(PDF) Comprehensive Technical Framework for Dual-Platform Unmanned Underwater Systems: Design, Hydrodynamics, and Autonomous Nav

Comprehensive Technical Framework for Dual-Platform Unmanned Underwater Systems: Design, Hydrodynamics, and Autonomous Navigation for Autonomous Underwater Competition Capt. Noah Ugur Kilinc Preface The operational landscape of modern underwater robotics, exemplified by the "Autonomous Underwater Competition-Europe," has evolved into a multi-disciplinary challenge requiring high-fidelity systems engineering. This study provides an in-depth technical analysis of a dualplatform Unmanned Underwater Vehicle (UUV) architecture, engineered to navigate the complex "Line Following and Closed Area Inspection" mission profiles. By integrating a high-performance "Main Ship" (UUV) with a compact "Daughter-Ship" (Mini ROV), the framework addresses the critical trade-off between computational power and maneuverability in confined spaces. The research investigates the nexus of computational fluid dynamics (CFD) for axi-symmetric hull optimization—targeting an ideal length-to-diameter (L/D) ratio of 7.0—and advanced sensor fusion techniques like Factor Graph Optimization (FGO) for precise navigation in GPS-denied, 20-meter deep environments. This preface introduces a robust methodology for developing cost-effective yet research-grade subsea systems, bridging the gap between theoretical hydrodynamics and practical autonomous intervention. Keywords: Dual-Platform UUV Architecture, Factor Graph Optimization (FGO), Near-Wall Hydrodynamics, Subsea Thermal Management, Autonomous Pipeline Inspection 1 1. SYSTEM ARCHITECTURE AND MISSION PARAMETERS The Autonomous Underwater Competition technical specification defines a multi-stage challenge where vehicles must navigate underwater environments to perform tasks such as line following, pipeline inspection, and targeted intervention.1 The dual-vehicle architecture is specifically designed to address the "Line Following and Closed Area Inspection" theme, where the Main Ship serves as a transport and communication hub, and the Mini ROV enters confined spaces that are inaccessible to the larger platform.1 The mission is divided into two distinct depth stages: a five-meter initial dive followed by a 20meter deep-submergence phase.1 This progression dictates the structural safety factors required for the pressure vessels and the selection of sensors capable of functioning in variable light and pressure conditions.4 The Main Ship must maintain a constant telemetry link with the remote operator, relaying both its own state and the data acquired by the Mini ROV.1 Figure 1: The BlueROV2 is a high-performance, highly customizable underwater ROV designed for inspections, research, and exploration. With a six-thruster vectored configuration, opensource software, and modular expandability, it delivers professional capability without the high cost of traditional systems. 36 1.1. Physical Constraints and Scoring Optimization Scoring in the competition heavily favors compact, lightweight designs.1 Vehicles with a maximum dimension under 60 cm and a mass below 12 kg receive the highest point allocations.1 This creates a significant engineering trade-off: the Main Ship must be large enough to house a high-performance compute module, a substantial battery array, and the deployment mechanism for the Mini ROV, yet compact enough to avoid scoring penalties.1 2 Feature Max Dimension Target Weight Compute Depth Rating Propulsion Main Ship (UUV) Mini ROV (Daughter-Ship) ≤ 90 cm (60 cm optimal) 1 Compact for 36" Pipe 2 ≤14 kg (12 kg optimal) 1 ≤ 2 kg (Excluded from weight score) 1 NVIDIA Jetson Orin Nano 7 Raspberry Pi Zero 2W / ESP32 8 20 Meters 1 20 Meters 1 6-8 Thrusters (Vectored) 9 3-4 Thrusters (H-Bridge) 6 Figure 2: Deep Trekker PIVOT ROV. 2 Photo credit: Frontier Subsea 3 2. UNDERWATER HYDRODYNAMICS AND HULL OPTIMIZATION The hydrodynamic design of a UUV determines its energy efficiency, stability during sensor acquisition, and maneuverability in high-current environments.4 For the Main Ship, a streamlined axi-symmetric body is preferred to minimize the drag coefficient (Cd), while the Mini ROV requires a geometry optimized for the complex flow fields found inside confined pipes.2 2.1. Drag Resistance and Geometry Fluid resistance underwater is composed of skin friction and form resistance.12 At low velocities typical of competition ROVs (0,5 to 1,5 m/s), skin friction accounts for 60% to 70% of the total drag.12 The drag force (Fd) is quantified as follows: where ƿ is the density of the fluid (~1025 kg/m³ for seawater), v is the velocity, and A is the characteristic frontal area.4 Research into axi-symmetric bodies suggests that an optimum length-to-diameter (L/D) ratio of 6.5 to 7.0 provides the best balance between surface area (friction) and flow separation (form drag).12 While a sphere has the minimum surface area, its high form resistance makes it unsuitable for efficient transit.12 Property Prismatic Coefficient (Cp) Value/Formula 0.60 - 0.70 12 Optimizes volume vs. drag Reynolds Number (Re) 12 Fineness Ratio (L/D) Wall Effect Distance 7.0 (Optimal) 12 h/D < 0,5 11 Impact on Design Determines boundary layer state Balances friction and form drag Increases drag in confined spaces 4 Figure 3: AUV ‘Cormoran’ at IMEDEA 12 2.2. Near-Wall Hydrodynamics for In-Pipe Navigation The Mini ROV faces unique challenges during the "Enclosed Space Inspection" task.1 When moving close to a pipe wall or seabed, the fluid velocity increases in the narrow gap between the vehicle and the surface, leading to a localized drop in pressure according to Bernoulli's principle.11 This "wall effect" generates an attractive force that can pull the vehicle toward the surface, complicating autonomous depth control.11 Computational fluid dynamics (CFD) simulations of mini ROVs indicate that total resistance increases as the distance to the wall decreases, particularly at higher speeds.11 To compensate, the Mini ROV's flight controller must implement an adaptive gain schedule that increases vertical thrust as the proximity sensors detect the pipe floor or ceiling.11 Figure 4: The 3D model (a) and coordinate systems (b) of the ROV 11 5 3. STRUCTURAL ENGINEERING AND PRESSURE VESSEL MECHANICS A UUV operating at 20 meters depth must withstand an external pressure of approximately 0,3 MPa (3 bar absolute).1 Material selection and wall thickness calculations are governed by the need to prevent catastrophic collapse (implosion) while facilitating heat dissipation and maintaining neutral buoyancy.17 Figure 5: Schematic of the ICTINEUAUV software architecture. 17 Figure 6: ICTINEUAUV, the VICOROB-UdG Team’s entry for the SAUC-E 2006 competition.17 6 3.1. Material Science: Aluminum vs. Acrylic For professional and research-grade AUVs, Aluminum 6061-T6 is the preferred material for primary pressure vessels due to its high yield strength (276 MPa), excellent corrosion resistance when anodized, and superior thermal conductivity.19 Conversely, acrylic (PMMA) is often used for observation-class vehicles due to its transparency and low cost.5 However, acrylic is subject to "creeping"—a phenomenon where the material slowly flows and deforms under sustained pressure, which can lead to failure in submersions lasting longer than two weeks at depths exceeding 25 meters. Figure 7: Thermal Conductivity of Aluminum 6061 21 The minimum wall thickness (t) for a cylindrical enclosure to avoid yielding is calculated using the hoop stress (δ h) formula: where P is the external pressure and r is the internal radius.16 However, for underwater robotics, the failure mode is typically elastic buckling rather than material yielding.6 Incorporating a safety factor (SF) of at least 4.0 is standard practice to account for manufacturing tolerances and material fatigue. 7 Material Al 6061-T6 Acrylic Carbon Fiber Density (g/cm³) Thermal Conductivity (k) Yield Strength 2.70 23 167 W/mK 23 276 MPa 19 1.19 25 0.2 W/mK 25 70 MPa 25 Variable High (Anisotropic) 1.60 26 Figure 8: Frame composed of sealed carbon fiber tubes designed for use on the ROV FiberFish 26 Figure 9: Fully assembled ROV FiberFish 26 8 4. THERMAL MANAGEMENT AND INTERNAL HEAT DISSIPATION Sealed underwater enclosures act as thermal traps for high-performance electronics. In the Advanced category, the Main Ship requires significant computational power for real-time SLAM and image processing, typically provided by an NVIDIA Jetson Orin Nano module, which can draw up to 25W in "Supermode".7 4.1. Thermal Resistance Networks Heat dissipation underwater occurs via a multi-stage process: internal conduction from the component to the air or chassis, convection within the internal air, conduction through the enclosure wall, and finally, external convection to the surrounding water.28 Newton's Law of Cooling governs the rate of heat transfer (Q): where h is the heat transfer coefficient, and A is the surface area.32 Because water has a much higher heat capacity and thermal conductivity than air, the external convective resistance is negligible compared to the internal resistance.32 In an aluminum enclosure, mounting the CPU directly to the hull with a thermal spreader effectively turns the entire vehicle into a heat sink.29 Experimental data show that components submerged in water remain significantly cooler than those in air; for instance, a receiver coil reaching 43.2°C in air may drop to 19.2°C when submerged. 32 4.2. Mitigation Strategies for High-Load Electronics 1. Passive Conduction: Utilize aluminum internal chassis and thermal interface materials (TIMs) with high thermal conductivity (k~2,2 W/mK) to create a solid bridge between the SoC and the aluminum hull.29 2. Active Internal Cooling: High-speed fans should be used within the enclosure to circulate air, ensuring that heat is transferred from components not in direct contact with the hull to the cooler walls.28 3. Component Relocation: Thermally sensitive components, such as batteries and IMUs, should be placed in separate compartments or at the bottom of the enclosure to avoid heat rising from the main processor. 9 5. UNDERWATER POSITIONING AND "BLIND" NAVIGATION A critical requirement for the Autonomous Underwater Competition Advanced category is the ability to navigate autonomously and report coordinates without access to GPS. 1 This "blind navigation" requires a multi-sensor fusion approach combining inertial, acoustic, and velocity measurements. 35 5.1. Integrated Navigation Systems (INS) The Main Ship utilizes an INS/DVL/USBL integrated navigation system. 35 1. Doppler Velocity Log (DVL): The DVL measures the vehicle's velocity relative to the seafloor by transmitting acoustic pings and measuring the Doppler shift in the reflected signal.36 The Water Linked DVL A50 provides high-frequency velocity updates (up to 10 Hz), which are essential for station-keeping and dead reckoning.36 Figure 10: The DVL A50 is the world’s smallest Doppler Velocity Log (DVL) 36 2. Ultra-Short Baseline (USBL): To bound the drift inherent in dead reckoning, a USBL system provides absolute position fixes.37 A surface transceiver calculates the distance and bearing to the Main Ship's transponder based on time-of-flight and phase difference.37 The coordinate systems used in the paper include a body frame (b-frame), navigation frame (n-frame), USBL frame (u-frame), and geographic frame (g-frame). Detailed definitions can be found in [8]. The AUV is equipped with a USBL and DVL, as shown in Figure 11. 37 10 Figure 11: Sketch of AUV positioning. 37 Figure 12: Configurable Underwater Group of Autonomous Robots 51 3. Inertial Measurement Unit (IMU): A MEMS-based IMU provides high-frequency acceleration and angular velocity data. Magnetic interference from the vehicle's thrusters can cause significant yaw drift, requiring careful calibration or the use of a North-seeking gyro in higher-budget applications.35 11 System DVL USBL MEMS IMU Pressure Precision 0.027 m/s 35 1-4 m 35 High (Short-term) 0.01 m 17 Update Rate 5-10 Hz 1-2 Hz 100+ Hz 20-50 Hz Operational Constraint Requires bottom lock (<50m) 36 Requires line-of-sight to surface 41 Cumulative drift over time 35 Highly reliable depth ( ) 43 The layout of relevant equipment and the definition of two reference frames are shown in Figure 13: Figure 13: Equipment layout. 35 The navigation frame ON XNYN ZN has its origin on the surface and its axes pointing North, East and Down (NED reference frame); the body frame Ob Xb Yb Zb is centered in the center of gravity of the AUV, with the x axis pointing in the direction of the forward motion of the vehicle, the z axis pointing down and the y axis completing a right-handed reference frame 38. The main names and corresponding symbols in the body frame are shown in the table (Main names and symbols in the body frame). 35 12 Motion Linear Rotation Xb Yb Zb Drift X Y Z Velocity u v w Angle Ө ɸ Ψ Angle Velocity p q r Figure 14: Exploded view of the CougUV 51 5.2. Factor Graph Optimization (FGO) and Sequential Filtering To fuse these disparate data sources, modern UUVs employ Factor Graph Optimization (FGO). Unlike the traditional Extended Kalman Filter (EKF), which only considers the current state, FGO optimizes the entire trajectory by iterating over historical measurements.37 This allows for the correction of historical drift when an absolute USBL fix is received or when a visual landmark is re-identified (loop closure).37 The flowchart of the algorithm is shown in Figure 15. IMU can pre-integrate its own gyroscope, accelerometer, and magnetometer to directly calculate the attitude angle and velocity in the body frame. DVL provides corrected velocity observations, and USBL provides corrected position observations. The navigation information is ultimately output through the Extended Kalman filter. 35 13 Figure 15: The flow chart of the algorithm.35 14 6. COMMUNICATION AND DATA RELAY ARCHITECTURE The design requires a two-tiered communication system: the Mini ROV reports to the Main Ship, and the Main Ship reports to the operator.1 In the underwater domain, traditional RF communication is ineffective beyond depths of one meter due to high attenuation.41 6.1. Ethernet-over-Tether via HomePlug AV (IEEE 1901) High-bandwidth communication (HD video and telemetry) between the vehicles is achieved using HomePlug AV technology.47 Systems such as the Blue Robotics Fathom-X interface use frequency modulation (2-30 MHz) to send 100 Mbps Ethernet signals over a single twisted pair of wires.47 In this architecture, the Main Ship houses an internal network switch. The Mini ROV is connected to this switch via a micro-tether, effectively making it a node on the Main Ship's local network.2 The Main Ship's compute module (Jetson) acts as a gateway, aggregating the data from both vehicles and forwarding it over the primary umbilical to the surface operator's console.2 Figure 16: Fathom-X Tether Interface Board Set. Blue Robotics SKU: RB-Blu-30 48 15 Figure 17: HPAV Architecture 49 6.2. Telemetry Reporting Protocol: MAV Link 2.0 To manage the two vehicles, the MAV Link 2.0 protocol is used due to its lightweight binary serialization and support for multi-vehicle systems.52 ● ● ● System and Component IDs: The Main Ship is assigned System ID 1, while the Mini ROV is System ID 2. 53 This allows the operator's Ground Control Station (GCS) to distinguish between incoming telemetry streams. 53 MAV Link Forwarding: The Main Ship runs a MAV Link router (e.g., MAV Proxy or MAV Link2 Rest) that intercepts packets from the Mini ROV and encapsulates them for transmission to the surface.34 Position Reporting: The position calculated by the Main Ship's INS is broadcast as GLOBAL_POSITION_INT messages, while the Mini ROV's relative position within the pipe is transmitted using custom LOCAL_POSITION_NED or NAMED_VALUE_FLOAT messages. 52 16 Connection Medium Protocol Operator to Main Ship Neutral Umbilical (50m) HomePlug AV Main Ship to Mini ROV Micro-Tether (15m) HomePlug AV Telemetry UDP / MAV Link 2.0 Binary Serial Video RTSP / H.264 Streamed Bandwidth 80-100 Mbps 48 80-100 Mbps 48 Low Overhead 53 4-10 Mbps 57 Figure 18: Security threats and attacks against the MAV Link Protocol 53 17 Below is the over-the-wire format for a MAV Link v2 packet. Figure 19: MAV Link 2 Packet Format 55 18 7. AUTONOMOUS COMPUTER VISION AND SLAM The Mini ROV must autonomously identify the pipeline and locate clues.1 Given the constraints of a mini ROV, the implementation must be computationally efficient yet robust to underwater visual degradation.45 7.1. Pipeline Detection and Line Following The vision pipeline for "Line Following" involves several critical stages.59 1. Preprocessing: Images are converted to grayscale and denoised using Gaussian or median filters to mitigate the effects of turbidity and scattering.45 2. Segmentation: A lightweight Convolutional Neural Network (CNN) is employed to detect the unburied pipeline structure, achieving up to 72% precision in identifying various pipe types. 59 3. Line Extraction: Once the pipe is segmented, Canny edge detection and the Hough Transform are used to identify the centerline of the pipe, which serves as the path reference for the vehicle's flight controller.14 4. Adaptive Control: The detected line provides a cross-track error (ey) and a heading error (Ψe), which are processed by a PID or super-twisting controller to generate thruster setpoints. 15 Figure 20: Dead reckoning-derived mosaics of raw sonar images during two pipeline tracking missions in the 16-foot-wide Stevens tank. 61 The estimated trajectory is shown by the green line, and the detected pipeline is shown by the orange region (intensity denotes number of detections). Left: tracking of a straight pipeline (100ft) with some local bends and curves, right: Tracking of a sharply curved pipeline. 19 7.2. Visual Odometry and SLAM in ROS2 The system is built on the Robot Operating System 2 (ROS2) framework, utilizing nodes for image acquisition, processing, and control.44 For "blind navigation" in confined spaces, the Mini ROV employs Visual-Inertial-Depth (VID) fusion.58 When visual features are sparse, or illumination is poor, the system leverages proprioceptive velocity prediction—a deep-learning model trained to estimate linear velocity based on thruster commands and IMU data.63 Real-time state estimation is enabled by algorithms such as ReAqROVIO (Refraction-Aware Visual-Inertial Odometry), which models the refractive index of water to ensure accurate depth and distance estimates without requiring complex underwater camera calibration.63 Figure 21: Overview of the software architecture of the autonomy stack running on Ariel. 63 Figure 22: Overview of the RUSSO system, where imaging sonar fusion is integrated with the VIO system. The purple box and lines denote a new add-on to the VIO system. 58 20 Figure 23: Sensing and computing setup of Ariel, the custom underwater robot. 63 The robot contains an Alphasense core synchronized with a 5-camera and IMU system as its main sensing suite, an Nvidia Orin AGX running the perception and planning software, and a Flir Blackfly S color camera. The robot is actuated by 8 thrusters that enable 6 DoF actuation. A Pixhawk 6x flight controller is used as the low-level autopilot. Figure 24: Snapshots from both the synthetic 45 (a) and real-world field datasets (b), alongside the multi-camera, multi-sensor (inertial and DVL) SLAM results. The first column shows the estimated trajectory and the sparse point cloud generated during SLAM, while the remaining columns depict various scenes captured during data collection, highlighting challenges such as uneven and low lighting, limited visibility, haze, and backscatter. 21 Figure 25: Control system interconnection diagram 15 22 8. PROPULSION, ACTUATION, AND PAYLOAD SYSTEMS The Main Ship and Mini ROV require specialized propulsion systems to meet their respective mission objectives.9 8.1. High-Performance Thrusters The Main Ship's propulsion is based on the Blue Robotics T200 or T500 thrusters.65 The T200 is a fully-flooded, water-cooled brushless DC motor that is naturally pressure-tolerant to depths of 300 meters.9 For heavy-duty operations, the T500 generates up to 16.1 kgf of thrust—over three times that of the T200—making it ideal for the Main Ship when carrying the Mini ROV in strong currents.64 Figure 26: T200 and T500 Thrusters 9 Pa6rameter T200 Thruster T500 Thruster Maximum Thrust (24V) 5.2 kgf 16.1 kgf 65 Maximum Power 350 W 1044 W 65 Weight in Air 340 g 1100 g 65 Depth Rating 300 m 67 300 m 65 Housing Material Polycarbonate Glass-filled Polycarbonate 64 23 8.2. Torpedo Launcher and Manipulator Systems The "Intervention to Target" task requires the launching of five torpedoes.1 To comply with safety regulations, the mechanism must be mechanical or electromechanical.1 A spring-loaded magazine system triggered by a waterproof PWM servo is the most cost-effective and reliable approach.1 The torpedoes themselves are 3D-printed and ballasted to be slightly negatively buoyant, ensuring they follow a predictable trajectory toward the target.1 The Mini ROV is equipped with a single-axis manipulator arm, providing precise gripping and recovery capabilities for the "clue identification" phase.68 Figure 27: Optimized Hydrodynamic Design for Autonomous Underwater Vehicles 10 24 9. SYNTHESIS OF PREVIOUS RESEARCH AND HISTORICAL PRECEDENTS The development of the current dual-platform model draws heavily from successful precedents in international competitions and offshore operations.2 9.1. Lessons from SAUC-E and MATE ROV In the Student Autonomous Underwater Challenge Europe (SAUC-E), the Universität de Girona's "Sparus" AUV demonstrated the effectiveness of an open-frame design for high maneuverability in confined spaces.17 The use of Factor Graph Optimization for pipeline tracking was a key innovation that allowed Sparus to win the competition multiple times.37 Similarly, the MATE ROV competition archives provide a wealth of data on the use of 3Dprinted components and low-cost tether interfaces (HomePlug AV) for student-built vehicles.22 9.2. Commercial Mother-Daughter Systems The "PIVOT ROV" hosted by a work-class ROV is a direct industrial analog for the Autonomous Underwater Competition requirement.2 In offshore pipeline inspections, the larger work-class vehicle transports the PIVOT to the site, providing power and stability through high-current zones, before deploying the PIVOT for internal pipe navigation.2 A key takeaway from this integration is the necessity of a subsea telemetry interface and a custom hydraulic tether reel to manage the daughter-ship's cable, preventing entanglement in the host vehicle's thrusters.2 25 10. PROJECT MANAGEMENT AND DETAILED COST ANALYSIS Achieving research-grade performance while maintaining cost-effectiveness requires a strategic selection of components and a reliance on open-source software and rapid prototyping.51 10.1 Bill of Materials (BOM) for Research-Grade UUV System The following BOM outlines the costs for a professional-grade competitive system as of late 2025/early 2026. Component Category Description Recommended Product Compute High-performance AI NVIDIA Jetson Orin Nano Micro-controller Raspberry Pi 5 / Teensy 4.1 Primary Pressure Vessel Al 6061-T6 Anodized Tube Viewports/End caps Cast Acrylic / Al Caps Vectored Array 6x T200 + Basic ESCs Vertical/Lateral 2x T500 (Heavy Config) Seabed Tracking WaterLinked DVL A50 Global Positioning Cerulean USBL Mark II IMU / Depth Bar30 / BNO055 $150 17 Stereo/Processing Jetson Orin + IMX219 $200 7 Enclosures Propulsion Navigation Vision Est. Cost (USD) $500 7 $150 51 $1,200 $600 5 $1,620 66 $1,580 66 $8,710 77 $3,400 26 Communication Topside/Internal Fathom-X Interface Set $160 48 Mini ROV Primary Tether 50m Fathom Neutral $500 57 Hull / Propulsion 3D Printed / 4x micro $500 51 Estimated Total $19,270 10.2. Market Trends and Procurement Outlook The UUV market is valued at $1.9 billion in 2025 and is expected to grow at a CAGR of 7.3% through 2035. The "Observation Class" segment, which includes vehicles similar to those in this study, is growing at the fastest rate due to demand for infrastructure inspection and aquaculture monitoring. However, tariff adjustments and semiconductor supply chain shifts have increased the cost of advanced control units and sensors, requiring teams to prioritize components with long-term availability (e.g., NVIDIA Jetson modules are guaranteed until 2032).30 27 11. TECHNICAL SYNTHESIS AND STRATEGIC RECOMMENDATIONS The successful design of a dual-vehicle system for the Autonomous Underwater Competition hinges on the synergy between the Main Ship's stability and the Mini ROV's agility.1 11.1. Final Design Recommendations 1. Structural: Utilize a 6061-T6 aluminum hull for the Main Ship to ensure survival at 20 meters and to provide an efficient thermal path for the Jetson Orin Nano.19 For the Mini ROV, a 3D-printed impact-resistant frame is suitable, provided the electronics are housed in a small aluminum or thick-walled acrylic capsule.5 2. Navigation: Invest in a DVL for the Main Ship. It is the only reliable way to achieve the sub-decimeter precision required for autonomous task docking in low-visibility environments.36 The Mini ROV should rely on visual odometry and VID fusion, relaying its relative position back to the Main Ship's EKF.62 3. Telemetry: Implement a multi-node HomePlug AV network. This allows both vehicles to share a single high-bandwidth pipe to the surface, simplifying the tether management system (TMS) on the surface.47 4. Software: Adopt ROS2 Humble on Ubuntu 22.04. This ensures compatibility with modern SLAM packages like DROID-SLAM and allows for the implementation of Behavior Trees to manage complex mission transitions between the two vehicles.51 By adhering to this framework, the team can produce a vehicle that not only meets the competition's technical requirements but also serves as a robust platform for long-term underwater research and exploration.2 The combination of PhD-level algorithmic rigor and professional-grade mechanical design ensures optimal cost-to-performance and mission success in the challenging environment of Autonomous Underwater Competition. 28 CONCLUSION The comprehensive evaluation of the dual-platform UUV system demonstrates that successful autonomous underwater missions depend on the synergy of structural integrity, thermal management, and algorithmic rigor. Our findings confirm that utilizing Aluminum 6061-T6 for primary pressure vessels not only ensures survival at 0.3 MPa (20m depth) but also provides an essential thermal path for high-load AI modules like the NVIDIA Jetson Orin Nano, preventing throttling during real-time SLAM processing. Furthermore, the study identifies that mitigating "wall effects" through adaptive gain scheduling in flight controllers is vital for stable "Mini ROV" operations within narrow pipelines. The transition from traditional EKF to Factor Graph Optimization (FGO) significantly enhanced trajectory accuracy by enabling loop closure and historical drift correction. In conclusion, the adoption of a modular ROS2-based software stack and high-bandwidth Ethernet-over-Tether (HomePlug AV) communication provides a scalable foundation for future research in infrastructure inspection and complex autonomous underwater intervention. 29 References 1. Unmanned Underwater System Competition Technical Specification. 2. Deep Trekker. Work-Class ROV Hosts PIVOT ROV for High-Precision Offshore Inspections. accessed April 7, 2026, https://www.deeptrekker.com/resources/work-class-rov-hosts-pivot-rov-forhigh-precision-offshore-inspections 3. Zhao, C., Thies, P. R., & Johanning, L. (2022). Offshore inspection mission modelling for an ASV/ROV system. Ocean Engineering, 259, 111899. https://doi.org/10.1016/j.oceaneng.2022.111899 4. Underwater Vehicle Dynamics Model: A. Introduction. Naval Postgraduate School. https://faculty.nps.edu/brutzman/dissertation/chapter6.pdf 5. Subsea Tether management system for BlueROV2. Blue Robotics Community Forums, accessed April 7, 2026, https://discuss.bluerobotics.com/t/subsea-tether-management-system-forbluerov2/19894 6. Mohd Aras, M. S., Yi Xin, T., Zohedi, F. N., Bahar, M. B., Harun, M. H., Khamis, A., & Rizman, Z. I. (2025). Design, Development, and Testing of a Compact Underwater ROV for Inspection and Exploration. Journal of Studies in Science and Engineering, 5(1), 245–266. https://doi.org/10.53898/josse2025523 7. NVIDIA Jetson-Orin-Nano Series. https://openzeka.com/wp-content/uploads/2023/03/jetson-orinnano-datasheet-r4-web.pdf 8. Feng, X., Zhang, Y., Gao, A., & Hu, Q. (2024). Research on AUV Multi-Node Networking Communication Based on Underwater Electric Field CSMA/CA Channel. Biomimetics, 9(11), 653. https://doi.org/10.3390/biomimetics9110653. 9. Blue Robotics. What is a Thruster?. https://bluerobotics.com/thruster-basics/ 10. Fan, G., Liu, X., Hao, Y., Yin, G., & He, L. (2025). Optimized Hydrodynamic Design for Autonomous Underwater Vehicles. Machines, 13(3). https://doi.org/10.3390/machines13030194 11. Jiang, M., Chen, C., Suo, Z., & Dong, Y. (2025). Hydrodynamic Calculation and Analysis of a Complex-Shaped ROV Moving near the Wall Based on CFDs. Journal of Marine Science and Engineering, 13(6). https://doi.org/10.3390/jmse13061183 12. Bertram, Volker & France, Brest & Fr, Volker & Alvarez, A (2006). Hydrodynamic Aspects of AUV Design. The Fifth Conference on Computer and IT Applications in the Maritime Industries (COMPIT). https://www.researchgate.net/publication/228395786_Hydrodynamic_Aspects_of_AUV_Design 13. Sato, E., Liu, H., Orita, Y., Sakagami, N., & Wada, T. (2022). Cooperative path-following control of a remotely operated underwater vehicle for a human visual inspection task. Frontiers in Control Engineering, 3, 1056937. https://doi.org/10.3389/fcteg.2022.1056937 https://www.frontiersin.org/journals/control-engineering/articles/10.3389/fcteg.2022.1056937/full 14. Bobkov, V., Shupikova, A., & Inzartsev, A. (2023). Recognition and Tracking of an Underwater Pipeline from Stereo Images during AUV-Based Inspection. Journal of Marine Science and Engineering, 11(10). https://doi.org/10.3390/jmse11102002 15. Zhao, B., Li, S., Wang, X., Yang, M., Yu, X., Meng, Z., & Wan, G. (2025). Design of Control System for Underwater Inspection Robot in Hydropower Dam Structures. Journal of Marine Science and Engineering, 13(9). https://doi.org/10.3390/jmse13091656 16. Meena Rezkallah. ASME B31.3 Sample Wall Thickness Calculations, accessed April 7, 2026, https://www.littlepeng.com/single-post/2017/07/21/asme-b313-sample-wall-thickness-calculations 17. D. Ribas, N. Palomeras, P. Ridao, M. Carreras, E. Hernández. ICTINEUAUV Wins the First SAUCE Competition - 2007 IEEE International Conference on Robotics and Automation, Roma, Italy, 1014 April 2007. http://vigir.missouri.edu/~gdesouza/Research/Conference_CDs/IEEE_ICRA_2007/data/papers/124 8.pdf 18. Thruster User Guide. Blue Robotics. https://bluerobotics.com/learn/thruster-usage-guide/ 19. Phillip Keane. 6061 Aluminum vs. 2024 Aluminum. Differences in Properties, Strength, and Uses. https://www.thomasnet.com/articles/metals-metal-products/6061-aluminum-vs-2024-aluminum/ 20. Engineering 18-Bar Safety: Why Aluminum 6061-T6 is Superior for High-Pressure Aerosols, accessed April 7, 2026, https://goldensoarpackage.com/fr/engineering-18-bar-safety-whyaluminum-6061-t6-is-superior-for-high-pressure-aerosols/ 30 21. Liu Che. Thermal Conductivity of Aluminum 6061: A Detailed Report. Fiber Laser Cutting Machine & Laser Welding Manufacturer | Arcus CNC, accessed April 7, 2026, https://arcuscnc.com/thermalconductivity-of-aluminum-6061/ 22. Tips/Tricks/Hacks - MATE ROV Competition, accessed April 7, 2026, https://materovcompetition.org/resources/tips-tricks-hacks 23. Aluminium Alloy 6061.T6 Extrusions. AALCO. https://www.aalco.co.uk/datasheets/AluminiumAlloy-6061-T6-Extrusions_145.ashx 24. Huang, J., Wang, Y., & Chen, Y. (2025). Raspi2 USBL: An open-source Raspberry Pi-Based Passive Inverted Ultra-Short Baseline Positioning System for Underwater Robotics. ArXiv. https://arxiv.org/abs/2511.06998 25. 6061-T6 Aluminum vs. PC. MAKEITFROM. accessed April 7, 2026, https://www.makeitfrom.com/compare/6061-T6-Aluminum/Polycarbonate-PC 26. Wood, Joshua F., "Design and Deployment of A Compact ROV For Deep-Sea Research And Exploration" (2025). Open Access Master's Theses. Paper 2586. https://digitalcommons.uri.edu/theses/2586 27. ROV Kits, accessed April 7, 2026, https://materovcompetition.org/rov-kits 28. Josiah Haas. How Enclosure Design Impacts Heat Dissipation & Thermal Management. Bud Industries, accessed April 7, 2026, https://www.budind.com/blog/2026/01/how-enclosure-designimpacts-heat-dissipation-thermal-management/ 29. NVIDIA Jetson Thermal Design Guide: Unleash Peak AI Performance. Things Embedded. accessed April 7, 2026, https://things-embedded.com/us/white-paper/nvidia-jetson-thermal-designguide-unleash-peak-ai-performance/ 30. John E. Caruso. Jetson Orin: NVIDIA performance boost opens up new possibilities. accessed April 7, 2026, https://www.syslogic.com/blog/jetson-orin-nano-nx-nvidia-performance-boost-opensup-new-possibilities 31. Michael Wayde. A guide to thermal management solutions for electronic enclosures. https://assets.phoenixcontact.com/file/f0027d39-73da-4ad8-a506124d0f7f11b6/media/original?US_Guide-to-Thermal-Management-Solutions-for-Enclosures-Whitepaper_20260112_U008480A.pdf 32. Zhao, Q., Niu, S., Shao, Z., & Jian, L. (2024). Hull-Compatible Underwater IPT System with Enhanced Electromagnetic–Thermal Performance for USVs. Energies, 18(2), 237. https://doi.org/10.3390/en18020237 33. C. M. Tarhan, A. Feyzioğlu, and G. Brenner, “Cooling analysis and innovative design to increase heat transfer in sealed electronic devices,” Journal of Mechatronics and Artificial Intelligence in Engineering, Vol. 6, No. 2, pp. 22–39, Oct. 2025, https://doi.org/10.21595/jmai.2025.25166 34. Navigating an Autonomous Underwater ROV With GPS using ROS2 and PYMAVLINK, accessed April 7, 2026, https://discuss.bluerobotics.com/t/navigating-an-autonomous-underwater-rov-withgps-using-ros2-and-pymavlink/19699 35. Liu, J., Yu, T., Wu, C., Zhou, C., Lu, D., & Zeng, Q. (2024). A Low-Cost and High-Precision Underwater Integrated Navigation System. Journal of Marine Science and Engineering, 12(2). https://doi.org/10.3390/jmse12020200 36. DVL with UGPS for BlueROV2 - Documentation - Water Linked, accessed April 7, 2026, https://docs.waterlinked.com/dvl/bluerov-integration-dvl-ugps/ 37. Li, P., Liu, Y., Yan, T., Yang, S., & Li, R. (2022). A Robust INS/USBL/DVL Integrated Navigation Algorithm Using Graph Optimization. Sensors, 23(2), 916. https://doi.org/10.3390/s23020916 38. Allotta, B.; Caiti, A.; Costanzi, R.; Fanelli, F.; Fenucci, D.; Meli, E.; Ridolfi, A. A new AUV navigation system exploiting an unscented Kalman filter. Ocean Eng. 2016, 113, 121–132. https://doi.org/10.1016/j.oceaneng.2015.12.058 39. Knowledge Base. Water Linked, accessed April 7, 2026, https://www.waterlinked.com/knowledge 40. Water Linked DVL. Marine Thinking, accessed April 7, 2026, https://store.marinethinking.com/products/water-linked-dvl 41. Yuehan Jiang. Bernd-Christian Renner. Low-Cost Underwater Swarm Acoustic Localization: A Review - IEEE Xplore. Volume 12, 2024. DOI:10.1109/ACCESS.2024.3357359. http://ieeexplore.ieee.org/iel7/6287639/10380310/10411871.pdf 42. Underwater Acoustic Positioning & Navigation Systems. Evo Logics. accessed April 7, 2026, https://www.evologics.com/solution/underwater-positioning 31 43. E. Hernández, P. Ridao, M. Carreras, D. Ribas, N. Palomeras, A. El-fakdi, F. Chung, T. Almohaya, X. Ribas, G. García de Marina, J. Massich, N. Hurtós. ICTINEUAUV, un-Robot per a Competir. CCIA 2006: Congrés Català d'Intel.ligència Artificial. https://www.francoischung.com/wpcontent/uploads/2006/10/Hernandez-et-al._CCIA-2006.pdf 44. Rahman, Md. Musfiqur & Al Mahmud, Jubayer & Ali, Firoj & Rahim, Md Abdur & Sarker, Subrata & Yan, Bing. (2026). ROS2-based real-time autonomous mapping and navigation: Integrating visual SLAM and sensor fusion, with performance analysis under varying lighting conditions. Measurement. 268. 120695. 10.1016/j.measurement.2026.120695. https://www.researchgate.net/publication/400387480_ROS2-based_realtime_autonomous_mapping_and_navigation_Integrating_visual_SLAM_and_sensor_fusion_with_ performance_analysis_under_varying_light 45. Pushyami Kaveti. Ambjørn Grimsrud Waldum. Hanumant Singh. Martin Ludvigsen. Enhancing Situational Awareness in Underwater Robotics with Multi-modal Spatial Perception. 2025 IEEE ICRA Workshop on Field Robotics. https://arxiv.org/html/2506.06476v1 46. Student Autonomous Underwater Competition-Europe, accessed April 7, 2026, http://robolex.com/sauce/sauce-introduction/ 47. Fathom-X Tether Interface Board - Marine Thinking, accessed April 7, 2026, https://store.marinethinking.com/products/fathom-x-tether-interface-board 48. Fathom-X Tether Interface Board Set – Robot Shop, accessed April 7, 2026, https://www.robotshop.com/products/fathom-x-tether-interface-board-set 49. Sherman Gavette. HomePlug AV Technology Overview. Microsoft Download Center. April 18, 2006 https://download.microsoft.com/download/a/f/7/af7777e5-7dcd-4800-8a0ab18336565f5b/homeplugavwp_shermangavette.doc 50. SAUC-E competition – François Chung, Ph.D. | English, accessed April 7, 2026, https://www.francoischung.com/en/my-projects/sauc-e-competition 51. D. Ribas, N. Palomeras., X. Ribas, G. García de Marina, E. Hernàndez, F.Chung, N. Hurtós, J. Massich, A. Almohaya and J. Vila. ICTINEUAUV Takes the Challenge. Institute of Informatics and Applications. University of Girona. 2006-01-01. https://www.francoischung.com/wpcontent/uploads/2006/01/Ribas-et-al._UdG-2006.pdf 52. Aziz F and Loya A (2025) Multi-Mode MAVLink Test Bench AeroQT for Aircraft Simulation and Real-Flight Validation. Aerosp. Res. Commun. 3:14524. doi: 10.3389/arc.2025.14524, https://www.frontierspartnerships.org/journals/aerospace-researchcommunications/articles/10.3389/arc.2025.14524/full 53. Koubaa, Anis & Allouch, Azza & Alajlan, Maram & Javed, Yasir & Belghith, Abdelfettah & Khalgui, Mohamed. (2019). Micro Air Vehicle Link (MAVLink) in a Nutshell: A Survey. 10.48550/arXiv. 1906.10641. https://arxiv.org/pdf/1906.10641 54. Vladimir Djapic. Thomas B. Curtin. SAUC-E: Fostering the Next Generation of Scientists and Engineers. IEEE Robotics & Automation Magazine • September 2011. https://www.researchgate.net/profile/Thomas-Curtin/publication/252062305_SAUCE_Fostering_the_Next_Generation_of_Scientists_and_Engineers/links/564c9af508ae7ac727e1fca e/SAUC-E-Fostering-the-Next-Generation-of-Scientists-and-Engineers.pdf 55. Protocol Overview. MAV Link Guide, accessed April 7, 2026, https://mavlink.io/en/about/overview.html 56. MAV Link Messaging. PX4 Guide (main), accessed April 7, 2026, https://docs.px4.io/main/en/mavlink/index 57. Tether. accessed April 7, 2026, https://www.ardusub.com/introduction/hardware-options/requiredhardware/tether.html 58. Shu Pan. Ziyang Hong. Zhangrui Hu. Xiandong Xu. Wenjie Lu. Liang Hu. RUSSO: Robust Underwater SLAM with Sonar Optimization against Visual Degradation. https://arxiv.org/html/2503.01434v1 59. Da Silva, Y. M., Andrade, F. A., Sousa, L., De Castro, G. G., Dias, J. T., Berger, G., Lima, J., & Pinto, M. F. (2022). Computer Vision-Based Path Following for Autonomous Unmanned Aerial Systems in Unburied Pipeline Onshore Inspection. Drones, 6(12), 410. https://doi.org/10.3390/drones6120410 32 60. MATE ROV Competition. Wikipedia. https://en.wikipedia.org/wiki/MATE_ROV_Competition 61. Jinkun Wang, Tixiao Shan, Muthukumaran Chandrasekaran, Timothy Osedach, Brendan Englot. Deep Learning for Detection and Tracking of Underwater Pipelines using Multibeam Imaging Sonar - Stevens Institute of Technology. https://personal.stevens.edu/~benglot/Wang_ICRA_2019_UWPerceptionWorkshop.pdf 62. Geng, C., Chen, J., & Li, H. (2025). A Lightweight and Low-Cost Underwater Localization System Based on Visual–Inertial–Depth Fusion for Net-Cage Cleaning Robots. Journal of Marine Science and Engineering, 14(1), 48. https://doi.org/10.3390/jmse14010048 63. Singh, M., Dharmadhikari, M., & Alexis, K. (2025). Ariel Explores: Vision-based underwater exploration and inspection via generalist drone-level autonomy. ArXiv. https://arxiv.org/abs/2507.10003 64. The T500 Thruster. Blue Robotics, accessed April 7, 2026, https://bluerobotics.com/new-productthe-t500-thruster/ 65. T500 Thruster for High-Power ROVs, USVs, AUVs, and marine robotics, accessed April 7, 2026, https://bluerobotics.com/store/thrusters/t100-t200-thrusters/t500-thruster/ 66. Thrusters, Actuators, and Lights. Blue Robotics, accessed April 7, 2026, https://bluerobotics.com/product-category/thrusters/ 67. T200 Thruster. Blue Robotics, accessed April 7, 2026, https://bluerobotics.com/store/thrusters/t100-t200-thrusters/t200-thruster-r2-rp/ 68. ROV Accessories. Gr8Dane Marine Services Inc., accessed April 7, 2026, https://gr8dane.ca/rovaccessories 69. Oceanus ROV Accessories and Parts. Marine Nav Ltd., accessed April 7, 2026, https://marinenav.ca/rov-accessories.html 33