Subsea robotics: what ROV and AUV inspection delivers today

Underwater inspection, maintenance and repair on civil offshore assets is carried out by two vehicle families that are frequently lumped together and should not be. A remotely operated vehicle (ROV) hangs on an umbilical. The tether supplies power and carries control signals and live video, which gives the system effectively unlimited endurance, a human in the loop at all times and the ability to apply force. An autonomous underwater vehicle (AUV) carries its own energy, follows a planned track and surfaces with the data. It covers ground efficiently, but it does not intervene.

Inside the ROV family, the split that decides a budget is between observation class and work class. Observation ROVs are compact, electrically driven, deployable from a workboat or a crew transfer vessel and carry cameras, lights, sonar and contact probes. Work-class systems are hydraulically powered, weigh from several hundred kilograms upwards, carry manipulators and normally require a dedicated vessel with a launch and recovery system and station keeping. The gap between the two is not incremental, it is structural: it drives vessel class, crew size, mobilisation time and weather sensitivity.

• Observation-class ROV: visual and acoustic inspection, cathodic protection readings, close-up work on foundations, scour protection and cable entry points.
• Work-class ROV: manipulator tasks, valve actuation, debris recovery, tooling deployment, cleaning of weld areas ahead of a visual inspection.
• Survey AUV: corridor and area coverage along cable routes and pipelines, repeatable tracks, bathymetry and sonar imagery over kilometres rather than metres.

Vendors in this market cover very different parts of that spectrum. Saab Seaeye offers both observation and work-class vehicles, and its Falcon belongs to the observation class rather than the work class, a distinction that specification documents get wrong with some regularity. Blueye in Norway and Deep Trekker build compact inspection ROVs. In the German-speaking market, Tethys Robotics in Zurich, Hydromea in Renens, iSiTEC in Bremerhaven and SubCtech in Kiel as a component supplier are active. These are market examples for orientation, not partners of werob. A structured overview of vehicle categories is collected in the robot catalogue.

Specification discussions about subsea robotics tend to start at the deep end, with 3,000 m ratings and hydraulic manipulators. For the German offshore wind fleet that is the wrong starting point. According to Deutsche WindGuard, as of 30 June 2026 the fleet has a mean water depth of 31 m and a mean distance to shore of 73 km. Since 2019 every German project has been built on monopile foundations.

The fleet behind that average is not small. As of the same date, 1,764 turbines with a combined 10,818 MW were feeding into the grid across 31 fully operational wind farms. The North Sea accounts for 8,987 MW across 1,455 turbines, the Baltic Sea for 1,831 MW across 309 turbines. In total 1,786 turbines stand off the German coast, commissioning is still pending for 22 of them, and 65 foundations have been set without a turbine on top. In the first half of 2026, 84 turbines with 1.1 GW fed in for the first time.

Put the depth figure next to commercially available hardware and the consequence is uncomfortable for anyone selling heavy systems into this segment. A common observation ROV is rated for 305 m. That is roughly ten times the mean water depth of the entire German fleet. The routine inspection case, meaning foundation and scour protection surveys, cathodic protection readings, cable entry point checks and general condition documentation, is covered by an electrically driven observation vehicle deployed from a small vessel. It needs no work-class system and no DP2 support vessel. The limiting factors offshore Germany are the 73 km transit, the weather window and the turbidity, not the depth.

That does not make the market trivial. Uniform depth and uniform foundation type mean uniform inspection procedures, and uniform procedures are exactly what makes repeatable robotic survey worthwhile over hundreds of structures. The subsea vertical is therefore less about extreme engineering and more about disciplined repetition. On the political framing it is worth being precise: the expansion targets of 30, 40 and 70 GW for 2030, 2035 and 2045 are current German law, but WindGuard itself considers the 2030 target no longer achievable in time. Planning a service business on the legal target rather than on the installed base would be a mistake.

The first thing to accept about underwater work is that there is no GPS. Radio frequencies do not propagate usefully in sea water, so position has to be built acoustically and inertially. In practice that means an ultra-short baseline (USBL) or long baseline (LBL) acoustic system, combined with an inertial navigation system and a Doppler velocity log (DVL) that measures speed over ground against the seabed. The quality of a survey is largely the quality of that navigation solution, because a defect without a reliable position cannot be relocated on the next campaign.

On top of the navigation layer sit the payloads. Acoustics carry the wide-area work, optics carry the detail, and each has a hard physical limit.

• Multibeam echosounder: bathymetry of the seabed around monopiles and along routes, the standard tool for scour development and for detecting exposure or free spans.
• Side-scan sonar: wide acoustic imagery to locate debris, trawl marks, boulders and changes in the seabed.
• Sub-bottom profiler: penetration below the seabed surface, used where the question is depth of cover rather than surface shape.
• High-definition optical imaging with dedicated lighting: corrosion, coating condition, marine growth, weld areas, anode consumption. Usable only within the visibility the water allows on the day.
• Laser profiling and subsea LiDAR: dimensional detail in low turbidity, for quantifying deformation, denting or gaps rather than merely showing them.
• Cathodic protection probes: contact or proximity measurement of the electrochemical potential against a reference electrode, verifying that the steel is still protected.

The practical constraints in this business are turbidity, current and weather windows, not water depth. A vehicle rated for far greater depths does not see any better in a suspended sediment plume after a storm, and it does not launch in a sea state its host vessel cannot work in. Campaign planning that ignores this produces schedules that look efficient on paper and are not achievable in the North Sea autumn.

An AUV earns its keep where the task is coverage: kilometres of cable corridor, pipeline route, or an area survey that has to be repeated in a comparable way every few years. The vehicle flies a planned track without a pilot, without a tether and without a vessel holding station above it, which removes the strongest cost driver of tethered work. What it does not do is react to what it finds. An AUV brings back the data set, and the anomalies inside it are worked afterwards, in many cases by an ROV on a targeted second visit.

Three vehicles illustrate how wide the class is. EvoLogics in Berlin builds the Quadroin, a micro AUV rated for 150 m with four thrusters, a payload capacity of up to 3 kg and an endurance of up to 12 hours at 1.5 m/s, navigating with a Nortek Nucleus1000 combining DVL and AHRS. It is in series production. Exail in France offers the A18-M, a survey AUV for 3 to 300 m, from 442 kg upwards, with up to 24 hours of endurance at 3 knots, an INS, DVL and GPS navigation suite and synthetic aperture sonar, described by the manufacturer as sea proven. Kongsberg in Norway sits at the other end with HUGIN Superior, rated for 6,000 m at roughly 2,200 kg and 62 to 76 hours of endurance.

Two points of housekeeping matter when reading this market. iXblue and ECA Group have traded as Exail since October 2022 and should not be listed as separate suppliers. Ocean Infinity, frequently named in the same breath, does not sell hardware at all: it is a survey contractor and is itself a customer of manufacturers such as Kongsberg. For a cable or pipeline programme the practical question is rarely which AUV is best in the abstract, but which endurance, sensor suite and data format fit the inspection interval that has been agreed. That question is taken further in the article on condition monitoring of subsea cables and pipelines.

Observation vehicles document. When something has to be moved, opened, cleaned or measured by contact, the requirement changes. Work-class ROVs exist for manipulator tasks: operating subsea valves, recovering dropped objects and debris, deploying and retrieving tooling, and removing dense marine growth from weld areas so that a visual inspection can see anything at all. The last of these needs care. Aggressive removal of an established fouling layer can damage the coating underneath, which is the same trade-off that governs hull work and is discussed in the article on robotic hull cleaning.

Cathodic protection measurement sits in between. Verifying that a foundation or a pipeline is still protected requires the probe to reach the steel or the anode and take a potential reading against a reference electrode. That is a contact task with a positioning requirement, but not a heavy one, and observation-class vehicles with the right probe fitted do it routinely.

The category that attracts the most speculation is the resident system: a vehicle that stays subsea in a docking station and is called on demand rather than mobilised with a ship. Here the market reality is narrower than the marketing. Saipem's Hydrone-R is genuinely commercially deployed, working for Equinor at the Njord field since June 2023. Saipem's FlatFish has German roots, having been developed at DFKI in Bremen between 2013 and 2017. Eelume, by contrast, does not sell a resident system: its S-Series is a portable vehicle without a docking infrastructure, and the resident concept remains a research and development topic. Modus Subsea Services, once a reference name for resident inspection, became insolvent in 2025. Anyone building a business case on permanent subsea residency in German waters should treat it as an emerging capability, not a procurable standard.

For inspection work on ships and floating structures, robotic survey touches classification rules, and this is where tender documents most often go wrong. The frequently written phrase "DNV IWS" describes something that does not exist. DNV does not operate an IWS notation. DNV has BIS and IW for bottom survey afloat, regulated in DNV-RU-SHIP Pt.7 Ch.1 Sec.5. The *IWS notation belongs to Lloyd's Register and is set out in the LR Rules, Part 1, Ch. 3, Sec. 4. A second recurring error is the citation of DNV-CP-0484 as a remote inspection document. Its actual title is "Approval of service supplier scheme".

The more important insight is structural. Buying a capable vehicle does not open this market. Access runs through recognition as a service supplier under IACS UR Z17, which is an approval of the organisation, its procedures, its personnel and its quality system, not of the hardware. Section 3 of UR Z17 covers in-water survey in lieu of docking survey carried out by diver or ROV, and Section 16 covers providers of remote inspection techniques. That is the entry barrier, and it is an organisational one.

The same logic applies on the asset side. Recurring examinations and class surveys remain with the classification society and the recognised expert. A robot supplies the record, not the certificate. Anyone promising that a survey vehicle replaces an inspection regime is describing something that does not exist in the applicable rules.

On the diving side, commercial diving work in Germany is governed by DGUV Vorschrift 40. DGUV Regel 101-023, often cited alongside it, is titled "Forschungstauchen" and covers scientific diving rather than commercial diving operations. No published DGUV statistic was found that establishes diver risk as a quantified driver for robotics adoption, so the argument for removing personnel from the water is worth making on its own terms rather than dressing it in a citation that does not exist.

werob is a manufacturer-independent systems integrator for service robotics and a brand of CITO GmbH in Hamburg. It is not a manufacturer and not a consultancy. In a subsea context that means the work starts with the inspection requirement rather than with a vehicle: which structures, at what interval, in what water depth, with what visibility, producing what evidence for whom.

From there the questions are practical. Which vehicle classes actually satisfy that requirement, and which are oversized for a 31 m fleet. Which sensor payload produces data that is comparable between campaigns rather than merely impressive once. How the data leaves the vessel and enters the operator's existing maintenance and asset systems, in which formats, with which positional reference. Which party holds the approvals under which the operation is carried out, since class surveys, expert assessments and, in aviation-supported work, operating permits stay with the organisations that hold them.

• Translate the inspection obligation into a technical specification that a supplier can quote against.
• Compare vehicle classes and payloads across manufacturers rather than inside one product line.
• Check interfaces, data formats and positional references before procurement, not after the first campaign.
• Plan the operation with the organisations that hold the relevant approvals and qualifications.
• Set the inspection interval and the baseline so that year-on-year comparison is technically possible.

The commercial case for repeatable robotic survey is built on avoided vessel time, avoided diving operations and earlier detection of scour or coating loss, and it is specific to a portfolio rather than universal. The ROI calculator is the starting point for putting numbers on a concrete asset base.