The New York Times has an article today about how cars overloaded with portable DVD players and computer keyboards are actually making the roadways considerably more dangerous.
Man. Computers? When I was ten, my parents took me on a thirteen hour roadtrip from California to Utah (Violin camp. Seriously). They played an unabridged audio book of Anne Rice’s Interview with the Vampire, which was both entertaining and educational.
According to the NY Times, the biggest causes of driver distraction leading to accidents include dialing phone numbers and texting on wireless devices. Of course, now that everyone has a wireless device, there have been a ton of TV spots advertising Ford’s voice-controlled Sync system (which uses Nuance’s speech rec technology).
[youtube]http://www.youtube.com/watch?v=t1WkTxNb7rY[/youtube]
The article quotes a Ford technical specialist: “Voice interface is sort of the way of the future…Voice dialing is clearly superior to dialing manually.”
Well, great. I wrote a case study about equipping police units with voice controls in the Jan/Feb issue of Speech Tech. Researchers at the University of New Hampshire, who implemented and continue to optimize the project, are also conducting research on how speech recognition influences driving ability. Andrew Kun, an associate professor in UNH’s Electrical and Computer Engineering Department describes the dilemma:
So someone gives you a system with a speech recognizer and you’re now supposed to be able to drive a little more safely. But what if the recognizer isn’t very good or just isn’t performing very well? Will that hinder your ability to drive? If the question is that you now have a second task, which is talking to the computer, and it’s not going well because the computer isn’t understanding you, how much does this influence your primary task? By how, I don’t mean what goes on in your heads, I mean what are the result regarding driver performance.
When I spoke with him, they’d only completed a few WOz tests using extreme cases–44 percent and 88 percent recognition rates. As expected, very low recognition rates hindered driving ability and high recognition rates helped it. I don’t know what Sync’s recognition rate is, but I’m pretty sure it’s higher than 44 percent, otherwise I doubt they’d have deployed it.
As Kun admits, “There are a lot of variables that have to be taken into account.”
One is speech recognition rate, the other is what kind of driving do you do? Is a highway more difficult than city driving? Do you have to use the press-to-talk button? We’re trying to come up with a list of things.
