The use of robotics in shipwreck discovery


As underwater drone technology advances, its applications in shipwreck discovery, exploration, and analysis broaden. Already, these machines have been a part of numerous milestones. Will robots someday replace human divers?

How Robots Discover and Analyze Shipwrecks

Battles, storms, and icebergs have sunk ships for as long as they’ve been sailing the open seas. According to some estimates, there are roughly 3 million shipwrecks submerged in oceans, rivers and lakes worldwide. Many wreckages resting at the bottom of the ocean floor are well preserved, providing the perfect opportunity for research expeditions.

Underwater robots come in all shapes, sizes, and styles. For instance, some glide on fins propelled by currents, while others use thrusters. Depending on their intended use case, they may or may not have arms, solar panels, cameras, batteries, engines, sonar, or tethers.

Generally, any machine designed for underwater exploration will need circuitry to process, communications technology to keep operators in the loop, and a shell to protect internal components. Specific inventions use accessories like cameras, thermometers, or grabbers to interact with and record their surroundings.

Modern research teams often use robotics in shipwreck discovery, exploration, and analysis over human divers since digital technologies can quickly process information and accurately record data. Moreover, many unmanned machines are small enough to navigate without disturbing their surroundings — an important feature since many known wreckages are historical sites.

Why Teams Use Robots Instead of Human Divers

There are several reasons why research teams, scientists, and governments rely mainly on robotics instead of human divers to find and explore shipwrecks.

Robots Don’t Need Oxygen

Humans can only hold their breath for so long. Even with the right equipment, they still have to resurface — scuba tanks only last for about one hour, and factors like heightened activity quickly drain oxygen. On the other hand, robots can stay underwater for hours.

Dives Don’t Risk Human Lives

Although repairing underwater robots can be expensive, it’s relatively simple. In contrast, if a diver were to accidentally injure themselves underwater, they might not recover. Deploying machines instead of humans helps ensure every expedition member remains safe.

Robots Don’t Need Humans Anymore

An autonomous underwater vehicle (AUV) is a robot that functions without receiving constant operator input. These machines are often pre-programmed to complete basic tasks independently, enabling expedition members to focus on other high-priority tasks.

Engineering Is Getting Cheaper

As the field of robotics progresses, components and processes become more affordable. For instance, according to one estimate, the average cost of an industrial robot will be $10,856 by 2025, down from $68,659 in 2005 — a 145% cost reduction in only two decades.

Obstacles to Using Robots in the Deep Ocean

Robotics engineers have had to develop inventive workarounds for many of the issues the deep sea presents. Water pressure is one of the biggest obstacles they face. According to the National Oceanic and Atmospheric Administration, it reaches about 596 atmospheres at 19,685 feet below sea level — equivalent to the weight of an elephant sitting on a quarter.

For reference, the ocean is around 12,200 feet deep on average, meaning there are few places where research teams can deploy underwater robots without considering deep sea pressure — even fewer when accounting for shipwrecks’ locations.

Light is another one of the leading development hurdles. Once sunlight hits water, it can travel up to 3,280 feet in the right conditions. In typical conditions, light barely makes it past 655 feet. Since machines can’t navigate in pitch blackness, they need lots of lights. However, engineers must account for positioning because seawater is often full of particles.

If light reflects off tiny organic debris in the water — marine snow — it would blind the camera, making navigation impossible. Robotics engineers must carefully consider their bulbs’ placement and strength.

Communication is the last major hurdle. While it’s not as significant of an issue with manned underwater vehicles, their unmanned counterparts need a way to communicate with operators. Unfortunately, radio waves don’t penetrate seawater. Pre-programming, autonomy, and fiber optic cables are the most common workarounds.

Examples of Robots Exploring Shipwrecks

As robotics progressed, research teams discovered new techniques, components, and materials, helping them go deeper than ever before.

1. The Humanoid Diving Robot

Stanford University researchers recently developed OceanOneK, a humanoid diving robot — the first of its kind. It uses a haptic feedback system — a technology that stimulates touch, pressure, and vibrations — so its operators essentially feel what it feels while it’s underwater.

This life-sized machine uses eight multidirectional thrusters and a three-dimensional vision system to navigate. It needs special components — like microspheres and compressed oil — to keep it from collapsing in on itself under intense water pressure.

OceanOneK has two cameras for eyes and two manipulators for arms. It was designed to remain buoyant and carry out tasks while experiencing intense water pressure, making it ideal for exploring shipwrecks.

2. The Autonomous Mobile Robot

Dr. Robert Ballard —who discovered the Titanic — developed a new class of AUVs for marine archaeology. These remote-controlled mobile machines are small, so operating them costs a few thousand dollars per day instead of tens of thousands. Better yet, they can dive deeper, move faster, and remain underwater for days between charges.

3. The Tethered Antarctic Robot

The Falklands Maritime Heritage Trust, a London-based charitable organization, launched an expedition to find a long-lost shipwreck. The ship Endurance was lost to Antarctic waves in 1914 — and it wasn’t until a century later that an underwater robot, Sabertooth, uncovered it.

Sabertooth dove around 10,000 feet, using sonar blasts to pinpoint the ship’s location. Then, it used its built-in camera system to capture video. This hybrid machine switched between operator and autonomous control throughout the mission to improve precision.

The expedition’s project manager called it a once-in-a-lifetime find, saying Endurance looks like it sunk yesterday. Since Antarctica is so cold and inhospitable, no underwater critters or microbes were around to feed on the wood or trigger rot, preserving the wooden ship remarkably well.

Examples of Emerging Marine Archaeology Technology

Underwater robotics engineers can use examples of emerging marine archeology technology to refine designs and accelerate development.

1. Robot Dives to Never-Before-Explored Depths

Inspired by the deep-sea snailfish, a China-based research team developed an untethered soft robot. Their soft-bodied, battery-powered invention was small enough that a person could hold it with both of their hands.

Its small size helped it dive 36,000 feet below sea level — the deepest a robot has ever gone. It explored for about 45 minutes without being damaged by the extreme pressure or cold. 98% of the ocean floor is less than 19,685 feet deep.

Similar machines only ever reached up to 26,245 feet, making this expedition the first of its kind. Unfortunately, the research team had to watch their snailfish-like invention glide into the darkness, never to be recovered since it didn’t have a navigation system.

2. Artificial Intelligence Identifies Shipwrecks

Leila Character, a geographer, developed artificial intelligence (AI) to detect shipwrecks with 92% accuracy in partnership with the U.S. Navy’s Underwater Archaeology Branch. The algorithm trained on images of wreckages and seafloor terrain to consistently distinguish between the two. This development marks a significant milestone in marine archaeology technology.

3. A Deep Learning Model Identifies Shipwrecks

Coastal waters are bright, light, and shallow, so most shipwrecks are found there. Since countless wreckages happen in open waters, researchers invest in technologies to simplify deep-sea exploration.

One team’s deep learning model automatically detected shipwrecks with a 90% precision value, which consistently produced true positives. The article’s authors discuss its accuracy to stress the importance of this machine learning subset.

A standard AI could easily misinterpret images if real-world conditions don’t align with its training data. Water clarity, wave choppiness, sunlight penetration, and underwater obstacles could produce false positives. Deep learning models are far more resilient to those errors.

The Future of Robotics in Shipwreck Discovery

Robotics engineers have been refining AUVs for decades. Today, they’ve reached a point where they can send smart underwater drones to extreme depths without relying on human intervention. The world should expect to see novel inventions within a few years as more research teams explore and adapt this technology.