ROBOTIC.INDUSTRIES

Field and Service Robotics

ROV vs AUV: Which Underwater Robot for Which Job

An ROV is tethered, powered from the surface and can intervene. An AUV runs free for hours and only surveys. The tether decides which one your job needs.

Work-class remotely operated underwater vehicle on the deck of a support vessel
Work-class remotely operated underwater vehicle on the deck of a support vessel

The tether decides it. An ROV takes power and commands down a cable, so it has unlimited endurance, full video bandwidth and an arm that can turn a valve, and it needs a vessel holding station above it. An AUV carries its own battery for 8 to 60 hours, covers ground far faster, and can only look. Intervention needs a tether; survey does not.

8 to 60 hAUV endurance on internal power
unlimitedROV endurance, powered from the surface
6,000 mworking depth of a work-class ROV
1thing an AUV cannot do: touch anything

The two families

ROV against AUV
PropertyROVAUV
PowerFrom the surface, unlimitedInternal battery, 8 to 60 h
ControlLive pilot, full videoPre-programmed, acoustic link only
ManipulationYes, one or two armsNo
Survey speed0.5 to 1.5 knots3 to 5 knots
Support vesselRequired, holding stationLaunch and recovery only
Depth rating, work classto 6,000 mto 6,000 m
Data returnLiveOn recovery, or summaries acoustically
Day rate including vesselvery highhigh, shorter vessel time
Typical useIntervention, close inspection, constructionPipeline and seabed survey, hydrography
The vessel is the cost, not the robot. A work-class ROV spread needs a support vessel with dynamic positioning holding station for the whole dive, and vessel day rates dwarf the robot. That is why an AUV that surveys four times faster and lets the ship transit between deployments changes project economics even though the vehicle itself is not cheaper.

The categories in between

  1. Observation-class ROV. A small tethered vehicle with cameras and no manipulator, launched from a small boat or a quayside. Cheap, useful for hull and harbour inspection, and the entry point for most operators.
  2. Hybrid ROV. Operates free-swimming for survey, then connects to a light fibre tether for intervention. It buys flexibility at the cost of complexity.
  3. Resident systems. A vehicle that lives on the seabed in a docking station, charging and uploading through it, deployed on demand without a vessel at all. This is the structural change in the industry, because it removes the vessel from routine inspection entirely.
  4. Underwater crawler. Drives on the structure or seabed rather than swimming, which gives stability for contact measurement in current.

Resident systems deserve the attention. If the vessel is the dominant cost, a vehicle that never needs one converts inspection from a mobilised campaign into a scheduled routine, and that is a different business rather than a cheaper one.

Sensor payloads and what they cost in endurance

Typical underwater sensor package
SensorMeasuresPowerRange
Multibeam echosounderSeabed bathymetry30 to 90 Wup to 200 m
Side-scan sonarSeabed imagery15 to 50 W50 to 200 m per side
Sub-bottom profilerLayers below the seabed20 to 80 W5 to 50 m penetration
Doppler velocity logSpeed over ground for navigation10 to 30 Wup to 200 m altitude
Camera and lightingVisual detail50 to 400 W1 to 10 m
Ultrasonic thickness probeSteel wall thickness5 to 20 Wcontact

The camera row explains why free-swimming vehicles survey with sonar rather than with pictures. Lighting is the single largest power draw on any underwater platform, and turning it on can halve the endurance of a battery-powered vehicle, which is exactly the resource an AUV cannot spare.

What makes underwater hard

  • No radio. Water blocks the frequencies used for radio and satellite navigation, so there is no GPS below the surface and no high-bandwidth wireless link. Acoustic communication offers low data rates with seconds of latency.
  • Navigation drifts. Without an absolute reference, position comes from inertial sensors and a Doppler velocity log, and error accumulates over the dive unless corrected against acoustic beacons or seabed features.
  • Visibility is often metres or less. Sonar does the work that cameras do on land, and interpreting sonar is a skill.
  • Pressure. Every 10 m adds roughly one atmosphere, so a 3,000 m vehicle sits under about 300 bar and every penetration and housing is a design problem.
  • Current. Station keeping in a 2 knot current consumes most of the thrust a small vehicle has, which is exactly when the inspection is needed.

Choosing in four questions

The decision
QuestionIf yes
Does anything need to be touched, turned or cut?ROV, work class
Is the job covering distance rather than examining a point?AUV
Does the data need to be seen live?ROV
Is this a repeating routine on a fixed asset?Resident system
Is it shallow, short and visual?Observation-class ROV
Is contact measurement needed in current?Crawler

Frequently asked questions

What is the difference between an ROV and an AUV?

An ROV is tethered, powered from the surface, piloted live and can carry manipulators. An AUV runs free on internal battery for 8 to 60 hours, follows a pre-programmed mission, surveys three to four times faster and cannot touch anything.

Which is cheaper?

Neither vehicle dominates on price; the support vessel does. An ROV needs a vessel holding station for the whole dive, while an AUV lets the ship transit between deployments, and that difference in vessel time usually decides project cost.

Why is there no GPS underwater?

Water absorbs the radio frequencies satellite navigation uses. Position comes from inertial sensors with a Doppler velocity log, corrected against acoustic beacons or recognised seabed features, and it drifts without those corrections.

What is a resident system?

A vehicle that lives on the seabed in a docking station, charging and transferring data through it, deployed on demand without a support vessel. It converts inspection from a mobilised campaign into a scheduled routine.

How deep do these vehicles work?

Work-class ROVs and large AUVs are commonly rated to 6,000 m, which covers the great majority of the ocean floor. At 3,000 m the vehicle sits under roughly 300 bar, so every housing and penetration is a pressure design problem.

Sources

  1. World Robotics 2025 report, service robotsInternational Federation of Robotics, inspection and maintenance robot segment
  2. Robotics at NISTNational Institute of Standards and Technology, robot performance and test methods
  3. arXiv robotics preprints, underwater robotics and navigationPrimary literature on acoustic navigation and autonomous underwater vehicles