Updated: 02.22.2014

Can You Understand Me Now?

As hands-free devices have become the norm in today’s world, I think we have all experienced the frustration felt by Howard in the following clip from The Big Bang Theory:

Although we have not yet achieved the level of sophistication anywhere near that portrayed by the Enterprise’s computer, voice recognition technology has certainly come a long way since its was first developed in the early 1950s.

The first voice recognition system, named Audrey, was developed by Bell Laboratories in 1952, and could only understand numerical digits spoken by a single voice.  By the mid-70s, Carnegie Mellon’s “Harpy” system had expanded voice recognition abilities to the point where it could understand 1011 words, which is approximately the vocabulary of a three-year-old. Advancements in both the hardware and software algorithms used for speech recognition helped to expand this vocabulary over the next several decades, but it wasn’t until Google introduced its cloud-based Voice Search application for the iPhone that voice recognition truly became a commercially viable product. Today, Google’s English Voice Search system uses a cloud-based database of approximately 230 billion words that have been extracted from actual user queries.

Although we no longer need to speak V-E-R-Y   S-L-O-W-L-Y for modern speech recognition systems to understand us, I know there are times when my phone totally misunderstands my verbal command and has me shaking my head as I wonder how “Call Home” could be interpreted as “Order Pizza.” Honestly, poor voice recognition/translation is nothing more than a humorous annoyance for those of us trying to place a phone call to our homes and instead end up being connected to the local pizza joint.

But speech recognition technology is now being incorporated into far more critical applications than just our being able to phone home, such as in airplane cockpits.

Traditionally, general aviation pilots have had to manually enter their flightplans into onboard GPS systems, waypoint by waypoint, using a knob to scroll through each letter of the alphabet to spell out the aviation call sign for each waypoint. The longer a flight is, the more waypoints are required. For example, when flying from Albany, NY, to Florence, SC, sixteen different waypoints are required – the first three of which are AGNEZ, SAGES, and LHY. Entering each of these into the GPS is a tedious, time consuming activity at BEST, but easily becomes a safety hazard when mid-flight changes to a flightplan are required due to changing weather conditions or some other unforeseen situation. Manually modifying GPS data diverts a pilot’s attention away from flight-critical activities such as scanning gauges and/or avoiding other aircraft.

Engineers at NASA have long recognized the importance of flight safety not only for their high-profile manned-space missions, but also as it applies to general aviation and increasing safety for the flying public. For over 50 years, NASA’s Vehicle Systems Safety Technology (VSST) program has studied ways in which cockpit safety can be improved for all aircraft. Randy Bailey, one of VSST’s lead aerospace engineers at Langley Research Center, recognized that one of the best ways to improve flight safety is to minimize pilot distraction.

“GPS navigation systems are wonderful as far as improving a pilot’s ability to navigate, but if you can find ways to reduce the draw of the pilot’s attention into the cockpit while using the GPS, it could potentially improve safety.”

— Randy Bailey, Langly Research Center
“Speech Recognition Interfaces Improve Flight Safety”
NASA Spinoff 2012, p56

Meanwhile, civilian pilot J. Scott Merritt had become increasingly frustrated with the “antiquated” method of twisting knobs when entering flightplans into his GPS. Because he understood the inherent danger of this, Merritt wondered if speech recognition technology could be used for waypoint entry. He partnered with several professors at Rensselaer Polytechnic Institute (RPI) in New York State in order to study the feasability of his idea. After developing and validating an initial design concept, a US Patent was awarded and soon thereafter Pragmasoft Inc (now VoiceFlight Systems LLC) was formed.

In 2005, Pragmasoft/VoiceFlight Systems LLC was awarded a NASA Small Business Innovative Research (SBIR) contract which enabled the company to deliver a fully-functional unit to Langley Research Center where it underwent successful flight testing. In July, 2009, the system became the first voice-recognition system to receive FAA-certification for use in civilian aircraft.

Since its launch in 2001, VoiceFlight®* has become commercially available for general aviation aircraft weighing less than 6000 pounds. What does this mean for private pilots who have installed VoiceFlight? Quite simply, increased flying safety and overall ease when entering or editing flight plan details. With a simple touch of a button installed on the yolk, pilots are now able to issue voice commands to:

  • Enter a waypoint using the standard phonetic alphabet;
  • Edit a flight plan;
  • Delete/erase a flight plan;
  • Enter direct-to destinations;
  • Enter Victor airways which are then automatically expanded.

An online demonstration of the ease with which information can be verbally entered using the phonetic alphabet is available on the VoiceFlight Systems webpage

VoiceFlight is a prime example of how one man’s idea can come to commercial fruition with the help of NASA’s SBIR program – an idea that will improve the safety of the flying public.

*VoiceFlight® is a registered trademark of VoiceFlight Systems LLC.

Leave a Reply

You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>