Bots to Outsmart Mine-Hunting Dolphins

The Navy's been trying for years to improve the slow and plodding process of underwater mine detection. Now, they're using the military's affinity for all-things AI to create a mine detection system that minimizes errors through "adaptive learning."The Navy's recent mine-hunting efforts have proved imperfect. In 2003, they credited the deployment of dolphins with clearing 100 underwater mines from the Persian Gulf. But the dolphins require lengthy training and a human handler at all times. And then there's that constant harping from animal advocacy groups. In 2007, Lockheed Martin created a Remote Minehunting System, which got a test run on the Bainbridge destroyer. A 2008 report from the ship's skipper, Commander Stephen Coughlin, cited "growing pains" using the unwieldy `bot, which is lifted from the water using a giant hook, precision timing - and luck.Now, the Navy is funding four projects that may not produce a sleeker system, but hope to produce one that learns from its mistakes. The idea is to program undersea `bots with sonar tools that are linked to adaptive algorithms. As the `bot accumulates new input from ocean-floor surveillance, the algorithms will improve accuracy according to correct and incorrect mine identification.The automated feedback learning mechanisms proposed herein will provide a unique capability to adapt the feature extraction, selection and classification process that can lead to improved false alarm and target identification rates as the system is matured.The `bots would be programmed with preliminary data, but they'd continue to train themselves once deployed. And no need to stockpile fish as rewards for a job well done.But despite the shortcomings of dolphin mine-seekers, the proposed AI alternative has a lot to live up to. In 2003, Navy spokesperson Tom LaPuzza said it was "doubtful anything man-made will ever match the dolphins' capabilities."By Katie Drummond June 23, 2009 3:45 pm Categories: Navy