Acoustic data lakes: a game changer for anti-submarine warfare?

For decades now, in anti-submarine warfare (ASW), the hunter and the hunted have been playing a game of cat and mouse with progressively rising stakes. Modern submarines are becoming quieter, even “near-invisible”—driving sonar manufacturers to improve, and expand sensor arrays, if they are to maintain detection ranges.

Navies’ ever-greater sonar needs

Traditional sonar-processing involves gathering incoming data and using on-board systems to present it in real time to operators in a form that enables them to classify and localise a potential target—and take the right decisions. But this conventional approach is under strain. Ever-developing sensor arrays now produce a deluge of data at the limits of what operators can process in real time. In addition, an emphasis on littoral operations, geographically-dispersed missions, and “new-normal”, budgetary constraints, add to the complexity and pressure. At a time when half the world’s submarines are soon expected to be operating in the Asia-Pacific and powerful navies are investing heavily, the search for a better approach to sonar processing is occupying the minds of many a naval planner. Yet, the answer to navies’ needs may be on the horizon, driven by rapid advances in computing power and disruptive technologies…

The acoustic data lake: a game-changing concept

The potential game changer is the acoustic data lake concept. It exploits massive advances in processing power and storage capability to enable vast volumes of data to be stocked and allows operators to rapidly call up historical information in addition to seeing real-time flows. This, in turn, opens up a raft of other possibilities. The data lake’s capacity enables all data captured from sensors to be stored, as well as relevant information from previous missions, or other ships, that has been pre-processed by onshore centres.

Unlocking a virtuous circle of improvement

With a large stock of relevant data in hand, the powerful statistical algorithms of Big Data Analytics can be applied to the problem, aided by Artificial Intelligence (AI). This allows the processing of both real-time and historical data, which results in a leap in the quality of information presented to operators—and the quality of decisions. The powerful combination of data lake and historical analysis also has the potential for a virtuous circle of rapid improvement, both on-board and post-mission—with the data lake’s contents downloaded at onshore processing centres. With large volumes of relevant, real data in hand, these techniques can be used to train both algorithms and operators to better interpret it, especially when gathered from weak signals or in complex environments. Improvements can then be exploited in the next mission, leading to even further advances.

Thales: harnessing digital for tomorrow’s sonar—today

But this step change in sonar processing is no pipe dream: Thales is already bringing the acoustic data lake concept to life. We are committed to investing heavily in key digital technologies like AI; and, backed by the expertise of Guavus, the Big Data Analytics pioneer that is now part of Thales, we are developing a sophisticated new platform incorporating both the lake and powerful processing capability.  In ASW then, the game of cat and mouse is far from over, but by harnessing the power of digital, the cat may just be sharpening its teeth…