How Autonomous Underwater Vehicles Work Without GPS

Autonomous underwater vehicles, commonly called AUVs, are untethered robots that travel through the ocean without a pilot or a physical cable connecting them to a surface ship. They range from torpedo-shaped survey platforms the size of a kayak to deep-rated machines built to withstand the crushing pressures of the deepest ocean trenches. What makes them distinctive is their independence: once launched, an AUV follows a pre-programmed mission or makes real-time decisions on its own, collecting data, photographing the seafloor, or sampling water chemistry before returning to be recovered. That independence, though, comes with a cascade of engineering problems that do not exist on land or in the air, from navigating without GPS to communicating through a medium that swallows radio waves almost instantly.

Navigating Without a Signal From the Sky

On land and in the air, GPS is taken for granted. Underwater, it is useless. Radio signals attenuate so rapidly in seawater that a GPS receiver cannot get a fix even a few meters below the surface. AUVs operating in shallow water can work around this by periodically surfacing to grab a GPS position and then diving again, but for vehicles running missions tens or hundreds of meters deep, surfacing is impractical or dangerous, especially under ice or in busy shipping lanes.

The workhorse solution is inertial navigation: accelerometers and gyroscopes that track every movement of the vehicle in three dimensions. The catch is that inertial sensors drift over time, and without correction that drift compounds, meaning the vehicle’s estimate of its own position grows steadily less accurate the longer it stays submerged. To keep that error in check, most AUVs pair their inertial system with a Doppler Velocity Log, a sensor that bounces acoustic beams off the seafloor to measure the vehicle’s speed over the ground. The velocity data from the DVL constrains how fast the inertial drift accumulates, keeping the position estimate usable for hours-long missions.

1Information Processing in Agriculture. Integrated navigation for autonomous underwater vehicles in aquaculture: A review

Other aiding sources supplement these two core sensors. Acoustic transponder networks on the seabed can provide position fixes in a small area, much like an underwater version of GPS but with a range of only a few kilometers. Terrain-matching algorithms compare the bathymetry the vehicle sees in real time against pre-loaded seafloor maps. And in some research settings, vehicles use simultaneous localization and mapping techniques, building a map of the environment on the fly while using that map to track their own position. No single method solves the problem completely; most operational AUVs fuse several inputs together, accepting that their position estimate will always carry some uncertainty that grows with mission length.

The Energy Problem

Battery life is the single biggest constraint on what an AUV can accomplish. Most vehicles run on lithium-ion battery packs, and a typical survey AUV might operate for eight to twenty-four hours on a charge, depending on speed, sensor load, and depth. When the batteries run low, the vehicle must surface and be recovered, usually by a crewed ship, for recharging and data download. That recovery cycle is expensive, weather-dependent, and slow.

Researchers and industry have been working for years on underwater docking stations that would let an AUV recharge without ever being hauled out of the water. The idea is straightforward: the vehicle guides itself into a dock mounted on the seabed or suspended from a buoy, where it can transfer data and receive power. The engineering, however, is not simple. Physical connectors corrode quickly in seawater, so most current designs use wireless inductive charging, the same basic principle as a wireless phone charger. Wireless power transfer is more reliable in the marine environment because there are no exposed metal contacts to degrade, though it is less efficient than a direct plug.

2Ocean Engineering. A review of underwater docking and charging technology for autonomous vehicles

Solving the docking problem is widely considered the key to keeping AUVs underwater for weeks or months at a time, which would unlock applications like persistent environmental monitoring and long-duration seabed inspection. A handful of commercial docking systems have been deployed, but most are still in early stages or limited to demonstration projects. Frequent recovery by crewed ships remains the norm for the majority of operational AUVs today.

Gliders and Biomimetic Fins

Not every AUV uses a propeller. Underwater gliders move by changing their buoyancy: a small pump shifts fluid or oil to make the vehicle slightly heavier or lighter than the surrounding water, and wings convert the resulting vertical motion into forward travel. Because the energy cost of adjusting buoyancy is tiny compared to spinning a propeller, gliders can stay at sea for months on a single battery charge, covering thousands of kilometers at a leisurely pace of roughly half a knot. The trade-off is speed and maneuverability; gliders are excellent for wide-area ocean surveys but poorly suited to inspecting a specific structure or holding station in a current.

3Applied Mechanics and Materials. Efficiency of Buoyancy Force Generation in a Pump-Based Buoyancy Engine for a Small-Scale Underwater Glider

At the other end of the design spectrum, some researchers are building AUVs that swim like fish. Biomimetic propulsion uses fins modeled on the tails and body movements of marine animals. These designs offer better stability and maneuverability in rough or confined environments than a traditional propeller, and they tend to produce less noise and turbulence, which matters when the mission involves getting close to marine life without disturbing it.

4PubMed Central. Underwater Undulating Propulsion Biomimetic Robots: A Review

Recent work on fin design has shown impressive gains. One study tested a progression of fin types, from rigid plates to flexible single-joint fins to high-aspect-ratio two-joint fins inspired by tuna-like tails. The flexible single-joint fin tripled the average thrust and improved the thrust-to-power ratio fivefold compared to the rigid fin. Switching to the two-joint design tripled the thrust again and quadrupled the thrust-to-power ratio relative to the flexible fin. Those numbers suggest that as biomimetic designs mature, fish-like AUVs could become competitive with propeller-driven vehicles for certain low-speed missions.

5Ocean Engineering. Thrust and efficiency enhancement scheme of the fin propulsion of the biomimetic Autonomous Underwater Vehicle model in low-speed flow regime

Talking Through Water

Communicating with an AUV while it is submerged is one of the field’s most persistent headaches. Radio waves, which carry data effortlessly through air, are absorbed almost immediately in seawater. The standard alternative is acoustic communication, which works over distances of several kilometers but at painfully low data rates, often just a few kilobits per second. That is fast enough to send short status updates or simple commands, but far too slow to stream video or transfer large sonar datasets in real time.

Optical communication offers a partial solution. Blue-green laser or LED links can transmit data at much higher speeds than acoustics, with lower latency. The drawback is range: optical signals scatter and attenuate in water, especially in turbid coastal environments, limiting practical link distances to tens of meters rather than kilometers. This makes optical links useful for close-range tasks like docking data transfer or short-range vehicle-to-vehicle communication, but not for keeping in touch with a surface ship while the AUV roams freely at depth.

6PubMed Central. Underwater Optical Wireless Communications: Overview

In practice, most AUVs operate with minimal communication during a mission. They surface periodically to relay data via satellite or radio, or they store everything onboard and hand it over when recovered. This communication gap shapes mission design in fundamental ways: the vehicle has to be autonomous enough to handle unexpected situations on its own, because there is no way to phone home for instructions in real time from deep water.

Working as a Team

A single AUV can survey only so much ocean in a given time. The logical next step is to deploy multiple vehicles that coordinate with each other, dividing a large search area into manageable sections and adapting their plans as data comes in. This is harder than it sounds. The communication constraints described above mean the vehicles cannot maintain a constant chatter the way a fleet of aerial drones can. They have to operate semi-independently, checking in with each other or with a central planner only intermittently.

Recent research has applied reinforcement learning to the coordination problem. One approach uses a hierarchical framework where a higher-level planner divides the search region among heterogeneous vehicles with different capabilities, and each vehicle then optimizes its own path locally, accounting for factors like turning radius and sensor coverage. The goal is to maximize how thoroughly the team covers the area while minimizing redundant overlap.

7Engineering Applications of Artificial Intelligence. Dynamic collaborative search for multiple autonomous underwater vehicles based on hierarchical multi-agent reinforcement learning

Another line of work uses digital twins, virtual replicas of each physical AUV that are continuously updated with real sensor data. The digital twin predicts the local ocean current field and lets a swarm controller test maneuvers in simulation before issuing commands to the real vehicles. This virtual-real loop reduces the amount of acoustic communication the swarm needs, because much of the coordination happens in software rather than through data-heavy exchanges between vehicles.

8PubMed Central. Digital twin-driven swarm of autonomous underwater vehicles for marine exploration

Multi-vehicle operations are still largely experimental. Most real-world AUV deployments today involve one or two vehicles at a time. But the technology is moving quickly, and the payoff for getting swarm coordination right is enormous: it could turn weeks of serial surveying into days of parallel coverage.

Mapping Coral Reefs and Tracking Ecosystems

Marine ecology is one of the areas where AUVs are starting to change what scientists can actually measure. Coral reef monitoring has traditionally depended on divers conducting visual surveys along fixed transects, a method that is slow, physically limited to shallow depths, and difficult to scale. AUVs equipped with cameras and acoustic sensors can cover far more reef area, systematically and repeatably, while producing high-resolution photographic mosaics and three-dimensional reconstructions of reef structure.

One recent project used an AUV with both passive acoustic sensors and visual cameras to autonomously locate biological hotspots on a Caribbean reef. The vehicle detected sounds associated with fish and invertebrate activity, steered toward the noisiest areas, and then photographed them in detail. It identified a concentration of marine life around a large pillar coral, and the combined acoustic and visual data cross-validated the finding. That kind of self-directed, multimodal survey would be extremely difficult to replicate with divers or towed camera systems.

9Science Robotics. Autonomous seeking and mapping coral reef biodiversity hotspots with a multimodal AUV

Beyond hotspot mapping, AUVs are being explored for species identification, disease detection, early warning of coral bleaching, and 3D structural mapping of entire reef systems.

10Ecological Informatics. Autonomous robotic systems for coral reef monitoring: Review and open research issues

Other work has shown that AUV-derived photogrammetric maps of coral reefs correlate well with rugosity, an established proxy for reef biodiversity and abundance, suggesting that rapid AUV surveys could substitute for some of the painstaking manual assessments currently used to track reef health over time.

11arXiv. Robot Goes Fishing: Rapid, High-Resolution Biological Hotspot Mapping in Coral Reefs with Vision-Guided Autonomous Underwater Vehicles

Under Ice and Into the Abyss

Some of the most dramatic AUV missions happen in places where humans simply cannot go. Polar operations are a prime example. AUVs have been running under sea ice since the mid-1990s, and missions of several hundred kilometers under ice cover have been achieved. The challenge is not just the cold; it is the fact that if something goes wrong, the vehicle cannot surface. The ice overhead is a ceiling with no exits. Recovery of a failed AUV under pack ice is, in most cases, impossible.

12Marine Technology Society Journal. AUVs Under Ice: A Four-Decade Retrospective on Strategy and Risk Through the Autosub Looking Glass

Despite decades of progress, there remains a large gap between what current AUVs achieve under ice and what glaciologists actually need. Missions of a thousand kilometers or more would be required to conduct large-scale surveys of sea ice extent, run multiple parallel transects beneath West Antarctic ice shelves, or complete sections under the vast Filchner-Ronne Ice Shelf. That gap is driven partly by battery limits and partly by the extreme conservatism that under-ice risk demands: operators choose shorter missions to reduce the chance of losing an expensive vehicle with no way to get it back.

At the other extreme of depth, engineers are working on AUVs rated for the hadal zone, the ocean trenches that reach down to roughly 11,000 meters. The pressures at those depths are staggering, about 1,100 times atmospheric pressure at the surface. Standard materials and electronics cannot survive. One engineering study evaluated spherical titanium alloy pressure vessels for a hydrogen-fueled AUV and showed that a Ti-10-2-3 alloy design could withstand 11,000-meter pressures with an acceptable safety margin, pointing toward a class of vehicles that could eventually conduct autonomous surveys of the deepest places on Earth.

13Results in Engineering. Numerical evaluation of type I pressure vessels for ultra-deep ocean trench exploration

Archaeology on the Seafloor

Shipwrecks and submerged cultural sites are scattered across every ocean, and many of them lie too deep for divers to reach. AUVs have opened up deep-sea archaeology by providing a platform for high-resolution acoustic and optical surveys that can document sites in detail without physically disturbing them. In the Black Sea, a major project used deep-sea AUVs to survey and create three-dimensional photogrammetric models of shipwrecks spanning roughly 2,500 years of seafaring history. One wreck from the early fourth century BC was documented with enough fidelity that archaeologists could interpret construction details and cargo from the 3D model, both during the expedition and in post-cruise analysis.

14Deep Sea Research Part I: Oceanographic Research Papers. Deep sea archaeological survey in the Black Sea – Robotic documentation of 2,500 years of human seafaring

This kind of non-invasive, repeatable documentation is valuable not just for discovery but for long-term preservation. A photogrammetric model created by an AUV can be revisited years later for new analysis without returning to the site, and repeat surveys can detect changes caused by currents, biological growth, or unauthorized salvage. For sites at depths of a thousand meters or more, AUVs are effectively the only practical survey tool.

Deep-Sea Mining Surveys

One of the more contentious emerging applications for AUVs is in deep-sea mineral prospecting. Polymetallic nodules, potato-sized lumps of manganese, nickel, cobalt, and other metals, carpet vast areas of the abyssal seafloor. Mining companies and researchers use AUVs to map where nodules are densest and to assess how mining disturbances might spread through the environment.

AUV-mounted cameras and high-resolution bathymetric sonar can resolve nodule coverage at a spatial scale of just a few meters, far finer than what ship-based sonar can achieve. One study in a nodule field found that increased nodule coverage correlated with slopes steeper than about 1.8 degrees and concave terrain features, details that would be invisible in coarser survey data. The same study used AUV imagery to track the resettlement of sediment kicked up by a disturbance experiment. Within 16 hours of the disturbance, suspended sediment had blanketed nearby nodules, but the visible impact was confined to about 100 meters downstream of the disturbance track.

15Biogeosciences. Understanding Mn-nodule distribution and evaluation of related deep-sea mining impacts using AUV-based hydroacoustic and optical data

That 100-meter figure is both reassuring and alarming, depending on your perspective. On one hand, it suggests that sediment plumes from small-scale disturbances may not travel as far as worst-case models predict. On the other hand, commercial mining would involve disturbances orders of magnitude larger and more continuous than a single sampling event, and cumulative effects over weeks or months of operations are not captured by a short experiment. AUVs will play a central role in monitoring those effects if commercial deep-sea mining ever moves forward at scale, but they are also the tools generating the baseline data against which environmental damage would be measured.

Legal Gray Zones

As AUVs become more capable and more widely deployed, they are running into legal frameworks that were written for crewed ships. International maritime law governs who can operate where, what rights a vessel has in different zones of the ocean, and how military and civilian assets interact. AUVs fit awkwardly into these categories. They are not crewed vessels, so it is unclear whether they enjoy the right of innocent passage through another country’s territorial waters. They are not traditional submarines, so existing rules about submarine navigation may not apply directly. And if a military operates an AUV, the question of whether it qualifies for sovereign immunity, the legal shield that protects warships from foreign jurisdiction, is genuinely unsettled.

16Journal of Conflict and Security Law. The evolving seascape of naval warfare: unmanned underwater vehicles and the challenges for international law

These are not abstract concerns. Countries have already seized or interfered with underwater vehicles operated by rival nations. As AUVs become smaller, cheaper, and more autonomous, the frequency of encounters in disputed waters is likely to increase. Whether the legal framework adapts through new treaties, case-by-case precedent, or creative reinterpretation of existing law is one of the open questions that will shape how freely AUVs can operate in the coming decades. The technology, as usual, is moving faster than the rules that govern it.