Speech Technologies Take to the Skies in Newest Fighter Jet
When the U.S. Air Force rolls out its newest fighter jet, the F-35 Lightning II Joint Strike Fighter, pilots will be talking to it instead of physically manipulating controls and display panels in the cockpit.
The planes, being built by Lockheed Martin for the U.S. Air Force, are expected to be in service by 2013. They will incorporate SRI International’s DynaSpeak speech recognition engine into their onboard computer systems, allowing pilots to give voice commands to control both communication and navigation systems and enter data into the flight management system. The expected benefit to the pilots is that they will be able to concentrate on their mission and maneuvering their planes without having to flip switches or press buttons to do things like change radio frequencies.
"There’s a big cognitive load on the pilots, and offloading some of these command and control functions will be a big help to them," explains Doug Bercow, director of business development at SRI.
The F-35 cockpit "was designed with speech recognition in mind," says Gary Pearson, vice president of advanced programs at Adacel, the Orlando, Fla.-based systems integrator that is designing and installing the speech application using the DynaSpeak technology.
The DynaSpeak system being deployed in the F-35 will feature additional noise-canceling technologies to filter out the many other sounds in the cockpit. Furthermore, the speech recognition technology is speaker-independent, meaning that any pilot flying the plane could use the system immediately. The pilot speaks to the system through a microphone integrated into his flight helmet. A press-to-recognize switch, when depressed, tells the system to begin recognizing the pilot’s voice inputs.
Pearson is quick to point out that the voice technology will not be used for critical flight or tactical functions. Also, "you can’t get into the system from outside. The only input is through the pilot’s mask," he says.
As an added layer of protection, pilots can still use manual controls to perform the same functions that they can with speech. "All the speech stuff is complementary to manual control," Pearson says. "The pilot can use one or the other."
Though currently being developed for the 3,000 to 6,000 F-35 planes the U.S. military is expected to buy, the system now is trained to recognize only English words and phrases, but can be adapted easily to accept acoustic models for other languages, according to Pearson. The defense departments from several other countries have already expressed an interest in the technology, he adds.
Adacel’s work on the F-35 is its first foray into the fighter jet arena. Earlier this year, the company won a contract with Rockwell Collins to develop a Voice Activated Cockpit (VAC) interface for its next-generation Pro Line avionics suite for civilian aircraft.
That system is expected to begin flight testing in 2008.