Coral Reef Restoration Robots and Where ABB Fits

Coral reef restoration has mostly been done by hand. Divers cut fragments from healthy corals, grow them in nurseries and plant them back one at a time, and Coral Maker in Western Australia says most projects restore less than one hectare a year with these manual methods. That pace looks small next to the news of 22 October 2026, when the IUCN classed 96 percent of the world's assessed shallow-water reef ecosystems as Critically Endangered. This post follows the work from coral gardens tended by divers, through QUT's RangerBot and LarvalBot on the Great Barrier Reef, to McLaren's OSCAR assembly robot, and asks where ABB robots could fit.

A diver works on a coral tree nursery during coral planting for the Sandals Foundation, 2022. Photo: UVI / Wikimedia Commons (CC BY-SA 2.0)

How divers learned to garden coral by hand

Coral restoration began in the 1980s, and for most of its history the work was done with a diver's hands. Divers cut fragments from a healthy colony; up to 75 percent of a colony can be removed without harming its growth. The fragments then grow in underwater nurseries, strung on lines, fixed to frames or hung from coral trees anchored to the seabed. When they are big enough, divers carry them back to the reef and fix them in place one piece at a time, which is why a single project covers so little ground.

A diver after wiring a coral back into place at Mona Island, Puerto Rico, in 1997, watched by a wrasse. Photo: Erik Zobrist, NOAA Restoration Center via Wikimedia Commons (public domain)

Land-based nurseries came next. In 2006 David Vaughan developed microfragmentation, cutting corals into tiny pieces that grow on ceramic or cement plugs in tanks. Larval methods followed. Peter Harrison of Southern Cross University pioneered coral IVF: his teams catch spawn during the annual mass spawning, rear the larvae in floating pools and settle them on damaged reef. After small trials in the Philippines, his larval reseeding technique was piloted on the southern Great Barrier Reef in 2016 and 2017. By 2018 his team was catching spawn with large floating spawn catchers, a big step up in scale.

Coral fragments growing on grids in tanks at ORA, a coral aquaculture farm, 2011. Photo: Dustin Dorton, ORA / Wikimedia Commons (CC0)

Robots arrived through pest control. Crown-of-thorns starfish are blamed for an estimated 40 percent of the Great Barrier Reef's decline in coral cover, and in 2015 QUT roboticists built COTSbot, which found the starfish with a camera and image recognition and injected each one with bile salts. In 2016 QUT and the Great Barrier Reef Foundation won the Google Impact Challenge to develop it into RangerBot. Two years later Matthew Dunbabin's team turned RangerBot into LarvalBot and trialled it with Harrison on Vlasoff Reef near Cairns. Tethered for that first trial and programmed from an iPad, it carried about 100,000 coral larvae per mission and released them over damaged patches like an underwater crop duster, as Dunbabin put it.

A crown-of-thorns starfish at Lizard Island on the Great Barrier Reef, the coral eater COTSbot was built to hunt. Photo: Anne Hoggett / Wikimedia Commons (CC BY 3.0)

What coral restoration robots do on the reef today

The newest machine works on land. McLaren Racing's engineering group, McLaren Accelerator, built OSCAR, the Operational System for Coral Assembly and Restoration, with the Great Barrier Reef Foundation. Fitting each young coral by hand into a small ceramic cradle that protects it from predators used to take about 90 seconds per device. In OSCAR a robotic arm places coral fragments into cradles moving along a conveyor, and the cradles are stacked for transport "a bit like a kebab," says McLaren's sustainability director Kim Wilson. Assembly now takes under 10 seconds. Earlier in 2026 the machine was installed at the National Sea Simulator in Townsville and tested on living corals from the 2025 spawning.

A rare deep-sea Hawaiian coral grown on a rack in a tank at the Ānuenue Fisheries Research Center, 2017. Photo: Meg456 / Wikimedia Commons (CC BY-SA 4.0)

Corals spawn on just a few nights each year, so the schedule is tight. For the 2026 season, expected around 28 November, the target is about 90,000 cradles, each holding a week-old coral, to be shipped north and planted on chosen reefs. Other groups are automating other steps. Coral Maker uses technology from masonry manufacturing to mass-produce stone coral skeletons, and its biologists are growing hardier corals that robotic arms can propagate and transplant. Marine biologist Taryn Foster has been working out how at her family's masonry factory in Geraldton. QUT's RangerBot family now includes surface drones as well as underwater robots for 3D mapping, pest surveillance and water sampling.

A scuba diver inspects elkhorn coral on Horseshoe Reef in the Florida Keys National Marine Sanctuary. Photo: Ananda Ellis / NOAA via Wikimedia Commons (public domain)

We found no ABB robot in a reef project, although the jobs look a lot like work ABB arms already do. OSCAR's core task, picking small items and placing them into carriers on a moving belt, is what the IRB 360 FlexPicker was built for. ABB says it handles 1 to 8 kg, tracks fast conveyors and comes in a stainless version validated to IP69K for wash-down with hot, high-pressure water. ABB's conveyor tracking module links up to four robots, four belts and eight cameras. In the lab, where spawn is fertilised and larvae are reared, the GoFa CRB 15000 cobot can work beside people without fencing, and at SLAS 2025 ABB showed a GoFa handling pipettes and tending instruments.

A wash-down ABB FlexPicker being sprayed clean, as food lines require. Photo: ABB

Where coral reef restoration robots go next

Every group in this story talks about scale. "We need to plant a million corals on the reef per year to give it the best chance of survival," Wilson says, and Coral Maker counts tens of millions of hectares of reef at risk. The Great Barrier Reef alone is roughly the size of Italy, so planting has to target areas where natural reproduction and currents can spread the gains. McLaren has had inquiries from reef nations in the Pacific, the Caribbean and the Coral Triangle. In 2018 Dunbabin said future LarvalBot missions would run on their own and at a much larger scale.

Part of the Great Barrier Reef seen by the Copernicus Sentinel-2A satellite on 1 April 2017. Image: ESA, contains modified Copernicus Sentinel data / Wikimedia Commons (CC BY-SA 3.0 IGO)

No ABB coral project has been announced, so an ABB cell here is a possibility. A FlexPicker line could load cradles, a GoFa could move trays and pipettes in a coral hatchery, and the whole cell could be planned in RobotStudio, where ABB says offline programs match reality up to 99 percent. Non-specialists can already help with the data. Coral Restoration Foundation says its CeruleanAI makes reef monitoring accessible to everyone: divers upload their photos to a cloud service that stitches them into reef mosaics and helps measure coral colonies, and stitching starts at a few dollars.

ABB IRB 360 FlexPicker robots packing small items into cases at high speed. Photo: ABB

Machines can speed up assembly and seeding, but a coral still has to survive the heat once it is in the water. The IUCN names heat stress from climate change as the most consistent threat to reefs and calls for emissions cuts alongside protection and restoration. The equipment also has to travel. A machine built for a research aquarium in Townsville may not suit a remote Pacific island, which is why McLaren is studying a smaller, rugged OSCAR for 2027. Projects will also need staff who understand coral biology and can keep the robots running. ABB's approach for new users is simpler programming: a GoFa can be taught by moving the arm by hand, and set up with graphical apps on the FlexPendant.

An ABB GoFa CRB 15000 cobot being guided by hand at an assembly table. Photo: ABB

Innovation

AI solutions for effortless ABB robot control.

Automation

Robotics

ceojohntran@chatwithrobot.net

+84905311611

© 2025. All rights reserved.

qtran1215@gmail.com