IIT Madras builds SEANAV, a digital twin to test and train self-navigating ships

Researchers at the Department of Ocean Engineering of the Indian Institute of Technology Madras have built SEANAV, a real-time digital twin in which autonomous navigation systems for ships can be tested, checked and trained before the ships go to sea.

SEANAV runs on the Unity simulation engine and recreates waves, ship traffic and bad weather, so that vessels can practise avoiding collisions and following routes. Several ships can sail crowded waters in it at the same time, and crews can train in it through virtual reality. The institute says thousands of hours of simulation cut the cost of testing.

The researchers also made DetectShips, a set of 14,000 real ship images, to train ship recognition with stereo cameras, and StereoLiD, which adds lidar data to estimate distances more precisely. The work was published in the ASME Journal of Offshore Mechanics and Arctic Engineering.