The search that ensued after the Titan submersible lost contact with its support ship in June 2023 covered 13,000 square kilometers of the North Atlantic seafloor, which is about twice the size of Connecticut. It was 420 nautical miles to the closest port. It was necessary to ship in equipment. There was a running clock. The best course of action was still a slow, methodical sweep of the deep ocean, mostly in the dark, in the hopes that something would turn up, despite all the technology that humanity had at the time. Four days later, a remotely operated vehicle from Canada discovered the debris field. Four days, for a comparatively small search area.
Ocean researchers are haunted by that timeline. And a group at MIT Lincoln Laboratory has been working for years to find a solution to precisely this kind of issue.
The Autonomous Sparse-Aperture Multibeam Echo Sounder, the system they have created, operates by deploying a fleet of small autonomous surface vessels that collectively create a sizable sonar array. Rather than a solitary vessel transporting equipment across the water, a well-organized collection of boats arranges themselves in a formation large enough to serve as a single, enormous listening device. The outcome, the researchers claim, is a system that can map the seabed at 100 times the resolution of a typical surface vessel and 50 times the coverage rate of an underwater vehicle. These aren’t small gains. That falls under a different category of capability.
This is especially intriguing because of the deployment model, not just the speed. These boats are light enough to be dropped straight into the ocean from an airplane. Theoretically, hours after a vessel goes missing, a search operation that now takes weeks to mobilize could start. Even the most meticulously designed systems are often humbled by actual ocean conditions, and this is just the beginning. However, the idea has a lot of potential.
A prototype, an 8-by-8-meter metal frame with several sonar subarrays floating in the water with the city skyline behind it, was tested by the team in Boston Harbor. That image has an almost modest quality. A humble-looking rig in a functional harbor, subtly showcasing what could eventually become the norm for mapping one of the hardest-to-reach places on Earth. Further sea trials were conducted in Woods Hole, Massachusetts, where an autonomous surface vessel with optical markers and lidar reflectors was tracked from shore while its onboard sonar scanned the seafloor below.
Here, the larger context is important. Less than 25% of the ocean floor on Earth has been thoroughly mapped. Through a national strategy, the US has pledged to finish mapping all waters deeper than 40 meters in its own exclusive economic zone by 2030. Similar global goals are carried out by the international Seabed 2030 initiative. These are important objectives. Given the size of the ocean and the cost and slowness of the current tools, they are also incredibly challenging. Although they cover ground slowly, underwater vehicles provide high resolution. Surface ships can’t see much detail, but they can cover ground quickly. The goal of the new sonar array is to bridge that gap by increasing resolution without compromising speed.
A complementary technique known as Sonar-MASt3R, which combines sonar and optical camera data to create real-time 3D maps even in murky, sediment-clouded water, is being developed separately by MIT engineers in collaboration with the Woods Hole Oceanographic Institution. Although the two technologies are unrelated, taken as a whole, they suggest that deep-ocean exploration tools are quietly and steadily improving.
Observing all of this, there’s a sense that ocean exploration is about to enter a phase that appears to be more agile, distributed, and deployable than the tedious, ship-bound work of the past. There are still unanswered questions about whether the funding will remain committed, whether the technology will scale, and whether the timelines will hold. However, the path seems obvious. The ocean floor will not remain unmapped indefinitely, and it may change more quickly than most people anticipate.
