a multilayer architecture composed of a damage-sensing layer and an underlying magnetoelectric sensing layer, both utilizing a self-healing elastomer with patterned liquid-metal conductors. The design enables the system to detect and recover from pricking, puncturing, and cutting damage while maintaining stable functionality. The SMES exhibits good sensitivity, rapid response, and robust durability in both air and water. Demonstrations with a smart diving glove and a soft robotic hand highlight its potential for noncontact communication and mechanoreception with damage feedback, paving the way toward next-generation amphibious soft machines that can feel and heal like living skin.

Rapid underwater repair using durable materials capable of withstanding extreme cold and saline marine conditions remains a significant unresolved challenge with broad implications for the maintenance and restoration of marine infrastructure. In this work, we developed a new class of functional materials, aluminosilicate-epoxy composites with self-initiated, frontal curing and seawater resistance, to address this challenge. A catalyzed stoichiometric frontal polymerization strategy was employed to overcome the fundamental chemical incompatibility between the frontal curing of epoxy and the alkali-activated geopolymerization of aluminosilicate. Distinct from traditional external photo- or thermal stimuli, calcium oxide hydration was used to trigger frontal curing in seawater, thereby eliminating external energy input during both initiation and curing. The underwater front-cured aluminosilicate-epoxy composite exhibits a compressive strength of 48.3 MPa and an adhesive strength of 5.57 MPa on steel within one hour. The key property indicator (KPI), defined as the ratio of minimal operational compressive strength to the required curing time, outperforms state-of-the-art performance by 87.4-fold. Moreover, the composites exhibit high ultraviolet resistance (92% strength retention) and seawater resistance (96.25% strength retention with 0.26% mass loss) after 720 hours of UV exposure and 7 days of seawater immersion. The frontal curing of seawater-resistant hybrid materials represents a paradigm shift in underwater repair, with substantial potential to transform marine infrastructure restoration.

Autonomous obstacle avoidance is a critical capability for Autonomous Underwater Vehicles (AUVs) to operate safely in dynamic and uncertain marine environments. Traditional AUV control methods rely on precise physical modeling and preset rules, yet they struggle to adapt to multiple sources of uncertainty, such as random initial states, dynamic obstacles, and varying currents. In recent years, deep reinforcement learning has provided a new avenue for data-driven adaptive policy learning. However, it remains insufficient for handling long-horizon tasks with sparse rewards. While hierarchical reinforcement learning can mitigate reward sparsity through temporal abstraction, it often faces challenges including exploration–exploitation imbalance, slow global convergence, and insufficient safety guarantees. Furthermore, most existing studies neglect dynamic environmental disturbances and task continuity, which further limits the practical application of these algorithms. To address these challenges, this paper proposes a hierarchical curiosity-driven AUV obstacle avoidance algorithm (HDAO), designed for autonomous obstacle avoidance in dynamic and uncertain underwater environments. The core design of HDAO incorporates several key innovations. Firstly, it introduces a Collision Threat Index for dynamic obstacles, which enables explicit risk perception and quantifies collision threats, thereby enhancing the policy’s generalization and robustness. Secondly, a task-decoupled hierarchical architecture is employed to synergistically optimize global path planning and local obstacle avoidance behaviors. This approach effectively manages long-horizon navigation tasks while alleviating high-dimensional training pressure. Finally, a novel reward mechanism is designed by integrating hierarchical active exploration with curiosity-driven passive exploration. This mechanism effectively incentivizes the agent to explore unvisited areas under sparse reward conditions and dynamically balances exploration and exploitation. Experimental results demonstrate that HDAO significantly outperforms existing methods in terms of obstacle avoidance success rate, training convergence speed and robustness against external disturbances.

Mechanical scanning sonar (MSS) plays an important role in high-precision object recognition and detection in underwater environments. However, existing research on MSS has focused on large objects, such as subsea structures and sunken ships, and often relies on unreleased datasets collected in confined water tank environments, limiting the study of small underwater objects and their application to real marine environments. Therefore, in this study, a Small Underwater Objects 3D Point Cloud (SUOP) Dataset was constructed using an MSS (BV5000) in the actual underwater environments at the seafloor. The dataset contains over 1,500 high-quality 3D point clouds for five objects, corresponding initial sonar scan data files, sonar system metadata, and 2D sonar images. The practicality of the proposed dataset was verified by applying it to an object recognition model. The results demonstrate that the SUOP dataset, with its object types, materials, and scanning conditions, enables accurate and robust evaluation of underwater object detection models, hence proving to be a valuable resource for research on marine underwater object detection.

Underwater wireless sensor networks (UWSNs) are essential to the work of the navy, as they are used to monitor objects (surveillance), to monitor the environment (environmental monitoring), and to defend the tactics (tactical defense). They are however challenged by serious issues in deployment because of limitation of underwater communication by acoustic means such as high latency, low bandwidth, high rate of packet loss and extreme energy limitation. The conventional centralized approach of data processing cannot survive under these circumstances and a shift towards the decentralized intelligence is needed. This paper will present an Energy-Aware Clustered Federated Learning (CFL) framework, which is specific to UWSNs in naval systems. The approach suggested will arrange sensor nodes into logical cluster, where local models are being trained and aggregated at cluster heads and sent to a central unit. In order to extend the network lifetime, an energy-conscious participation scheme is used to make sure that only nodes that are energy-reliant participate in model training. Moreover, we propose a powerful median-based aggregation approach at the cluster level in order to overcome the impacts of underwater communications that are noisy and lossy. Simulations conducted under realistic conditions in the underwater environment prove that the proposed CFL architecture is much more accurate in models, can boost communication overhead, and increase the energy efficiency of the system as opposed to conventional federated learning tools. It is also demonstrated that the results are much more robust to packet loss and communication failures, confirming the relevance of the framework in autonomous underwater operations. This paper points out the potential transformations that federated learning can bring to allow the development of intelligent, resilient, and energy-efficient, underwater sensor networks, and create new opportunities in the future in the fields of naval and maritime applications in challenging underwater conditions. Keywords: Underwater Wireless Sensor Networks (UWSNs); Federated Learning; Energy Aware Systems; Clustered Aggregation; Naval Applications.

With the continuous advancement of marine development, underwater operational tasks are becoming increasingly diverse and complex. Addressing the limitations of traditional methods and intelligent planning—which focus solely on acquiring task skills while separating grasp planning from force planning—this paper proposes a modeling approach integrating impedance control with deep reinforcement learning. Using a five-finger humanoid underwater dexterous hand as the grasping execution platform, this method achieves collaborative decision-making between grasp planning and force control for underwater dexterous hands. First, a modular underwater dexterous grasping scenario is established. Its kinematic model and inverse solution are analyzed, and the grasping problem is modeled as a Markov decision process. Second, based on the dexterous fingertip impedance control model for simulation, a grasping strategy learning method grounded in deep reinforcement learning is constructed to address the complex control challenges posed by the high degrees of freedom of the dexterous manipulator. Finally, the Proximal Policy Optimization (PPO) algorithm is employed for grasping strategy learning. An underwater dexterous grasping parallel training and testing environment is established using the Isaac Lab simulation platform to rapidly validate the learning method. Simulation results demonstrate the proposed method’s excellent dexterous compliant control performance and strong robustness to underwater variable environments: the PPO-based impedance control scheme reduces contact force variance by 56% compared to pure position control. The average maximum contact force is suppressed to 3.26 N, representing a 60.4% reduction compared to pure position control. This study achieves the organic integration of underwater hydrodynamic compensation, adaptive impedance control, and grasping strategy learning, providing a novel and effective solution for compliant grasping control of underwater dexterous manipulators.

Adaptations of seal louse nits to underwater life: morphology, respiration and attachment.

The suppression of direct forward scattering is challenging for object detection in the underwater environment under the incidence of quasi-planar waves. In this work, we study the lateral spatial phase of forward acoustic scattering in a vortex beam from a sphere. By subtracting a stable incident contribution, the forward scattering is extracted from the total field with good agreement between our experimental measurements and numerical simulations. A cross correlation method of the spatial phase is proposed, and results show a linear dependence of the phase singularity shift versus the sphere's offset to the beam axis, which may be potential for the underwater target detection and positioning by using the spatial phase information as the degree of freedom. © 2026 Acoustical Society of America.

Autonomous navigation in underwater environments is challenged by the absence of GPS, degraded visibility, and submerged obstacles. This article investigates these issues using the BlueROV2, an open platform for scientific experimentation. We propose a deep reinforcement learning approach based on the Proximal Policy Optimization (PPO) algorithm, using an observation space that combines target-oriented navigation information, a virtual occupancy grid, and raycasting along the boundaries of the operational area. This information is encoded into a high-dimensional observation space of 84 dimensions, providing the agent with comprehensive local and global situational awareness. The learned policy is compared against a reference deterministic kinematic planner, the Dynamic Window Approach (DWA), a robust baseline for obstacle avoidance. The evaluation is conducted in a realistic simulation environment and complemented by validation on a physical BlueROV2 supervised by a 3D digital twin of the test site, reducing risks associated with real-world experimentation. The results show that the PPO policy consistently outperforms DWA in highly cluttered environments, notably thanks to better local adaptation and reduced collisions. Finally, experiments demonstrate the transferability of the learned behavior from simulation to the real world, confirming the relevance of deep RL for autonomous navigation in underwater robotics.

Underwater structured-light three-dimensional (3D) imaging is essential for the precise reconstruction of submerged objects in scientific and engineering applications. However, the observed fringe patterns are often degraded by underwater light attenuation, scattering, and refractive distortion, leading to fringe blurring and aliasing and, consequently, reduced phase-reconstruction accuracy. To address these issues, we analyze the propagation characteristics of structured light in underwater environments and their effects on fringe image quality. With that, we engineer a diffusion-model-based fringe restoration neural framework, referred to as FP-DiffNet, in which fringe restoration is formulated as a probability-driven iterative denoising process. Specifically, a U-Net model is trained to progressively map degraded fringe observations to clear fringe patterns, incorporating physics-guided constraints and an adaptive noise-annealing mechanism to effectively separate scattering-induced noise while preserving fringe structures during the reverse diffusion process. The proposed framework is evaluated in terms of fringe image quality, wrapped-phase accuracy, and 3D reconstruction performance. Under extremely high turbidity, it achieves a PSNR of 42.42dB and a wrapped-phase MAE of 0.0354rad. By combining diffusion-driven phase reconstruction with dynamic compensation, sub-millimeter absolute measurement is achieved in turbid water, with a 3D reconstruction RMSE below 0.1mm. Importantly, no prior modeling of water optical parameters is required, enabling robust and practical underwater 3D imaging.

Three-dimensional tracking that relies solely on one-dimensional bearing measurements from multiple horizontal linear arrays is particularly challenging in underwater environments. Depth-dependent variations in sound speed cause acoustic rays to bend, leading to deviations in the incident bearing angles. Moreover, the relatively low sound speed in water makes the propagation delay non-negligible, a bearing angle measured at the current time instant typically corresponds to an earlier target state. To address these issues, we formulate a bearing measurement model under an isogradient sound speed profile that jointly accounts for ray bending and propagation delay. Bearing angles measured by multiple arrays at a common time correspond to target states at different emission times, which are coupled with unknown propagation delays through an implicit constraint. To handle this coupling, we propose a centralized particle filter with an iterative procedure to jointly estimate the propagation delay and the target state at the emission time for each array. The resulting emission-time state estimates are then marginalized to obtain the posterior estimate of the target state at the common measurement time. Finally, we derive the posterior Cramér-Rao lower bound to provide a performance benchmark. Simulation studies and sea-trial data demonstrate the effectiveness of the proposed approach.

Marine debris poses an alarming threat to ocean environments. Conventional methods of sea and ocean cleaning rely heavily on manual collection, a process that has repeatedly demonstrated its inefficiency and extensive demand for resources. This paper presents the SeaClear system, a novel multi-robot platform designed to autonomously detect and collect marine debris, thereby offering a more efficient solution to this environmental challenge. An overview of the system is presented, followed by a detailed description of each robot’s capabilities. Leveraging artificial intelligence, the system employs the deep-learning-based computer vision algorithm You Only Look Once (YOLO) for the detection of underwater litter, addressing the challenges of poor visibility and hydrodynamic disturbances of underwater environments. Additionally, the paper explores the implemented navigation and control methodologies, which are an essential part of the workflow of the system. The performance of the designed system is validated via field tests conducted in a real-world underwater environment. Finally, directions for future work are proposed. • The infrastructure of a multi-robot system for autonomous underwater debris detection, mapping, and collection is presented. • The suitability of YOLO-based deep learning for real-time debris detection in shallow waters is validated. • The sensing and control scheme that enables the operation of the multi-robotic platform is presented. • The practical performance of the system is confirmed with field experiments to validate the feasibility of AI-driven underwater operations.

ABSTRACT Diving is critical for maritime operations but poses severe physiological risks to divers in cold, high‐pressure underwater environments. Continuous monitoring of diver's vital signs, particularly respiration, is essential for safety but remains challenging due to seawater interference and the limitations of existing devices. Here, we present a multifunctional self‐powered monitoring system (MSM system) based on flexible triboelectric sensors that achieves real‐time measurement of a diver's respiratory rate and breathing depth underwater, while simultaneously tracking body motion. The system integrates lightweight wearable sensors, a portable signal‐processing module and a deep‐learning analysis terminal. By analyzing multichannel respiratory and motion data, the MSM system is capable of providing a comprehensive assessment of the diver's physiological status. During the training process, the provision of real‐time feedback facilitates the immediate adjustment of breathing and movement patterns, thereby enhancing both the safety of the training and its efficacy. Furthermore, in the event of danger during underwater operations, the system provides immediate warning, thus enhancing the safety of the diver. Underwater experiments demonstrate the system's robust performance in capturing respiratory patterns and motion. This MSM system offers a novel and practical approach to diver safety monitoring and personalized training.