SeaVis: Modeling and Control of a Remotely Operated Towed Vehicle for Seabed Visualization and Mapping
Abdelhakim Amer, Aske Alstrup, Frederik Rasmussen, Yury Brodskiy, Andriy Sarabakha, Erdal Kayacan · 2026
Read SeaVis: Modeling and Control of a Remotely Operated Towed Vehicle for Seabed Visualization and Mapping by Abdelhakim Amer, Aske Alstrup, Frederik Rasmussen, Yury Brodskiy, Andriy Sarabakha, Erdal Kayacan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-14683.txt.
Introduction
Amer and team model and control SeaVis, a towed robot vehicle run from afar for seabed maps and views. Steady exact positioning matters for sharp seafloor maps. They derive a math model and build a gain-scheduled linear-quadratic regulator for depth and attitude control. In a high-fidelity simulation against a standard PID controller, the new controller is claimed to be tougher, more efficient, and far lighter on flap actuation.
Catalog record
| CREATOR | Abdelhakim Amer, Aske Alstrup, Frederik Rasmussen, Yury Brodskiy, Andriy Sarabakha, Erdal Kayacan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 31227 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.14683
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
SeaVis: Modeling and Control of a Remotely Operated Towed Vehicle for Seabed Visualization and Mapping Abdelhakim Amer, Aske Alstrup, Frederik Rasmussen, Yury Brodskiy, Andriy Sarabakha, Erdal Kayacan Abstract High-resolution seafloor mapping necessitates stable and precise positioning for underwater robots. This paper introduces a novel mathematical model for SeaVis remotely operated towed vehicles (ROTVs) and develops a gain-scheduled linear-quadratic regulator (LQR) for robust depth and attitude control. We validate the approach in a high-fidelity simulation, benchmarking the LQR against a conventional PID controller over a challenging seabed profile. The presented results demonstrate the LQR’s superior performance, with significantly enhanced robustness to disturbances, greater control efficiency, and substantially reduced flap actuation. The gain scheduling also confirms the controller’s effectiveness across the full operational velocity range. The complete simulation environment and controller are open-sourced 1 1 1 https://github.com/frederikt0ft/ROTV-modeling-and-control . I Introduction The exploration of subsea environments and accurate seabed mapping have become increasingly critical for environmental monitoring, scientific research [ 7 , 2 , 4 , 6 ] , offshore infrastructure development, and defense applications [ 12 ] . Events such as the Nord Stream 2 explosion, Baltic Sea internet and power cable damage [ 20 ] and foreign mapping of Nordic subsea infrastructure have heightened concerns about underwater infrastructure security and the need for enhanced surveillance measures. With approximately 80% of the ocean floor remaining unexplored [ 14 ] , the development of efficient and reliable underwater mapping systems has never been more urgent. Remotely operated towed vehicles (ROTVs) offer significant advantages over traditional remotely operated vehicles (ROVs) for seabed mapping applications. Unlike ROVs, which rely on onboard propulsion systems, ROTVs are passively towed behind surface vessels, enabling efficient coverage of large survey areas without requiring additional propulsion power [ 24 ] . Their streamlined hydrodynamic design ensures stable depth control and improved energy management during operation, making ROTVs highly suitable for long-duration seabed surveying missions. This successful application is demonstrated by commercial systems like EIVA’s ScanFish and ViperFish ROTVs [ 1 ] , which are widely deployed in the industry for seabed mapping. However, despite these advances, a gap remains in the development of highly maneuverable, compact ROTVs that are easy to deploy and capable of high-accuracy seabed mapping. Several technical challenges must be addressed to realize such systems. First, modeling ROTV dynamics is inherently complex due to their high-speed operation and the presence of wing-like control surfaces that introduce coupled hydrodynamic lift and drag forces. Accurate representation of these phenomena is essential for effective model-based control design. Recent work has explored hybrid approaches combining physics-based modeling with data-driven techniques to better capture nonlinearities and handle model uncertainties [ 3 , 5 , 13 , 23 ] . However, such data-driven methods require substantial quantities of experimental data for training, which presents practical challenges for novel vehicle designs. Furthermore, before deploying advanced learning-based control strategies or conducting hardware experiments, it is essential to establish detailed nominal models that enable realistic simulation and support initial controller development. These analytical models provide physical insight into system dynamics, facilitate systematic controller design using established control theory, and serve as a validated baseline for understanding vehicle behavior across various operating conditions. This work addresses these challenges through the following key contributions: (i) development of a detailed analytical model for SeaVis, a novel ROTV designed for high-accuracy seabed visualization and mapping; (ii) design of a model-based gain-scheduled linear quadratic regulator (LQR) for precise depth control; (iii) implementation of a gain-scheduled PID controller with comparative performance analysis; and (iv) open-source simulation framework to support further research. II SeaVis System Description II-A Mechanical Design The SeaVis ROTV employs an aircraft-inspired configuration optimized for hydrodynamic performance. The vehicle measures 122 cm in length with a wingspan of 120 cm, comprising a streamlined fuselage, two forward wings, and a horizontal tail stabilizer. Depth and attitude control are achieved through three independently actuated control surfaces: port flap, starboard flap, and elevator. Hydrodynamic forces and moments are modulated by varying the deflection angles of these control surfaces. This is achieved via dedicated servo actuators, as illustrated in Fig. 1 . Fig. 1: SeaVis ROTV with electronics canister mounted on the left wing, showing sonar and control surface actuators. II-B Sensor Suite The vehicle is equipped with a comprehensive sensor suite for state estimation and environmental monitoring. A forward-looking sonar is mounted beneath the towing attachment point for bathymetric mapping and obstacle detection. An inertial measurement unit (IMU) integrated within the navigator board provides acceleration and angular rate measurements for attitude estimation. Hall-effect sensors embedded within the hull monitor control surface deflection angles, while leak detection sensors within the pressure vessel ensure structural integrity during operation. The sensor and actuator layout is depicted in Fig. 2 . Fig. 2: Bottom view showing actuator placement, sensor locations, and electronics canister. II-C Communication The communication architecture enables bidirectional data exchange between the surface control station, embedded computing platform, sensors, and actuators, as shown in Fig. 3 . The embedded system consists of a Raspberry Pi single-board computer running the BlueOS operating system, which provides the software framework for vehicle control and sensor integration. The IMU is interfaced through the Navigator board, with data transmission implemented using the MAVLink communication protocol [ 8 , 16 ] to ensure reliable telemetry between the surface laptop and the Raspberry Pi. An Arduino Nano microcontroller serves as an analog-to-digital conversion interface for the Hall-effect sensors, since the Raspberry Pi lacks native analog input capabilities. Communication between the Raspberry Pi and Arduino Nano is established via the UART serial protocol, enabling low-latency transmission of sensor data for closed-loop control. Fig. 3: Communication architecture diagram. III SeaVis mathematical model To characterize the complex dynamics of the SeaVis ROTV, a detailed mathematical model is developed based on rigid body mechanics and hydrodynamic theory. The ROTV is modeled as a rigid body with 6 DOF: three translational motions, surge ( x ), sway ( y ), and heave ( z ), and three rotational motions, roll ( \phi ), pitch ( \theta ), and yaw ( \psi ). Given the operational constraints of the towed vehicle configuration, several simplifying assumptions are made. The ROTV is towed at a constant surge velocity with negligible acceleration in the longitudinal direction. The presence of vertical stabilizing winglets effectively constrains lateral motion, allowing sway and yaw velocities and accelerations to be neglected. Consequently, the model focuses on three active degrees of freedom: heave, roll, and pitch, which are the primary controllable motions for depth and attitude regulation. III-A Nonlinear ROTV Dynamics The equations of motion for the ROTV in the body-fixed coordinate frame are derived from Newton-Euler formulation, incorporating rigid body dynamics and hydrodynamic effects [ 10 ] : \begin{multlined}\boldsymbol{M}_{RB}\boldsymbol{\dot{\nu}}=-\boldsymbol{C}_{RB}(\boldsymbol{\nu})\boldsymbol{\nu}-\boldsymbol{M}_{A}\boldsymbol{\dot{\nu}}-\boldsymbol{C}_{A}(\boldsymbol{\nu})\boldsymbol{\nu}\\ -\boldsymbol{D}(\boldsymbol{\nu})\boldsymbol{\nu}-\boldsymbol{g(\eta)}+\boldsymbol{\tau},\end{multlined} (1) where \boldsymbol{M}_{RB}\in\mathbb{R}^{6\times 6} represents the rigid body inertia matrix, \boldsymbol{C}_{RB}(\boldsymbol{\nu})\in\mathbb{R}^{6\times 6} denotes the rigid body Coriolis and centripetal matrix, \boldsymbol{M}_{A}\in\mathbb{R}^{6\times 6} is the added mass matrix accounting for hydrodynamic inertia, \boldsymbol{C}_{A}(\boldsymbol{\nu})\in\mathbb{R}^{6\times 6} represents the Coriolis and centripetal effects from added mass, \boldsymbol{D}(\boldsymbol{\nu})\in\mathbb{R}^{6\times 6} is the hydrodynamic damping matrix (including linear and quadratic terms), \boldsymbol{g}(\boldsymbol{\eta})\in\mathbb{R}^{6} captures gravitational, buoyancy, and cable restoring forces, and \boldsymbol{\tau}\in\mathbb{R}^{6} represents the control forces and moments generated by the control surfaces [ 18 ] . The vehicle’s pose is described by the