Assessment 1 – Research Proposal on Digital Twins in Autonomous Ship Design

Posted: September 29th, 2025

Assignment Brief: Assessment 1 – Research Proposal on Digital Twins in Autonomous Ship Design

Dear Students,

It’s always a pleasure to see you grappling with the cutting edge of our field, where the blend of simulation and real-world maritime demands really comes alive. This assessment builds on our discussions in the module about how digital tools are reshaping vessel autonomy, much like the hands-on simulations we’ve run in lab sessions. Imagine proposing a system that could predict a ship’s behaviour before it even leaves drydock – that’s the kind of forward-thinking work I hope you’ll bring here.

Module Details: Autonomous Marine Systems (AMS 4021), Level 4, 15 Credits, Semester 1, 2025-26 Academic Year. Assessment Type: Individual Research Proposal (contributes 40% to overall module mark). Word Count: 1,500 words (excluding references and appendices; 10% tolerance, with deductions beyond). Submission Deadline: Friday, 14 February 2026, 12:00 noon via Turnitin on Moodle. Late work penalised at 5% per working day, capped at 5 days; zero thereafter without approved extensions.

Assessment Overview: You’ll develop a research proposal for applying digital twin technology to the design of an autonomous surface vessel (ASV), drawing from case studies like those in North Sea monitoring or port logistics. This echoes industry-led projects from places like the University of Southampton or Strathclyde, emphasising simulation-driven validation and IMO guidelines on MASS (Maritime Autonomous Surface Ships). It’s a stepping stone to your group design later in the module.

Learning Outcomes Addressed:

1. Evaluate digital twin frameworks for hydrodynamic and control system modelling in naval architecture.
2. Formulate research questions and methodologies for autonomous vehicle integration.
3. Assess ethical, regulatory, and operational challenges in simulated maritime environments.
4. Present a feasible proposal aligned with UK Maritime Autonomous Systems Regulatory Framework.

Task Instructions:

1. Introduction and Rationale (approx. 400 words): Outline the role of digital twins in ASV design (e.g., real-time fault prediction or route optimisation). Identify a specific challenge, such as collision avoidance in congested waters, and justify its relevance for a 500 GT ASV. Reference IMO’s 2025 MASS Code interim guidelines.
2. Literature Review (approx. 500 words): Survey 4-6 sources from 2019-2025 via Google Scholar or IEEE Xplore. Critically appraise applications in ship stability or propulsion control, highlighting gaps like data integration in harsh conditions.
3. Methodology and Feasibility (approx. 400 words): Propose a mixed-methods approach, e.g., MATLAB/Simulink for twin development and CFD for validation. Include timelines, resources (e.g., open-source datasets), and risk mitigation. Discuss alignment with UN SDG 9 (Industry, Innovation and Infrastructure).
4. Expected Outcomes and Impact (approx. 200 words): Predict contributions to safer autonomy and suggest scalability for fleet operations.

Support and Resources:

• Fortnightly office hours (Wednesdays, 10am-12pm, via Zoom).
• Tools: Access to ANSYS Twin Builder trial and sample ASV models on Moodle.
• Reading pack: “Digital Twins for Maritime Autonomy” curated list. Feel free to bounce ideas off me – these proposals often evolve into standout final projects.

Marking Criteria (out of 100%):

Distinction (70%+) demands proactive insight; pass (40-69%) requires solid basics. Feedback within 20 working days on Moodle, including annotated proposal.

All the best with this – your proposals shape the module’s direction.

Warm regards, Professor Elena Hargrove Module Lead, Department of Naval Architecture & Ocean Engineering

Resources References (Harvard Format)

These draw from recent peer-reviewed works on digital twins and autonomous marine systems, sourced via Google Scholar and MDPI/IEEE databases, to underpin your proposal’s literature review.

1. 1. Negreiros, J., Santos, T.A. and Guedes Soares, C. (2023) ‘Digital Twins in the Marine Industry: A Review’, Electronics, 12(9), p. 2025.
   2. Raptis, T.P., Faiella, G., Giallanella, L., Kyriakopoulos, G.L. and Kameas, A. (2022) ‘Ship’s Digital Twin—A Review of Modelling Challenges and Applications’, Journal of Marine Science and Engineering, 10(8), p. 1136.
   3. Lee, S., Kim, N. and Han, S. (2024) ‘Review Digital twins enable shipbuilding: Applications, challenges, and future directions’, Alexandria Engineering Journal, 89, pp. 1-18.
   4. Wang, J., Zhang, H., Liang, Y. and Zhang, Y. (2025) ‘The application and development of digital twin in the marine domain’, Ocean Engineering, 295, p. 116947.

________________________________________________________________________________________________

Sample Research Proposal in Marine Engineering.

Research Proposal: Digital Twin Integration for Collision Avoidance in a 500 GT Autonomous Surface Vessel

Abstract

Digital twin technology promises predictive fidelity across design, control and operational envelopes. The proposal sets a narrow ambition: develop a digital twin that materially improves collision-avoidance decisions for a 500 gross-tonne autonomous surface vessel operating in congested coastal approaches. The work will pair physics-based hydrodynamic models with a control-layer twin and a probabilistic situational-awareness twin that ingests sensor uncertainty. The hypothesis holds that a multi-tier twin workflow will reduce near-miss events by improving pre-emptive control strategies and by enabling runtime scenario rehearsal. The proposal outlines literature gaps, a mixed-methods development strategy, validation plans using CFD and hardware-in-the-loop (HIL), resource needs, regulatory alignment and measurable success criteria.

Problem Statement and Rationale

Congested waters expose autonomous vessels to ambiguous encounters where sensor occlusion, wake interactions and close-quarter dynamics generate rapid, non-linear risk. Existing autonomy stacks rely on perception-to-planning loops that often collapse under degraded sensing or under-specified hydrodynamic coupling, producing evasive commands that either violate COLREGs or produce unstable motion in low-speed, high-interaction scenarios. A 500 GT ASV presents design constraints: sufficient displacement to carry payload while retaining manoeuvrability to execute avoidance within port approach channels. The proposed digital twin focuses on collision avoidance because it sits at the intersection of safety, regulatory compliance and commercial viability. A twin that simulates near-future vessel responses under plausible sensor and environment uncertainty will let designers and operators quantify margins before committing to hardware. The IMO 2025 interim MASS guidance provides a regulatory frame but leaves implementation details unresolved, specifically how validated simulation outputs should evidence safe operation. Addressing that gap will make the twin not only a design tool but a compliance artefact suitable for regulatory review. The project therefore treats the twin as an engineering proof instrument and a decision-support system for control law verification and validation.

Literature Review

Recent reviews document digital twin applications across shipbuilding and operations but reveal methodological fragmentation. Surveys identify conceptual frameworks and case studies where twins model propulsion and structural health, but few provide rigorous approaches linking hydrodynamic fidelity with runtime control validation (Raptis et al., 2022; Negreiros et al., 2023). Studies that couple CFD and control simulations often stop at component-level verification and do not address end-to-end data fusion under sensor failure modes. Empirical work on autonomous maritime safety uses synthetic scenarios or limited field trials, and the majority of published twins remain proof-of-concept with constrained datasets (Lee et al., 2024). Work by Bolbot et al. offers safety evaluation using digital-twin simulation series but places emphasis on scenario enumeration rather than on integrating probabilistic perception models into control verification (Bolbot et al., 2020). The most recent synthesis shows growing interest in operational twins, yet open questions persist: how to calibrate twin parameters with sparse at-sea telemetry, how to incorporate wake and viscous interaction effects at small separations, and how to manage computational cost while retaining fidelity for real-time decision support (Wang et al., 2025). The literature therefore supports a targeted contribution: an architecture that layers model fidelity adaptively so that high-fidelity CFD validates safety-critical control modes while reduced-order surrogates drive real-time planning during operations.

Research Questions

Which twin architecture yields the best trade-off between predictive fidelity and runtime tractability for collision avoidance? How does integrated uncertainty from perception and hydrodynamics alter control recommendations compared to conventional deterministic planners? What validation protocol will satisfy both engineering robustness and the evidentiary needs of the UK and IMO MASS frameworks? The research will answer these through model development, controlled simulation experiments and incremental validation using HIL and limited sea trials.

Methodology and Feasibility

The approach uses a mixed-methods pathway combining computational modelling, software engineering and experimental validation. First, develop a modular twin composed of three coupled layers: (a) a high-fidelity hydrodynamic core using CFD for off-line validation, (b) a reduced-order behavioural surrogate obtained via system identification for on-line prediction, and (c) a probabilistic perception twin that models sensor uncertainties and detection latencies. MATLAB/Simulink and ANSYS Twin Builder will host the control integration and surrogate models; OpenFOAM or commercially available CFD will handle viscous interaction cases. Second, construct scenario libraries reflecting port-approach interactions: cross-traffic wakes, restricted-visibility encounters and multi-actor constrained passages. Each scenario will be parameterised and sampled using probabilistic approaches to capture environmental variability. Third, validate the twin through a three-stage protocol: unit verification at component level, integrated simulation validation comparing twin predictions against high-fidelity CFD and reference experiments, and HIL trials where the autonomy stack runs against the surrogate in real time. The design will adopt continuous integration practices and automated test benches to ensure reproducibility. Data sources include public AIS datasets, synthetic sensor logs, and open environmental datasets; risk to data quality will be mitigated through cross-validation and synthetic augmentation. Resource needs are modest: access to an HPC node for batch CFD runs, ANSYS Twin Builder trial license available via the module, a small HIL rig with inertial measurement sensors and a scaled model for limited tow-tank or basin tests. Timeline spreads over 10 months with iterative deliverables: model prototypes at month three, surrogate validation at month six, HIL trials at month eight and final validation at month ten. The methodology aligns with UN SDG 9 by advancing industrial digitalisation and resilient infrastructure through validated simulation workflows.

Validation Strategy and Metrics

Validation will use quantitative collision-avoidance metrics and statistical hypothesis testing. Key performance indicators include time-to-conflict reduction, rate of COLREG-compliant manoeuvres, false-positive avoidance triggers and control-induced instabilities. Each metric will be computed across Monte Carlo draws of sensor and environmental uncertainty. Success criteria set target reductions: a minimum 30 percent improvement in time-to-conflict detection relative to baseline planners and fewer than 5 percent unstable recovery events in constrained scenarios. Confidence bounds will be reported with 95 percent intervals derived from bootstrap sampling. Regulatory acceptability will be evaluated by mapping evidence artefacts to the UK Maritime Autonomous Systems Regulatory Framework and to IMO guidance clauses cited in the proposal.

Expected Outcomes and Impact

Deliverables include a documented twin architecture, validated surrogate models, a reusable scenario library and an evidence bundle intended for regulatory review. The immediate technical impact will be clearer quantification of margins in collision-avoidance under sensor and hydrodynamic uncertainty. Operationally, the twin will allow designers to trade hull form and controller aggressiveness against safety metrics before sea trials. The broader impact touches procurement and certification: operators will gain a repeatable method to produce compliance evidence, reducing costly iterative sea tests. Contributions to scholarship will clarify integration pathways between CFD-based validation and runtime twin operation.

Ethical, Regulatory and Operational Considerations

Safety remains the primary ethical constraint. The project embeds conservative decision thresholds during HIL and sea validations to avoid risk to assets. Data governance will anonymise AIS-derived traces and follow institutional policies for sensitive information. The twin will not be used in live autonomous operations without a human-in-the-loop supervisory regime. Regulatory alignment will document how model updates are versioned and how retraining or parameter recalibration is traceable. The proposal accepts that regulators will require accessible audit trails and offers versioned artefacts and reproducible test cases to satisfy that requirement.

Project Risks and Mitigations

1. 1. • Risk: CFD-to-surrogate fidelity loss. Mitigation: use adaptive sampling and error-bounding techniques; retain conservatism in control limits.
      • Risk: Insufficient high-quality telemetry. Mitigation: augment with synthetic data and perform sensitivity analysis to demonstrate robustness to data gaps.
      • Risk: Computational cost for real-time usage. Mitigation: employ reduced-order models and surrogate caching for critical manoeuvres.
      • Risk: Regulatory mismatch. Mitigation: early stakeholder mapping and deliver evidence formats used by UK and IMO reviewers.

Resources and Timeline

1. 1. 1. Months 1–3: Requirements capture, scenario library development, initial CFD cases.
      2. Months 4–6: Surrogate model training, integration with control stack, mid-project validation.
      3. Months 7–8: HIL trials and iterative refinement.
      4. Months 9–10: Final validation, documentation and regulatory evidence preparation.

Conclusion

Practical constraints mandate a restrained scope. Focusing on collision avoidance for a 500 GT ASV produces a tractable problem with clear industry relevance. The proposed layered twin architecture links the fidelity of CFD validation with the tractability of reduced-order models, while treating perception uncertainty as a first-class citizen in control verification. The plan produces repeatable evidence for safety claims and establishes a route for future expansion to fleet-level twins.

Bibliography

Bolbot, V., Theotokatos, G., Bujorianu, L.M., Boulougouris, E. and Vassalos, D., 2020. A novel method to evaluate the safety of autonomous marine systems through digital twin simulation series. Safety, 6(4), p.61.

Lee, S., Kim, N. and Han, S., 2024. Digital twins enable shipbuilding: Applications, challenges, and future directions. Alexandria Engineering Journal, 89, pp.1-18.

Lu, Q., Parlikad, A.K. and Woodall, P., 2021. Developing a dynamic digital twin at a system level: A marine engineering case study of an offshore vessel powertrain. Ocean Engineering, 232, p.109122.

Negreiros, J., Santos, T.A. and Guedes Soares, C., 2023. Digital Twins in the Marine Industry: A Review. Electronics, 12(9), p.2025.

Raptis, T.P., Faiella, G., Giallanella, L., Kyriakopoulos, G.L. and Kameas, A., 2022. Ship’s Digital Twin—A Review of Modelling Challenges and Applications. Journal of Marine Science and Engineering, 10(8), p.1136.

Wang, J., Zhang, H., Liang, Y. and Zhang, Y., 2025. The application and development of digital twin in the marine domain. Ocean Engineering, 295, p.116947.

Xu, C., Wang, X., Li, J. and Yang, Y., 2022. Cyber–physical integration of digital twins for intelligent ship design and operation. Journal of Marine Science and Engineering, 10(12), p.1854.

___________________________________________________________________________________________________

Sample (ii).

Research the critical methodological gaps in current digital twin literature regarding data-fusion resilience and predictive validity under harsh maritime conditions.

Evaluate a mixed-methods methodology, integrating MATLAB/Simulink and CFD, to validate ASV control systems against IMO MASS compliance criteria.

Discuss the alignment of advanced ASV control validation via simulation with UN SDG 9 and the necessity of auditable ‘black box’ decision logs.

Illustrate the expected contributions of a digital twin framework to safer autonomy, focusing on pre-emptive fault mitigation and scalable fleet resilience.

Research Proposal: Real-Time Predictive Control in Autonomous Surface Vessels via High-Fidelity Digital Twins

1. Title and Abstract

Digital Twin Frameworks for Real-Time Predictive Control in Autonomous Surface Vessels: Enhancing Collision Avoidance Protocol.

Maritime Autonomous Surface Vessels (ASVs) require validated, pre-emptive control systems to manage operational risks in dynamic environments. This proposal outlines the development of a high-fidelity digital twin for a 500 GT ASV, explicitly targeting a critical failure point: collision avoidance in congested waterways. The study will fuse physics-based modeling (CFD) with real-time control system simulation (MATLAB/Simulink) to create a self-correcting twin, capable of maintaining synchronicity under data degradation and environmental volatility. Furthermore, the research addresses the inherent methodological gaps in data fusion and regulatory alignment with emerging IMO MASS Codes. The methodology involves a mixed-methods approach using open-source AIS data for scenario generation and a structured timeline for iterative model validation. Expected outcomes include a certifiable framework for ASV control and significant contributions toward establishing industry standards for simulated autonomy validation.

2. Introduction and Rationale

The shift toward sustained, unsupervised operation of ASVs mandates a design paradigm centered on predictive safety, which traditional naval architecture principles often fail to fully accommodate. Digital twin technology provides the requisite fiducial mechanism to represent a vessel’s operational state and performance envelope in a virtual context, allowing pre-emptive fault mitigation (Gao et al., 2024). Building an autonomous vessel requires that its core control system does more than react to an imminent threat; it must anticipate system failures or environmental non-conformities before they translate into catastrophic events. This capability relies heavily on the twin’s ability to ingest high-frequency sensor telemetry and consistently output a reliably validated state vector mirroring its physical counterpart. The research focuses on a specific, high-risk operational challenge: collision avoidance in densely trafficked areas, such as the English Channel or major port approaches, where human-centric decision latency is already a primary failure mode. Such scenarios demand a control system capable of rapid, optimized path adjustments under constrained maneuvering characteristics. Focusing the proposal on a 500 GT ASV, an intermediate commercial class, provides a scalable and economically relevant testbed for refining autonomous control logic.

Justification for this intervention is compelling because current regulatory frameworks demand assured performance validation. The International Maritime Organization’s (IMO) interim guidelines on the Maritime Autonomous Surface Ships (MASS) Code, expected to crystallize around 2025, place a significant burden on the industry to provide evidence of safe, predictable behavior from unmanned systems (IMO, 2023). A reliable digital twin offers the only economically and safely viable route to generate the volume and diversity of stress-test scenarios necessary for certification. Moreover, failure to implement a robust, validated collision avoidance system inhibits the adoption of ASVs in precisely the high-value logistics corridors where the technology promises the greatest efficiency gains. Consequently, this research is positioned not merely as a technical exercise but as a critical enabling factor for the regulated commercialization of autonomous shipping.

Problem Statement

Current digital twin implementations in naval architecture fail to maintain predictive validity for collision avoidance protocols in ASVs under degraded operational states, specifically when facing high-density, stochastic maritime traffic coupled with significant sensor data quality deterioration (O’Hare & Heavey, 2024). The core problem lies in the inability of existing models to accurately capture the non-linear coupling effects between aggressive control inputs and hull hydrodynamics while simultaneously compensating for real-time data uncertainty. This deficiency creates a certification barrier, as the control logic cannot be guaranteed safe when operating outside the clean, benign conditions often used in initial simulations. A fundamental need exists for a digital twin framework that rigorously models and validates ASV maneuvers against the complexity of real-world, congested maritime environments.

Research Objectives and Questions

The primary objective is to develop and validate a high-fidelity digital twin framework for a 500 GT ASV to demonstrably improve the reliability of its collision avoidance protocol under high-traffic and low-data-quality conditions.

The study is guided by the following research questions:

1. To what degree can a high-fidelity digital twin, integrating MATLAB/Simulink control logic with CFD-validated hydrodynamics, maintain predictive accuracy for ASV maneuvering under simulated congestion and severe environmental forces?
2. How must the data-fusion architecture within the digital twin be structured to tolerate and compensate for real-time sensor data degradation (e.g., loss of radar returns, intermittent AIS feeds) while sustaining collision risk assessment reliability?
3. Which set of performance metrics can rigorously validate the ASV’s pre-emptive collision avoidance maneuvers against COLREGs and proposed IMO MASS Code safety thresholds within the digital twin environment?
4. What are the key scalability considerations—in terms of computational resources and data flow—for deploying this validated digital twin framework across a diverse commercial ASV fleet?

3. Literature Review

The foundational work on digital twins in maritime contexts often focuses on asset integrity and predictive maintenance, treating the twin as a diagnostic tool for isolated components like propulsion systems or hull stress monitoring (Coraddu et al., 2019). Studies successfully apply twins to optimize propulsion control under stable conditions, demonstrating efficiency gains by comparing sensor-read engine performance against modeled baselines. However, these applications typically involve steady-state conditions and do not engage with the instantaneous, complex, non-linear dynamics required for emergency maneuvering. Furthermore, while the general theoretical utility of digital twins for ASV systems is well-established, moving from a descriptive or diagnostic twin (Level 1-2) to a true predictive and prescriptive twin (Level 4-5) remains a computational and conceptual challenge (Gao et al., 2024). The current literature thus suffers from a lack of integrated models where the control system, the vessel’s hydrodynamics, and a stochastic external environment interact in real-time, especially for high-risk, time-sensitive functions like collision avoidance.

A persistent methodological gap centers on validating ASV behavior against external, dynamic entities under regulatory constraints. Existing collision avoidance research frequently relies on simplified mathematical models or static representations of the COLREGs (Collision Regulations) environment, failing to account for the unpredictable human element in surrounding traffic. Consequently, most studies neglect to model the cyber-physical security risks alongside the operational risks inherent in remote data transmission and control loop closure, a clear deficiency for regulatory acceptance (O’Hare & Heavey, 2024). To be fair, modeling the intricate, rapid force changes exerted on a hull during an evasive maneuver requires specialized hydrodynamic analysis. Therefore, a reliance on purely data-driven models, which lack the underlying physics constraint of computational fluid dynamics (CFD) analysis, introduces a significant and unacceptable degree of uncertainty when attempting to validate safety-critical functions. The current research aims to synthesize the descriptive power of data-driven models with the deterministic constraint of physics-based models to address this critical void.

4. Methodology and Feasibility

The research will adopt a mixed-methods simulation approach composed of two interdependent modeling streams: Control System Development and Hydrodynamic Validation. The core digital twin will be built in MATLAB/Simulink to facilitate the development, tuning, and closed-loop testing of the ASV’s control architecture, including the advanced collision risk algorithm. This environment provides the necessary flexibility for implementing the pre-emptive trajectory planning logic based on fusing real-time simulated AIS/LiDAR data. The model will incorporate non-linear vessel dynamics equations (such as the Norrbin or Abkowitz models) to simulate vessel motion accurately.

The second stream involves Computational Fluid Dynamics (CFD) analysis using software like ANSYS Fluent or OpenFOAM. The CFD modeling will focus specifically on validating the high-stress hydrodynamic responses of the 500 GT hull form, particularly during aggressive rudder actions, heel angle limits, and acceleration/deceleration under irregular wave patterns. The resulting force coefficients and hydrodynamic damping parameters, derived from the computationally expensive CFD, will then be injected back into the less computationally demanding, real-time Simulink model. This two-way coupling ensures the Simulink control twin possesses high fidelity for critical maneuvering scenarios. Feasibility is high because the project leverages existing ANSYS Twin Builder trial access and readily available open-source marine datasets (e.g., Kaggle AIS datasets, wave buoy data) for training and scenario injection, mitigating initial cost and data collection burdens.

5. Ethical Considerations

Ethical responsibility in ASV development is centered on ensuring the “digital twin” accurately reflects a conservative standard of safety greater than, or equal to, that of a human-crewed vessel. The primary ethical consideration revolves around liability and algorithmic decision-making concerning collision scenarios. The research must ensure the collision avoidance algorithm, tested in the twin, adheres strictly to the spirit and letter of the COLREGs, prioritizing the safety of life at sea, even at the potential expense of vessel damage. Data privacy presents a secondary, although necessary, concern; all raw AIS and environmental data used will be anonymized or aggregated, protecting the operational security of commercial vessels used for the scenario generation. Furthermore, the final framework must include transparent logging of the twin’s decision-making process—creating an auditable ‘black box’—to satisfy future regulatory demands on accountability for autonomous systems failures.

6. Timeline and Resources

The project is structured into three sequential phases over a nine-month period, ensuring iterative development and validation cycles.

1. Phase I (Months 1–3): Model Construction and Baseline Calibration. Focuses on building the core Simulink dynamics model, developing the CFD mesh for the 500 GT hull, and performing initial hydrodynamic validation runs to calibrate the Simulink model coefficients against CFD results. Resources primarily involve computational access to the specified software packages and open-source data acquisition.
2. Phase II (Months 4–6): Control Logic Integration and Scenario Testing. Involves implementing the probabilistic collision avoidance algorithm, integrating the data-fusion risk mitigation layer, and running high-density traffic simulations using historical AIS data. This phase requires significant processing power for thousands of simulation runs.
3. Phase III (Months 7–9): Fidelity Quantification and Regulatory Alignment. Concentrates on quantifying the twin’s predictive accuracy under degraded data quality, compiling the results, and formulating the final proposal that maps the framework outputs to IMO MASS Code requirements. Personnel include the primary researcher and supervisory academic support.

7. Expected Outcomes and Impact

This research will deliver two concrete contributions: a validated digital twin architecture and a quantifiable fidelity metric for ASV control systems. The twin framework, capable of predicting collision risk with high temporal resolution and validated against real-world hydrodynamics, moves the industry beyond simple reaction-based control toward true pre-emptive autonomy. The fidelity metric will allow designers and regulators to objectively assess a twin’s performance under non-ideal, safety-critical conditions, a vital step toward regulatory acceptance. The study’s focus on verifiable safety directly contributes to the United Nations Sustainable Development Goal 9 (Industry, Innovation, and Infrastructure) by promoting resilient, sustainable industrial technologies through scientific research and upgrading. Scalability to diverse fleet operations is assured because the proposed MATLAB/Simulink framework is inherently modular; the core control logic can be adapted to larger vessel classes by simply re-running the initial CFD validation phase to generate new hydrodynamic parameters. This approach offers a practical blueprint for integrating predictive analytics into existing vessel management systems, driving fleet resilience.

8. References/Bibliography

Coraddu, A., Oneto, L., Baldi, F., & Anguita, D. (2019). Development of a digital twin for the ship energy management. Energy Conversion and Management, 196, 159-173.

Gao, Z., Duan, B., Wang, R., & Yin, Z. (2024). A review of digital twin technology for maritime autonomous surface ship (MASS) systems. Ocean Engineering, 300, 117210.

IMO. (2023). Interim Guidelines on the Generic Goal-Based MASS Code. MSC.1/Circ.1648. International Maritime Organization.

O’Hare, S., & Heavey, C. (2024). Cyber-Physical Security Challenges in Maritime Digital Twins: A Conceptual Framework. Journal of Marine Science and Engineering, 12(4), 587.

Raza, N., Wang, J., & Wang, J. (2022). Towards Integrated Digital-Twins: An Application Framework for Autonomous Maritime Surface Vessel Development. Journal of Marine Science and Engineering, 10(10), 1469.

United Nations. (2023). The Sustainable Development Goals Report 2023: Special Edition. United Nations, New York.

_______________________________________________________________________________________________

Write a research proposal addressing digital twin applications in autonomous surface vessel design for multi-jurisdictional Asian maritime environments.

• Assess regulatory compliance pathways for digital twin-based autonomy systems across divergent Asian port authority frameworks and IMO guidelines.
• Compose a methodology for validating digital twin architectures through hardware-in-the-loop testing and computational fluid dynamics simulation.

Sample Research Proposal (iii).

Research Proposal: Digital Twin Integration for Port Congestion Management in a 500 GT Autonomous Surface Vessel Operating Across Asian Maritime Corridors

Problem Statement and Regional Context

Asian ports move half of global container throughput, yet infrastructure expansion lags behind traffic growth. Mumbai handles over four million TEU annually within constrained berth availability. Shanghai processes more than forty million TEU across terminals that contend with tidal windows and narrow approach channels. Singapore coordinates thousands of vessel movements weekly through strait passages where margins shrink under poor visibility. Dubai integrates heavy Suezmax traffic with regional feeder services in confined turning basins. Autonomous surface vessels promise efficiency gains, but existing control architectures treat harbours as static obstacles rather than dynamic systems where queuing, weather and traffic interact. A 500 GT ASV sized for feeder routes or short-sea operations faces particular pressure: insufficient displacement to override cross-currents during slow-speed manoeuvres, yet large enough to generate liability if collision occurs. Port congestion amplifies risk because berthing delays force vessels into holding patterns where fuel consumption rises and collision probabilities increase under spatial compression. Current autonomy stacks rely on perception-to-planning loops that respond to immediate obstacles but lack predictive capacity for harbour-scale congestion events. Digital twin technology offers a route to embed forward simulation into control decisions, allowing the vessel to anticipate berth availability, adjust approach timing and select routes that minimise exposure to congested zones.

The IMO 2025 interim MASS guidance establishes safety expectations but leaves validation methods underspecified, particularly for harbour operations where environmental coupling dominates vessel behaviour. Regional port authorities across Asia maintain heterogeneous regulatory frameworks: Indian ports emphasise pilotage retention, Chinese terminals prioritise throughput optimisation, Singapore enforces strict traffic separation, Middle Eastern hubs balance commercial speed with security protocols. A digital twin that supports compliance evidence across these jurisdictions will provide operators with a repeatable method to demonstrate safety without costly iterative sea trials. The proposal therefore positions the twin as both an engineering tool for design refinement and a compliance artefact for regulatory review. Focusing on port congestion management narrows the scope to a tractable problem with measurable outcomes, avoiding the diffusion that undermines broader autonomy research.

Literature Review

Recent work documents digital twin applications in shipbuilding and operations but reveals uneven treatment of harbour-specific challenges. Zhou et al. (2023) present a digital twin framework for autonomous surface vessels in complex environments, emphasising sensor fusion and real-time decision support, yet the validation cases focus on open-water scenarios rather than confined harbour manoeuvres where hydrodynamic interaction effects become pronounced. The study demonstrates reduced-order surrogate models for vessel dynamics but does not integrate port-level scheduling data or berth availability forecasts, leaving a gap in operational planning capacity. Kumar et al. (2022) examine digital twin adoption in Indian shipyards for predictive maintenance and design optimisation, highlighting integration challenges with legacy systems and sparse telemetry, though the case studies remain shore-based and do not address runtime control validation for autonomous operations. The authors identify data interoperability as a persistent barrier when linking design-phase twins with operational systems, a concern that compounds in multi-jurisdictional contexts where data standards vary.

Al-Shehhi and Al-Habsi (2021) explore digital twin deployment for smart port infrastructure in the Middle East, particularly for vessel traffic management and autonomous berthing support. The research demonstrates value in integrating harbour environmental models with vessel position tracking, yet the twin architecture lacks dynamic coupling to vessel control systems, treating the twin primarily as a monitoring tool rather than a predictive decision aid. Wang et al. (2020) integrate artificial intelligence with digital twin simulation for autonomous ship control, proposing neural network surrogates to capture complex manoeuvring behaviour under parameter uncertainty. The methodology shows promise in reducing computational load for real-time applications, but validation remains limited to single-vessel scenarios without consideration of multi-actor interactions typical of congested ports. Zhang and Liu (2024) develop a multi-scale digital twin for ship propulsion optimisation, achieving real-time performance prediction through hierarchical model decomposition. The framework addresses computational tractability but does not extend to harbour operations where external environmental forces dominate propulsion efficiency.

Rahman and Basheer (2023) survey digital twin trends across Asia-Pacific maritime autonomy, identifying regulatory fragmentation and limited cross-border data sharing as impediments to scalable twin deployment. The review highlights successful pilot projects in Singapore and China but notes that most implementations address single operational aspects such as fuel optimisation or structural monitoring, rather than integrated mission planning that spans route selection, berthing and congestion response. The literature therefore supports a targeted contribution: a twin architecture that layers harbour-scale environmental prediction with vessel motion surrogates and schedule integration, designed explicitly for multi-jurisdictional deployment across Asian ports. The gap centres on validation protocols that satisfy both engineering robustness requirements and the evidentiary needs of diverse regional regulatory frameworks.

Research Questions

The investigation addresses three core questions. Which digital twin architecture provides sufficient predictive fidelity for port congestion management while maintaining computational tractability for runtime deployment aboard a 500 GT autonomous surface vessel? How does integration of harbour-specific environmental models and berth scheduling data alter route planning and berthing timing decisions compared to conventional reactive autonomy frameworks? What validation pathway will generate compliance evidence acceptable to port authorities across Mumbai, Shanghai, Singapore and Dubai, given their divergent regulatory requirements and operational priorities? Answering these questions requires model development, scenario-based simulation experiments and incremental validation through computational fluid dynamics, hardware-in-the-loop testing and limited harbour trials.

Methodology and Feasibility

The approach adopts a mixed-methods development strategy combining computational modelling, software integration and experimental validation. The twin architecture comprises three coupled layers: a harbour environmental module that forecasts tidal currents, wind patterns and traffic density using historical port data and real-time updates; a vessel motion surrogate derived from computational fluid dynamics validation and system identification techniques, providing rapid prediction of manoeuvring response under environmental forcing; a route planning and scheduling interface that ingests berth availability forecasts and optimises approach timing to minimise holding patterns and collision exposure. MATLAB/Simulink will host control integration and surrogate models, ANSYS Twin Builder will manage multi-physics coupling, OpenFOAM will handle viscous interaction cases for harbour hydrodynamics. The harbour environmental module will draw on publicly available datasets including tidal records from regional meteorological agencies, automatic identification system logs from Maritime and Port Authority archives and port berth schedules where accessible through terminal operator partnerships. Synthetic augmentation will address data gaps, with sensitivity analysis quantifying robustness to missing or degraded inputs.

Scenario libraries will reflect operational realities across target ports: tidal gate constraints in Mumbai, restricted visibility during monsoon transitions, high-frequency ferry traffic in Singapore Strait, berth allocation conflicts in Dubai’s multi-terminal complex. Each scenario will be parameterised to capture variability in traffic density, environmental conditions and schedule disruptions, with Monte Carlo sampling generating ensembles for statistical validation. Validation proceeds through a three-stage protocol. Component verification confirms individual module accuracy against reference datasets and first-principles models. Integrated simulation validation compares twin predictions against high-fidelity computational fluid dynamics for representative harbour approach cases, with error bounds established through systematic grid convergence and uncertainty quantification. Hardware-in-the-loop trials execute the autonomy stack against the surrogate twin in real time, injecting realistic sensor noise and latency to stress-test control robustness. Limited harbour trials in controlled sections of participating ports will provide empirical benchmarking, subject to port authority approval and safety protocols. The timeline spans ten months: months one through three establish requirements and develop initial environmental models; months four through six integrate vessel surrogates and complete component verification; months seven through eight conduct hardware-in-the-loop testing; months nine through ten execute harbour validation and prepare regulatory documentation. Resource requirements include access to high-performance computing for batch computational fluid dynamics runs, ANSYS Twin Builder trial licenses available through the module, partnerships with port authorities for data access and trial permissions, a small hardware-in-the-loop facility with inertial sensors and actuator emulators. The methodology aligns with UN Sustainable Development Goal 9 by advancing industrial digitalisation and resilient infrastructure through validated simulation frameworks that reduce physical testing costs and accelerate regulatory approval cycles.

Validation Strategy and Success Metrics

Validation employs quantitative metrics tied to operational efficiency and safety. Port entry latency measures elapsed time from harbour approach to berth allocation, with targets set at a minimum twenty percent reduction relative to baseline reactive autonomy. Collision risk exposure quantifies cumulative time spent within predefined proximity thresholds to other vessels or fixed structures during harbour transit, targeting a thirty percent reduction through improved route timing. Fuel consumption during harbour operations provides an economic metric, with expected reductions of fifteen percent through elimination of extended holding patterns. Control stability during low-speed manoeuvres assesses propulsion and rudder response under environmental coupling, with success defined as fewer than five percent of simulated cases exhibiting unstable oscillations or control saturation. Each metric will be computed across Monte Carlo ensembles spanning environmental variability and schedule perturbations, with confidence intervals derived from bootstrap resampling. Regulatory acceptability will be evaluated by mapping validation outputs to IMO MASS guidance requirements and to specific documentation standards maintained by participating port authorities, producing evidence packages formatted for review by maritime administrations and classification societies.

Expected Outcomes and Regional Impact

Deliverables include a documented twin architecture with open interfaces for third-party integration, validated surrogate models for harbour hydrodynamics and vessel motion, a reusable scenario library covering four major Asian ports, and compliance evidence bundles tailored to regional regulatory frameworks. Technical impact centres on demonstrable improvements in port congestion management, providing vessel designers and operators with quantified margins before committing to sea trials. Operational benefits extend to terminal efficiency: predictive route planning reduces holding pattern congestion, lowering harbour fuel consumption and freeing berth capacity for additional traffic. The broader contribution touches procurement and certification pathways, offering port authorities and vessel operators a repeatable method to generate compliance evidence that reduces reliance on costly iterative testing. For Asian maritime stakeholders facing infrastructure bottlenecks, the twin provides a scalable tool to optimise autonomous vessel integration without extensive physical modification to existing port facilities. Scholarly contributions clarify integration pathways between design-phase computational fluid dynamics validation and runtime twin operation, establishing precedents for multi-jurisdictional deployment where regulatory heterogeneity currently impedes technology adoption.

Ethical, Regulatory and Operational Considerations

Safety constraints dominate ethical obligations. The project embeds conservative decision thresholds throughout validation stages, maintaining human oversight during harbour trials and restricting autonomous operations to controlled test zones until validation targets are met. Data governance protocols will anonymise automatic identification system traces and harbour schedule data, adhering to institutional policies for sensitive commercial information and ensuring compliance with regional data protection frameworks. The twin will not be deployed in live autonomous operations without supervisory regimes that allow human intervention at all stages. Regulatory alignment requires traceability for model updates and parameter recalibration, with version control and audit trails designed to satisfy port authority review processes. The proposal acknowledges that regulators across Mumbai, Shanghai, Singapore and Dubai maintain divergent requirements, necessitating modular evidence packages that can be tailored to specific jurisdictions without fundamental rework of the underlying twin architecture. The design therefore prioritises transparency and auditability over optimisation, accepting performance trade-offs where necessary to maintain regulatory acceptability.

Project Risks and Mitigation

Several risks threaten project viability. Computational fluid dynamics to surrogate fidelity loss may produce unacceptable prediction errors; mitigation involves adaptive sampling techniques and error-bounded surrogate training, with conservative control margins retained where surrogate uncertainty remains high. Insufficient high-quality telemetry from participating ports could degrade environmental model accuracy; mitigation strategies include synthetic data augmentation validated through sensitivity analysis, demonstrating robustness to sparse inputs. Computational cost for real-time twin deployment aboard a 500 GT vessel may exceed onboard processing capacity; mitigation employs reduced-order models and surrogate caching for computationally intensive modules, with cloud offloading as a fallback for non-latency-critical functions. Regulatory acceptance barriers may emerge if port authorities reject simulation-based evidence without extensive physical validation; early stakeholder engagement and incremental validation releases will address this risk, building trust through transparent documentation and iterative feedback incorporation. Data access restrictions from commercial terminal operators may limit scenario library completeness; partnerships established during project initiation will include data-sharing agreements with fallback provisions for anonymised or aggregated datasets where proprietary concerns persist.

Resources and Timeline

Months one through three focus on requirements capture, stakeholder engagement with port authorities, initial environmental model development and scenario parameterisation. Months four through six integrate vessel motion surrogates, complete component-level verification and initiate integrated simulation validation against computational fluid dynamics benchmarks. Months seven through eight conduct hardware-in-the-loop trials, refine control interfaces and address integration issues identified during testing. Months nine through ten execute limited harbour validation trials, finalise compliance evidence documentation and prepare dissemination materials including technical reports and regulatory submissions.

Conclusion

Port congestion across Asian maritime corridors imposes operational costs that undermine the commercial case for autonomous surface vessels. Focusing on a 500 GT vessel navigating high-density terminals in Mumbai, Shanghai, Singapore and Dubai produces a tractable research problem with measurable outcomes and clear stakeholder value. The proposed layered digital twin architecture integrates harbour environmental forecasting with vessel motion prediction and schedule optimisation, addressing gaps in existing autonomy frameworks that treat ports as static environments. The validation pathway balances engineering rigour with regulatory acceptability, producing evidence suitable for review by diverse port authorities while maintaining computational tractability for runtime deployment. Success will be measured through quantified reductions in port entry latency, collision exposure and fuel consumption, with statistical confidence bounds and regulatory evidence packages demonstrating readiness for operational trials. The project establishes a foundation for scalable autonomous vessel integration across Asian ports, offering a repeatable methodology that can be extended to additional harbours and vessel classes as regional adoption of maritime autonomy accelerates.

Bibliography

Write My Paper