The oceans, vast and enigmatic, cover more than 70% of our planet’s surface yet remain among the least charted frontiers due to their complexity and inaccessibility. Traditional methods of underwater exploration, often reliant on manned expeditions or singular automated devices, face limitations in scale and risk. By deploying a swarm of AUVs, each equipped with cutting-edge sensors and communication protocols, researchers can undertake massive parallel missions that map seascapes, monitor wildlife, and gather critical environmental data. However, coordinating such numerous, independent robots in a cooperative manner introduces immense challenges in autonomy, navigation, and data integration.

This is where the digital twin framework intervenes as a transformative solution. In essence, every physical AUV in the swarm has a corresponding digital double operating within a high-fidelity simulation environment. These digital counterparts synthesize real-time data inputs, including positional coordinates, sensor readings, hydrodynamic conditions, and system health metrics, to construct a coherent and dynamic virtual model of the swarm’s collective behavior. This continual feedback loop enables adaptive mission planning and rapid response to unforeseen circumstances, such as shifting currents or mechanical malfunctions.

The power of a digital twin-driven swarm is its capacity for emergent coordination without centralized control. By leveraging machine learning algorithms housed within the virtual arena, individual AUVs negotiate movement patterns, task allocations, and collision avoidance strategies independently yet harmoniously. This decentralized intelligence allows swarms to scale effectively, deploying hundreds or even thousands of units, all while maintaining operational integrity and mission coherence. The concept draws inspiration from biological collectives such as fish schools or bird flocks, where local interactions yield complex global dynamics.

Technically, implementing this system required breakthroughs in communication and computational architecture. Underwater communication notoriously suffers from bandwidth constraints and latency issues. To overcome this, the research introduced an optimized acoustic communication protocol coupled with intermittent surface relays for data synchronization. High-performance edge computing modules embedded within each AUV process raw data locally, diminishing the load on central servers and ensuring rapid decision-making even in communication sparse regions.

The researchers also applied advanced hydrodynamic modeling to enhance the accuracy of the digital twins. Understanding fluid dynamics is critical for predicting vehicle trajectories and energy consumption in diverse underwater currents and turbulence. The virtual models continuously assimilate sensor feedback to refine these simulations, leading to more realistic and reliable predictions. As a result, energy expenditure is minimized, extending operational endurance and allowing longer, more complex missions.

One of the most remarkable achievements of this research is the swarm’s robust fault tolerance. In laboratory and field trials, individual AUV failures, whether mechanical or software-driven, did not compromise the mission. The digital twin network identifies malfunctioning units, recalibrates swarm configurations accordingly, and reallocates tasks among remaining vehicles. This resilience is vital for long-duration expeditions in harsh environments, where maintenance opportunities are scarce.

From an applications perspective, the digital twin-driven swarm paves the way for transformative advances in marine science and industry. It enables high-resolution seafloor mapping crucial for understanding geological processes and locating underwater resources such as rare minerals or archaeological artifacts. Environmental monitoring benefits immensely by detecting pollution plumes, assessing coral reef health, and tracking migratory marine species on scales unachievable by current methods.

Furthermore, the autonomous nature of the swarm significantly reduces human risk and operational costs. Deep-sea expeditions, traditionally expensive and time-consuming, can now be conducted continuously and remotely with automated oversight. This democratization of ocean exploration unlocks opportunities not only for large research institutions but also smaller entities and developing nations seeking to broaden their marine knowledge.

However, the study also acknowledges remaining challenges. The complexity of digital twin synchronization across vast spatial scales requires further refinement to handle extreme environmental variability and ensure fail-safe autonomy. Ethical considerations around autonomous systems operating in sensitive marine zones are emphasized, prompting calls for comprehensive governance frameworks balancing innovation with conservation imperatives.

Looking ahead, the intersection of digital twins and swarm autonomy raises exciting prospects beyond oceanography. Similar principles could be adapted for terrestrial robotics, atmospheric monitoring, and even extraterrestrial exploration, where distributed systems operate in hostile or inaccessible domains. The modularity of the digital twin architecture allows rapid customization and scaling for diverse tasks, signaling a paradigm shift across multiple technological sectors.

In conclusion, the digital twin-driven swarm of autonomous underwater vehicles represents a monumental leap forward in marine exploration capabilities. By harnessing the synergy of virtual-real integration, distributed intelligence, and adaptive control, this platform unveils a new epoch where the secrets of the deep sea can be unraveled comprehensively, safely, and sustainably. The visionary work by Yan, Zhang, Guan, and their team not only pushes the boundaries of engineering but also enriches humanity’s quest to understand and protect our planet’s blue heart.

As this technology continues to mature, its societal implications will be profound. Enhanced marine data will inform climate models, fisheries management, and disaster response strategies, crucial for addressing global challenges such as biodiversity loss and ocean acidification. The fusion of digital twin technology and robotic swarms encapsulates how interdisciplinary innovation can transform exploratory science from an arduous endeavor into a seamless, intelligent operation, inspiring a new generation of researchers and explorers to dive deeper than ever before.

Subject of Research: Digital twin integration with autonomous underwater vehicle swarms for enhanced marine exploration.

Article Title: Digital twin-driven swarm of autonomous underwater vehicles for marine exploration.

Article References:
Yan, J., Zhang, T., Guan, X. et al. Digital twin-driven swarm of autonomous underwater vehicles for marine exploration. Commun Eng (2026). https://doi.org/10.1038/s44172-025-00571-7

Image Credits: AI Generated

Traditional underwater vehicles typically rely on fully actuated mechanisms, which allow them to control movements along all degrees of freedom independently. However, underactuated AUVs, which lack such comprehensive control authority, are often preferred due to their lower energy consumption, reduced mechanical complexity, and enhanced robustness in challenging aquatic environments. The downside of such designs lies in the intricate control problems they present, as maintaining a desired depth with a restricted control surface or actuation degree requires sophisticated guidance and estimation techniques. The research by Qu and colleagues pushes the envelope by introducing a cascaded control architecture that integrates guidance, state estimation, and control layers, thereby allowing the vehicle to adapt dynamically to changes in underwater conditions while steadfastly maintaining depth.

At the core of this approach is a hierarchical system where the guidance module sets the depth trajectory based on mission parameters and environmental inputs. This trajectory serves as a reference for the subsequent state estimation layer, which employs advanced algorithms to ascertain the vehicle’s current depth and velocity, accounting for sensor inaccuracies and environmental disturbances. Crucially, these estimates inform the control layer, which manipulates the vehicle’s limited actuators to correct deviations from the target depth. By cascading these modules, the system isolates individual task complexities while maintaining an integrated workflow, resulting in more reliable and efficient depth tracking performance.

The state estimation technique used in this framework leverages probabilistic filtering strategies that are robust to noise and uncertainties inherent in underwater sensing. Given that sensor readings underwater are prone to various disturbances—from salinity gradients to pressure variations—accurately determining the vehicle’s depth and vertical velocity demands sophisticated filtering techniques. The study effectively utilizes nonlinear state observers, which adaptively refine depth estimates by incorporating real-time data and known vehicle dynamics, leading to substantial improvements over classical estimation methods.

Complementing the state estimation, the guidance component adapts dynamically by monitoring sea conditions and mission objectives. Traditional guidance strategies often assume steady-state conditions or predictable underwater currents, which hardly reflect the stochastic and turbulent nature of real oceanic environments. By incorporating adaptive path planning mechanisms, the proposed cascaded guidance system modifies the depth trajectory in real time, optimizing the vehicle’s route for energy efficiency and mission success probability. This flexibility is particularly valuable for long-duration missions or complex operational scenarios like deep-sea exploration and subsea inspections where environmental parameters can shift unexpectedly.

The control scheme itself represents a significant leap forward in handling underactuated dynamics. Underactuation implies that control inputs cannot directly modulate every movement direction, frequently resulting in system nonlinearities and constrained maneuverability. To tackle this, the research team developed a nonlinear controller based on feedback linearization principles combined with adaptive gain tuning. This advanced controller not only compensates for nonlinear system behavior but also adjusts in response to changing hydrodynamics and payload variations, ensuring robust depth tracking without requiring extensive manual retuning.

Crucially, the study validates its approach through extensive numerical simulations and experimental trials using an underactuated AUV prototype. These rigorous performance tests demonstrate the system’s ability to maintain tight depth tracking accuracy even under heavy disturbances such as turbulent currents and sensor noise. The experiments also highlight the controller’s rapid convergence speed and minimal overshoot, both critical factors in practical underwater missions to avoid collisions with seabed structures or marine life. The selective filtering strategies in conjunction with adaptive control allowed the vehicle to cruise at different operational depths with unprecedented stability and energy efficiency.

The implications of this research extend well beyond academic interest, promising transformative impacts on commercial and scientific domains. Autonomous vessels equipped with this cascaded architecture will not only operate more safely but will also reduce operational costs thanks to optimized energy usage and enhanced mission reliability. For environmental monitoring applications, where long-term deployment without human intervention is essential, stable depth control can significantly improve data accuracy and sensor longevity. Moreover, subsea infrastructure maintenance, including pipelines and telecommunication cables, will benefit from precise depth control to conduct inspections more effectively and avoid costly operational failures.

One of the particularly innovative aspects of this work is its modular architecture, rendering the solution adaptable and scalable. The cascaded framework can accommodate future enhancements such as the integration of machine learning-based predictive models to further improve guidance or the pairing with advanced sensor suites for enhanced environmental awareness. Additionally, the modular design fosters easier troubleshooting, maintenance, and upgrades, critical for field deployments that demand reliability over extended periods in remote underwater locations.

The multidisciplinary nature of this research integrates principles from control theory, ocean engineering, robotics, and artificial intelligence. Such convergence is essential when addressing the multidimensional challenges posed by underwater navigation in complex water columns. By bridging theoretical developments with practical applications, the study sets a new standard in AUV systems engineering and offers a blueprint for future innovations in autonomous maritime technologies.

In conclusion, the cascaded guidance, state estimation, and control strategy developed by Qu and colleagues presents a pivotal advancement in the domain of underactuated autonomous underwater vehicles. Through its intelligent and adaptive management of underactuated dynamics, it signifies a leap toward more capable, reliable, and efficient AUVs that can navigate the depths with unprecedented precision. This advancement not only enhances scientific exploration but also strengthens industrial underwater operations central to modern maritime economy and ecological stewardship.

This work exemplifies how combining layered control architectures with robust estimation and adaptive guidance unlocks new performance dimensions for underactuated robotic platforms. As AUVs take on increasingly sophisticated and vital tasks beneath the oceans’ surface, innovations like this will define the next era of underwater robotic autonomy—potentially reshaping our ability to understand and protect vast and largely inaccessible marine environments.

The comprehensive testing outcomes and theoretical analyses reported in the study reflect meticulous attention to real-world challenges and an ambitious vision for future autonomous underwater operations. Continued refinements and practical deployments of this cascaded control framework are expected to propel the autonomous underwater vehicle field into new realms of efficiency and reliability, contributing meaningfully to ocean science, defense, energy, and environmental efforts worldwide.

—

Subject of Research: Autonomous underwater vehicles (AUVs) depth tracking using cascaded guidance, state estimation, and control.

Article Title: Cascaded guidance, state estimation, and control for depth tracking of underactuated autonomous underwater vehicles.

Article References:
Qu, Y., Zhang, Q., Xiang, X. et al. Cascaded guidance, state estimation, and control for depth tracking of underactuated autonomous underwater vehicles. Commun Eng 4, 191 (2025). https://doi.org/10.1038/s44172-025-00521-3

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s44172-025-00521-3

This robotic fish represents a convergence of advanced mechatronics, control systems, and underwater communication technologies, resulting in an autonomous device uniquely suited to marine ecosystem inspection and maintenance. Its biomimetic design incorporates flexible fins that mimic natural fish locomotion, allowing for efficient and agile navigation under water. By deploying cutting-edge actuators that replicate fin movements, the robot achieves propulsion with minimal noise and disturbance to aquatic life, a significant advantage over traditional underwater vehicles relying on propellers.

Central to the robot fish’s functionality is its umbilical communication system, facilitating reliable data transmission between the submerged platform and surface control units. This system supports real-time wireless communication, a capability critically tested in recent trials alongside a surface robot at the experimental facilities of Port Castelló. These wireless exchanges enable synchronized operations and remote command execution, essential for intricate tasks such as inspection of sensitive environments or deployment of scientific sensors.

Equipped with an auxiliary sonar system, the robot fish navigates complex underwater terrains and performs detailed mapping of submerged structures. This sonar capability enhances situational awareness, allowing the platform to detect obstacles and features within its operational radius. Coupled with a specialized visual inspection subsystem designed explicitly for the internal examination of fish farm nets, the robot offers unprecedented insights into aquaculture infrastructure health, addressing a crucial need for sustainable aquafarming practices.

The integration of sensor deployment and retrieval mechanisms within the robot opens new frontiers for marine monitoring. By autonomously positioning environmental sensors and subsequently collecting them, the robotic system facilitates long-term data acquisition without extensive human intervention, increasing operational safety and efficiency while reducing costs. This aspect is vital in challenging aquatic environments where traditional sensor maintenance is logistically complex and hazardous.

Professor Raúl Marín, a leading researcher at CIRTESU, emphasizes the importance of incrementally rigorous testing regimes. The developmental pathway began with controlled university lab experiments and progressively extended to real-world marine settings like Port Castelló. This methodological approach ensures that each facet of the technology is validated under increasingly realistic conditions, refining system robustness and performance. Collaborative efforts with Port Authority personnel have been instrumental in this process, enabling access to diverse operational scenarios that simulate actual deployment challenges.

Environmental sustainability and animal welfare in aquaculture are central motivators behind this research. By providing a sustainable, non-invasive platform for inspection and maintenance, the robotic fish minimizes human disturbances to underwater habitats while ensuring the structural integrity of fish farming nets. These advances translate directly into improved animal safety and welfare, as timely detection of net integrity issues prevents escapes and protects farmed species from predation or disease transmission.

Looking ahead, CIRTESU’s strategic roadmap involves enhancing the robot fish’s capabilities to autonomously perform net repairs, a complex task that demands precise manipulation and sophisticated control algorithms. Developing robotic interventions to conduct maintenance operations underwater not only promises to revolutionize aquaculture logistics but also positions the technology as a scalable solution for broader marine infrastructure management.

The collaboration between Universitat Jaume I and the Port Authority of Castelló underscores a shared vision for the port as a living laboratory—a dynamic innovation hub facilitating advanced marine technology experimentation. Since formalizing their partnership in July 2024, the provision of port facilities as an isolated testbed environment has accelerated the transition of technologies from experimental prototypes towards operational readiness, elevating the Technology Readiness Level of CIRTESU’s projects.

Academic contributions underpinning this development include doctoral research led by Andrea Pino Jarque, complemented by supervision from María Rosario Vidal and Raúl Marín Prades, with overall coordination by Professor Pedro J. Sanz Valero. Additionally, the involvement of recent graduates such as Max Puig Sariñena, who contributed to underwater communication experiments, exemplifies the centre’s commitment to integrating educational initiatives with applied research.

The robot fish initiative stands as a testament to the transformative potential of interdisciplinary engineering, combining mechanics, electronics, software, and marine sciences to address real-world challenges. Its success resonates beyond national boundaries, offering a model for future autonomous underwater systems designed for environmental monitoring, infrastructure surveillance, and sustainable aquaculture.

This award-winning project not only elevates the profile of CIRTESU and Universitat Jaume I within the global robotics community but also reflects broader trends toward environmentally conscious automation. As marine industries seek innovative tools to balance productivity with ecological stewardship, technologies like the robot fish are poised to lead a new era of marine exploration and maintenance, fostering a safer and more sustainable aquatic future.

Subject of Research: Advanced biomimetic underwater robotic fish for marine environment monitoring and aquaculture maintenance.

Article Title: Award-Winning Biomimetic Robot Fish Enhances Marine Automation and Sustainable Aquaculture Monitoring at Port Castelló.

News Publication Date: September 2024

Web References: Not provided

References: Not provided

Image Credits: Universitat Jaume I of Castellón

Keywords: underwater robotics, biomimetic design, marine automation, aquaculture monitoring, sensor deployment, wireless underwater communication, sonar inspection, sustainable aquaculture, autonomous marine vehicles, robot fish, Port of Castelló, CIRTESU, Universitat Jaume I

The team behind this groundbreaking synthesis, led by Dr. Weidong Zhang of Shanghai Jiao Tong University in partnership with scholars from Tongji University and Hainan University, presents a sweeping analysis of current technologies essential to heterogeneous aquatic systems. Their work delves into six core technical domains that underpin these robotic networks: communication, perception, navigation, control, decision-making, and energy management. Through a deep understanding of these interrelated areas, this review not only encapsulates the present state of the art but also offers insightful perspectives on future innovation trajectories.

Aquatic robotic networks inherently involve heterogeneity, blending underwater remotely operated vehicles (ROVs), autonomous underwater vehicles (AUVs), surface vessels, and aerial drones into orchestrated task forces. Each robotic type specializes in unique interaction modes and sensory capabilities, ranging from high-precision sonar imaging below the waves to aerial surveillance above. When combined, these vehicles can perform coordinated operations impossible for any single platform. This heterogeneous integration not only enhances operational flexibility but also significantly mitigates the limitations intrinsic to individual robots navigating complex and dynamic aquatic environments.

One of the most formidable challenges addressed in this survey is communication. Underwater communication remains notoriously problematic due to the physical properties of water, which strongly attenuates electromagnetic signals commonly used in terrestrial wireless networks. Acoustic communication emerges as a primary medium underwater, yet it struggles with limited bandwidth, signal delays, and multipath interference. To overcome these obstacles, current research has been exploring hybrid communication frameworks that integrate acoustic, optical, and radio-frequency methods tailored to varied operational depths and environmental conditions. These efforts are laying the groundwork for robust, high-fidelity data exchange across heterogeneous robot teams.

Perception systems are equally critical in enabling aquatic robots to navigate and interact with their surroundings effectively. Given the distortions and visual occlusions common in underwater environments, conventional cameras often fall short. Therefore, researchers have turned to advanced multi-modal sensing techniques—combining sonar, LIDAR, radar, and environmental sensors—to build richer environmental models. Machine learning, particularly deep learning frameworks, is increasingly leveraged to fuse sensor data in real-time, facilitating enhanced object recognition, obstacle avoidance, and scene understanding. Such perceptual sophistication is crucial for autonomous decision-making and safe maneuvering in hazardous or cluttered maritime settings.

Navigation and control constitute core pillars in orchestrating the movement of these robotic units. Precise localization underwater remains a complex issue as GPS signals cannot penetrate water. Alternative approaches, such as inertial navigation systems (INS), Doppler velocity logs (DVL), and acoustic beacons, are combined to provide reliable positioning. Control algorithms must also account for the nonlinear dynamics of water currents, changing buoyancy, and external disturbances. Advanced adaptive control strategies and cooperative formation control enable these robots to maintain coordinated trajectories and perform group maneuvers. The fusion of robust navigation with adaptive control enhances the overall reliability and effectiveness of the system in mission-critical scenarios.

Decision-making within heterogeneous robotic teams integrates data gathered from diverse sensors and communication channels, presenting unique computational challenges. This review outlines recent advances in distributed decision frameworks that can manage uncertainty, incomplete information, and operational constraints across multiple vehicles. Reinforcement learning approaches enable robots to adapt to changing mission goals and environmental factors autonomously. Furthermore, multi-agent systems research contributes techniques for negotiation and task allocation among heterogeneous platforms, optimizing cooperative behavior without human supervision. Intelligent decision-making is thereby becoming an indispensable capability for maximizing mission success in complex aquatic tasks.

Energy management in aquatic robots is a persistent bottleneck impeding mission duration and operational range. Limited onboard energy capacity, coupled with the high power demands of communication, sensing, locomotion, and computation, necessitates innovative solutions. The review highlights state-of-the-art low-power hardware designs and energy-efficient algorithms tailored for long-term deployment. Promising advancements in energy harvesting technologies—such as wave energy converters and solar panels for surface vessels and aerial drones—offer pathways toward self-sustaining systems. Additionally, coordinated energy management strategies enable heterogeneous teams to share workloads efficiently, extending the collective operational lifespan of the robotic network.

Beyond these technical domains, this survey draws attention to compelling real-world applications demonstrating the potential impact of heterogeneous aquatic robotic systems. Oceanographic monitoring benefits from multi-modal data collection enabled by coordinated vehicles, providing unprecedented insights into marine ecosystems and climate change indicators. In underwater archaeology, combinations of surface drones and AUVs facilitate detailed site mapping, artifact detection, and controlled excavation with minimal human intervention. Emergency response operations also leverage rapid deployment and collaboration of various robotic types to navigate hazardous environments, locate victims, and provide situational awareness in disaster zones.

Despite these promising strides, significant obstacles remain to be addressed for fully mature heterogeneous aquatic robotic systems. The survey underscores issues such as signal degradation at different depths, sensor reliability in turbulent waters, the complexity of real-time multi-robot coordination, and limitations arising from hardware constraints in small unmanned vehicles. The authors emphasize that future research must continue integrating innovations across hardware design, software algorithms, and system architectures to surmount these challenges. Cross-disciplinary collaboration will be paramount to bridging gaps between theory and applied engineering in this dynamic field.

Importantly, the review advocates for the adoption of unified frameworks that can integrate diverse robotic agents into cohesive operable ecosystems. Standardization in communication protocols, data formats, and control methodologies will accelerate interoperability and scalability. The authors also encourage the infusion of artificial intelligence advancements—especially in deep reinforcement learning and multi-agent cooperation—to elevate autonomy and robustness. Such integrative approaches will empower robotic systems to handle the increasing complexity and scale of marine missions envisioned in the near future.

Moreover, the paper points toward exciting possibilities in expanding the role of heterogeneous aquatic robots for sustainable marine resource management. Intelligent, cooperative fleets can monitor fish populations, assess coral reef health, and track pollution levels with greater coverage and precision. These capabilities are vital for informed policy-making and conservation efforts. The potential for robotic systems to reduce human environmental footprints by undertaking risky or labor-intensive tasks is a transformative prospect that aligns with global sustainability goals.

For emerging researchers venturing into this multidisciplinary and technically demanding arena, this review offers a valuable roadmap. It consolidates a wide spectrum of research findings, technological trends, and open questions, facilitating a comprehensive understanding of heterogeneous aquatic robotics. By highlighting both successes and persistent limitations, the article equips scientists and engineers with a clear vision of paths forward and areas where innovation can make profound impacts. The collaboration between established institutions in China underscores the significance and global relevance of this research topic.

In summary, the study titled "Survey on heterogeneous aquatic robot systems: communication, perception, navigation, control, decision-making and energy management" charts the frontier of intelligent aquatic robotic systems with a balanced, systematic, and detail-oriented perspective. It signals an era where diverse robotic agents operate in fluid, synergistic modes, overcoming individual constraints through collective intelligence and adaptive coordination. This promises to revolutionize how humanity studies, protects, and utilizes the oceans, fostering new modes of interaction with one of Earth’s most critical environments.

Subject of Research:
Not applicable

Article Title:
Survey on heterogeneous aquatic robot systems: communication, perception, navigation, control, decision-making and energy management

News Publication Date:
30-May-2025

Web References:
10.55092/rl20250003

References:
Liu R, Hu X, Jiang Z, Wang J, Zhang W. Survey on heterogeneous aquatic robot systems: communication, perception, navigation, control, decision-making and energy management. Robot Learn. 2025(1):0003.

Image Credits:
Shanghai Jiao Tong University, Hainan University

Keywords:
Robot control