Cargo ship sailing at sea (representational image)Getty Images
Scientists at Texas A&M University have developed an AI system that will help ship captains avoid deadly collisions at sea.
The system, called ‘Ship collision avoidance of Machine learning And Radar Technology for Stationary Entities and Avoidance’, or SMART-SEA for short, will reduce ship operators’ reliance on captain experience.
By providing “human-in-the-loop” advisors with real-time instructions, the system will combine human expertise with AI precision to make maritime navigation easier and safer for those involved.
Mitigating human error in maritime operations
The team, led by Dr. Mirjam Fürth, an assistant professor of ocean engineering, developed their system under a one-year contract by the US Department of the Interior (DOI) and the US Department of Energy (DOE) through the Ocean Energy Safety Institute (OESI). The purpose of SMART-SEA is to reduce collisions caused by human error in maritime navigation. Unlike fully autonomous navigation systems, it will require a “human-in-the-loop” advisor.
As Texas A&M University points out in a press statement, collisions between sea vessels and stationary structures such as oil rigs are becoming “increasingly common.”
“Many of these collisions are caused by human error,” Fürth added. “By using data to provide seafarers with real-time instructions, we hope to reduce marine collisions.”
Due to this increase in sea collisions, the Texas A&M scientists set out to build an AI-powered copilot, despite the fact that there are so many more variables at sea than on the road—and we can barely get an autopilot for working for road vehicles.
Large ships have so much forward momentum that their stopping distances are often measured in miles rather than meters.
They also turn incredibly gradually, meaning that any collision system has to factor in this slow, lumbering movement. Add ocean currents, wind, sensor time lags, and a host of other factors to the mix, and it’s clear that developing autonomous vessels is far from an easy problem.
Combining radar data with advanced machine learning
The A&M team explained that SMART-SEA combines raw radar imaging data with advanced machine learning. The radar imaging allows it to detect moving objects, even in adverse weather conditions. The machine learning algorithm, meanwhile, classifies and identifies stationary objects that could lead to collisions.
The AI system uses “state-of-the-art computational fluid dynamics models and machine learning trained on past vessel motions,” the team said in their statement.
The logic system that underpins SMART-SEA is based on seafarer experience compiled via focus groups at the Texas A&M Galveston faculty. According to a New Atlas report, it works with a Modified Velocity Obstacle (VO) algorithm combined with an Asymmetric Grey Cloud (AGC) model to assess risks and avoid collisions. It does all of this while complying with International Regulations for Preventing Collisions at Sea (COLREGs).
