View non-flash version
projects, we must offer them something that they can?t do, or we can do better than others.? Real ArtiÞ cial IntelligenceCMRE draws on its legacy of knowledge, expertise and ex- perience in the undersea environment. But the research has applications in all aspects of the maritime domain. CMRE is studying autonomous unmanned underwater systems to nd enemy submarines and neutralize underwater mines, requiring multidisciplinary approaches involving autonomy, arti cial intelligence, and system interaction and collaboration.?The systems today are limited to human-guided systems or systems that execute preplanned missions. Human opera- tors must intervene to observe events, make decisions and guide the vehicle,? said Warren Fox, program manager for the overall autonomous naval mine countermeasures program at CMRE. ?They are very operator-intensive. ? ?We?re looking at a more automated robotic solution. We?re trying to eliminate tethers, and allow multiple vehicles to com- municate with each other and work together while people are kept at a safe distance from potentially mined areas,? said Fox. These systems must have the ability to learn about their sur- roundings, and be able to think.?One of the big limitations on sensing the environment, which is the intelligence you need on which to make a deci- sion for intervention, is dependent on how big an aperture you have. And that used to mean a physical aperture towed by a big ship, which could be kilometers long,? said Potter. ?Now, with lots of autonomous vehicles, you have an option for cre- ating an aperture out of discrete elements that could also be kilometers along or across. AUVs are smaller and cheaper than ships, allowing us to develop new intelligent-adaptive ways of gathering environ- mental data, forming a big-picture of what is going on and responding to it. But that comes with some very signi cant challenges. Now that the technology of autonomous vehicles and robots is rel- atively mature, the challenge has really shifted toward their intelligent behavior. It?s one thing to make a system autono- mous. It?s quite another to make it intelligent.? ?Communication in that realm of the world is very dif cult and prone to frequent disruption,? said Been. ?It?s complex and dynamic due to the sheer physics of the problem. That means that the bandwidth is quite low for communications. And so, locally, these systems have to be intelligent. They have to know what to do.? ?How do you have a collection of autonomous, intelligent assets control themselves as a team to produce more than the sum of the parts? That?s not a solved problem,? said Potter. A recent experiment that exempli es this multi-disciplinary approach sounds like a sci- scenario. ?To say ?We?ve got all these autonomous underwater robots, and they can listen for and detect submarines, and cooperate with each other to hunt the submarines down,? sounds very Matrix-like. It?s a very sci- kind of vision. And, indeed, it?s a big leap forward to devolve the intelligent actions of hunting a submarine down to an automated system?it?s a huge step in machine intel- ligence.? ?We have just now reached the point where all the pieces have been tested and we?re actually putting the whole thing together and having a couple of vehicles in the water in a very realistic scenario where they will do exactly what I just de- scribed,? says Potter. Operational S&T?Our charter is to organize and execute an S&T program in the maritime environment. Which for me is made of two parts: the physical environment and its synthetic counterpart. You know when you?re in the physical maritime environment because it tastes like salt. You know when you are in the syn- thetic environment when you are looking at a screen. Both are important,? said Gough. We have acoustic engineers, physicists, computer scientists and mathematicians all working to better understand the mari- time environment and nd solutions for maritime autonomous unmanned systems, underwater communications and mari- time information,? said Gough. ?It?s our job to help NATO operationalize that S&T in the maritime and turn it into what the commanders and decision-makers advantage. It opens up options for war ghters.? ?We don?t know enough about the ocean to know what it will do next. We?re always working at the frontier,? Gough says. ?When we?re at the frontier, we never really know what?s next.? Military ?We?re looking at a more automated robotic solution. We?re trying to eliminate tethers, and allow multiple vehicles to communicate with each other and work together while people are kept at a safe distance from potentially mined areas? Warren Fox, program manager for the overall autonomous naval mine countermeasures program, CMRE46 MTRNovember/December 2012MTR #9 (34-49).indd 46MTR #9 (34-49).indd 4612/3/2012 11:34:32 AM12/3/2012 11:34:32 AM
