Searchers combing the southern Indian Ocean for remains of Malaysia Airlines Flight 370 are hampered by a large array of natural factors. First, of course, is the large amount of "ground" to cover. It's a big ocean, and where to look has proved difficult to determine. The second factor is that airplanes don't float. Once they hit the water, they break, then leak, then sink.
Since light only penetrates at most a couple of hundred feet of water, and the Indian Ocean is a whole lot deeper than that, the chances of seeing a wrecked anything sunken in the ocean from a satellite are essentially nil. The only thing visible on the surface will be floating bits that get broken off or flushed out of the aircraft body during the crash. If the plane goes down more or less intact, there ain't gonna be a whole lot to see.
Longer wavelengths of electromagnetic energy used by, for example, radar don't penetrate even that far. The only technology capable of actually finding a wreck on the bottom of the ocean is side-scan sonar. Like any sonar system, side scan sends out pulses of sound energy, which propagate very nicely through water for very large distances. Under good conditions in open ocean, low frequency (say, about 75kHz) sound waves can reach several hundred meters, bounce off a solid object, and return with useable echos. One can assume the hard shell of a fiber-reinforced plastic aircraft (such as the Boeing 777 that MA370 flew) will prove to be an ideal sonar target, providing nice, bright, sharp echos to contrast with the soft crud blanketing most of the bottom of the world's oceans.
With this in mind, the U.S. Department of Defense is sending a suitably equipped automated underwater vehicle (AUV) to join the search -- at the appropriate time. The appropriate time will be when searchers have narrowed down the search area sufficiently to make a side-scan sonar search feasible.
You can't search a whole ocean with side-scan sonar. As I mentioned above, the technology's range is on the order of a few hundred yards, which is on the order of a half mile. So, we've got to get the device within a half mile of the wreck to see anything. Currently searchers believe the aircaft went down in an area thousands of miles across. At 4.5 knots (the AUV's maximum speed), it would take a couple of hundred hours to search a single track 1,000 miles long and a mile wide. That's eight days, not counting time to retrieve the AUV to recharge its batteries!
Even covering an area a hundred miles wide and a thousand miles long would take years.
That's why we need to narrow the search area down before sending the little robot sub off to go looking for the plane.
The reasonable search strategy is to comb the search area with low-flying aircraft hunting for something -- anything -- floating on the surface that can be identified as having become dislodged from the aircraft when it went down. Working backward with known ocean currents, one can then calculate the likely impact location. Then, and only then, it will be time to send in the AUV to search the bottom for the aircraft's remains.
Why use an AUV rather than a manned craft? You want to use a submersible, anyway, to reach close enough to the bottom. The depth in the search area ranges from 3,770 to 23,000 feet. The AUV in question (a Bluefin 21 manufactured by Bluefin Robotics of Quincy, Mass.) has a depth rating of nearly 15,000 feet.
Operations at such depths are dangerous, and we've already established that the search is going to be long and dull. That makes it a hit on two of the three Ds of robotics. The three Ds are Dull, Dirty, and Dangerous. A hit on any of those three flags an application as a likely candidate for robotics. A hit on two means you'd better have a good reason not to try an automated system first.