Innovating on the Fly and over the BattlefieldJuly 30, 2010
The Pentagon has portrayed its use of drones in Afghanistan as an unparalleled success. But recently the military acknowledged that some of its unmanned aerial vehicles suffer from computer glitches and human error that lead to system failures, according to an article in the Los Angeles Times.
Unarmed drones have been used since the mid-1990s, but armed drones were flown into combat just nine months after being retrofitted, according to the Pentagon. These armed drones, it acknowledged, just weren’t ready when the military began its campaign to drive the Taliban and Al Qaeda from Afghanistan.
"It was never designed to go to war when it did," Lt. Col. Travis Burdine, a manager for the U.S. Air Force Unmanned Aircraft Systems Task Force, told the Times. "We didn't have the luxury of ironing out some of the problems."
The article reports that military engineers actually purchased off-the-shelf components at Radio Shack and Best Buy to improvise a system to capture and stream video from the drones, and since they were not designed for the rigors of warfare, it is not surprising that they have relatively high failure rates. Other problems suggest that the designs were not thought through completely. For example, the article notes that at least one drone crashed because it had no fuel gauge and the operators did not realize it was running out of fuel.
Most weapons used by the U.S. military spend years in research and trials before being deployed in or over the battlefield, so such mistakes are found earlier in the process. In this case, the military took a chance with an untested system because it bet the benefits might outweigh any problems.
Christian Terwiesch, a professor of operations and information management at Wharton, and co-author of Innovation Tournaments, a book that deconstructs the innovation process, called this a classic "learning-by-doing" strategy. This strategy is often useful in situations with a high level of uncertainty. Terwiesch said the uncertainty might reflect complex technologies or unpredictable market requirements — like the use of a drone in a new role in an untested environment.
"In many high-uncertainty settings, you are better off just launching a beta version and then iterate afterwards," he said. In fact, the Air Force says that design and training improvements — along with lessons learned about their problems — have lowered the drones' accident rate.