Autonomous Undersea Robots Are Using Lasers to Make Offshore Operations Safer and More Efficient

Offshore equipment is built to endure, but nothing lasts forever in the extreme conditions of undersea operations. Even with protective coatings, all metal eventually corrodes in salt water, which is why offshore industries need to vigilantly inspect their assets on a regular basis.

Typically, inspecting underwater equipment for things like corrosion or other defects has been a painstaking job done by humans or manually controlled machines. But a new partnership between an American artificial intelligence (AI) company and a Canadian marine technology company could make deep-sea maintenance far easier to complete.  

On October 26, at the Minds + Machines conference in San Francisco, Boston’s Avitas Systems, a GE venture company, and St. John’s, Newfoundland-based Kraken Robotics Systems announced they were partnering to give companies in the offshore oil and gas sector access to autonomous robots equipped with high-precision sensors that feed data to machine-learning software. These digital tools will make inspections easier, safer and more cost-efficient than ever before.

A better way to dive

Underwater equipment inspection has, typically, been a dangerous job, with divers risking life-threatening injuries from underwater pressure. As offshore operations have moved into increasingly deeper waters, it’s becoming that much more difficult for humans to descend to the necessary depths to inspect equipment.


Read More Icon V3-01

Meet the Robots Getting Lots of Love in Digital Industries


Now, offshore industries primarily use Remote Operated Vehicles (ROVs) equipped with cameras, which require teams of professionals to carefully study video footage for signs of rust, fouling from marine life and other forms of equipment strain. It’s a difficult and time-consuming task that’s vulnerable to human error.

When these inspection teams do manage to spot a potential defect on the underwater asset, they need to take a measurement. To do that, the ROV driver has traditionally used the vehicle’s manipulator arm, for instance, to hold a ruler in front of the camera, and then steer the vehicle up to the object.

“Now you’re trying to control a 1,000-pound robot with a large hydraulic manipulator, trying to get close to this object to measure something, without potentially damaging it,” says David Shea, vice president of engineering at Kraken Robotics. It can take hours to perform this operation safely and accurately.

Seeing the light in the ocean’s depths

Clearly, this is a process that could be improved with digital technology, which is where Kraken Robotics and Avitas Systems come in. Rather than rely on problematic manual techniques, the companies have developed subsea robots equipped with lasers and robust AI technology.  

Here’s how it works. When an inspection is needed, an operator would send one of Kraken’s robot divers deep into the ocean. Its SeaVision sensor would then map the entire area within its field of vision and create a 3D model of the equipment that needs to be checked. “The pilot now has a 3D model that they can spin around and take any measurement they want, instead of trying to do it manually through the video,” says Shea. 


Read More Icon V3-01

Tough sensors keep industry operating in Canada’s harsh conditions


To create the model, Kraken’s SeaVision sensor emits visible light, in the form of three different coloured lasers: red, green, and blue. The three lasers help it recognize different types of corrosion, identify marine life like barnacles or mussels, and diagnose structural concerns. It can even read markings and labels on pipelines, something that other lasers can’t do.

Coupled with other hardware, like its complementary high-resolution sonar technology, which provides a much larger, but less fine-grained image of the underwater surroundings, Kraken Robotics’ sensor offers vastly superior inspection capabilities to offshore industries. But it’s only when coupled with autonomous inspection software by Avitas Systems that customers gain access to a truly comprehensive solution.

An AI inspector that only gets smarter

Avitas Systems specializes in bringing the power of artificial intelligence to industrial inspection. In this partnership, Avitas Systems synthesizes Kraken’s sensor data with other types of information, applies machine-learning algorithms to perform advanced analytics and provides results to customers in real-time over a web-based portal with user-friendly dashboards.

For example, instead of scheduling inspections on a regular basis, Avitas Systems’ analytics empowers operators to track their assets responsively, based on risk probabilities determined by the AI algorithms. It also has Automated Defect Recognition software that allows the AI to recognize and identify anomalies on the underwater assets. Because it is based on deep-learning models, the AI improves with experience. “It gets better as we get more data in the field,” explains Dominique Mann, communications manager for Avitas Systems.

Imagine an Autonomous Undersea Vehicle swimming smoothly at a depth of more than 1,000 metres. Using AI-powered navigation and sonar-derived volumetric imaging, it tracks the path of a pipeline. The sonar notices a potential defect, so the pilot decides to produce a laser scan of the area. An expert in a different part of the world can look at a colourized 3D model and not only identify the problem, but also precisely measure its scope and begin planning a solution.

This type of rapid problem-solving is only the beginning. As Kraken Robotics’ sensors gather more data, Avitas Systems’ machine-learning algorithms will only get smarter. With these end-to-end solutions, industries will gain much better awareness of undersea processes that affect the performance of their assets, and they can then put into place strategies to optimize their operations. The bottom of the ocean is still a harsh environment for machines, but with this new partnership, it’s about to get a whole lot friendlier.

Want more stories like this delivered to your inbox?


Related Stories