1. Background and Motivation
Subsea intervention today still relies heavily on remotely operated vehicles (ROVs) and, in some cases, human divers. While ROVs have evolved significantly in sensing, navigation, and dexterity, their fundamental limitation remains unchanged: they operate as free-floating systems. In practice, this becomes a problem the moment physical interaction is required. Position hold mode of ROV in a current while trying to turn a valve, maintain tool contact, or inspect a specific point on a pipeline is far from trivial. Even with skilled pilots, there is always a compromise between stability, visibility, and control effort. Thrusters compensate continuously, sediment gets disturbed, visibility drops, and precision suffers. This issue becomes more pronounced in:
Soft or silty seabeds
High-current regions
Confined or cluttered subsea environments
From an operational standpoint, a significant portion of intervention time is spent not on the task itself, but on trying to maintain a stable working condition. This conceptualization starts from a simple question: Instead of forcing a hovering system to behave like a stable manipulator, can we separate these roles?
2. Concept Overview
The proposed system introduces a separation between positioning and interaction. A work-class ROV or Human Occupied Vehicle (HOV) is used as a hovering support platform, maintaining a controlled stand-off distance above the seabed typically in the range of a few meters to avoid sediment disturbance and preserves visibility, which is often critical during inspection or intervention. An Amphibious Quadruped Robot (AQR) is deployed from the ROV from this stable hovering position. The AQR is designed to:
Descend gently to the seabed using a tether
Establish a stable base using legged contact
Perform inspection or manipulation tasks directly on the asset
The ROV remains above, acting as a deployment system, a communication relay and a supervisory observation platform. In effect, the system shifts from a single vehicle doing everything to a cooperative system, where each component operates within its strengths.
2.1. Deployment and Transition to Seabed
The ROV first reaches the worksite and stabilizes at a chosen altitude. This height is selected based on visibility, current conditions, and seabed characteristics. Importantly, the ROV does not attempt to descend further to avoid disturbing the seabed. The AQR is then released using a controlled mechanism. The descent is not a free drop it is guided or damped to ensure soft landing, without sediment plume while maintaining ROV’s orientation. Once the AQR reaches the seabed, it transitions into a stable configuration. Unlike wheeled systems, the quadruped design allows it to adapt to uneven terrain, distribute load across multiple contact points and maintain stability even on soft ground. This transition from water column to seabed is the key operational shift in the system.
2.2. Seabed Operation and Interaction
Once grounded, the AQR becomes the primary agent of intervention using the seabed as a reaction base. Legged locomotion allows the robot to move along subsea assets such as pipelines or cables without requiring continuous thrust. It can step over obstacles, adjust its posture, and reposition itself with relatively low energy expenditure. Thus, it enables stable visual inspection, consistent tool contact and controlled application of force. For manipulation tasks, such as turning a valve, positioning a sensor, or performing localized maintenance, the robot is no longer limited by drift.
<figure><img title=”Concept image AQR with AUV_final.png” src=”https://cdn.filestackcontent.com/V0QlhLEiQQme82cdwbWA” alt=”Hybrid Subsea Intervention for Contact-Stable Tasks Using Hovering ROVs and Seabed-Deployed Legged Robots (Image generated using AI-based visualization tools based on the authors’ concept.)” width=”1280″ height=”853″ />
<figcaption>Hybrid Subsea Intervention for Contact-Stable Tasks Using Hovering ROVs and Seabed-Deployed Legged Robots (Image generated using AI-based visualization tools based on the authors’ concept.)</figcaption>
</figure>
3. System Architecture
The system follows a layered structure, aligned with existing subsea operations. At the surface, a surface vessel houses the operator console, where mission planning, monitoring, and control are carried out. Communication and power are transmitted to the ROV via a tether. The ROV acts as an intermediate layer to handle navigation and positioning, visual feedback and deployment and retrieval of the AQR. The ROV operates in a controlled hovering mode, using thrusters primarily for station-keeping rather than interaction. The AQR forms the seabed layer with a legged locomotion system using a manipulation or tooling interface, a sensor suite (vision, sonar, force feedback) and an onboard controller. The communication between the ROV and AQR can be handled through tethered links or short-range communication, depending on the design.
4. Mission Workflow
A typical mission follows a repeatable sequence:
The ROV is deployed and navigates to the target site.
It stabilizes at a safe altitude above the seabed.
The AQR is deployed and descends to the seabed.
The AQR establishes a stable posture and begins operation.
Inspection or intervention tasks are carried out.
The AQR is retrieved, and the system returns to the surface.