Ford Aims to Make Voice Recognition the "Primary Interface" for Cars
In-car voice recognition has always been a "killer app," given its advantages in convenience and the prospects for avoiding distracted driver syndrome, but it's just never worked all that well. That's changed with the latest generation of SYNC, which powers the in-car infotainment system called MyFord Touch. Back in 2007, SYNC could barely handle 100 commands, but the latest handles 10,000. And with "natural voice recognition," you can talk to the car in a normal voice.
With Brigitte Richardson, a Ford voice recognition engineer, behind the wheel, here's a video demo: Ford developed the new version of SYNC with voice control leader Nuance, and debuted it July 15. The company says it wants to make voice recognition "the primary user interface" in the car, which means a lot less reaching for buttons and dials. The improved technology is in line with a receptive marketplace. According to a Harris Interactive survey, consumer acceptance of voice recognition grew eight percent year-over-year in 2010.
Have you ever laughed out loud when a voice recognition system (on the phone or in the car) got something ludicrously wrong? Me too, and I've generally found them more trouble than they're worth. There's no guarantee that this latest voice recognition system will overcome the inherent problems -- slow reactions, incorrect identification, a weakness working with colloquial speech patterns -- and become intuitive for users.
Likewise irksome are some dial-and-screen-based interfaces, such as BMW's iDrive, which wants a lot of unnecessary driver attention. Systems like iDrive, far from solving the problem of driver distraction, can actually add to it. A wag at The Truth About Cars wrote that the system "encourages you to take your eyes off the road long enough to plough into a solid object." According to Richardson, you're not stuck in a rigid pattern with the SYNC system (which debuts in the 2011 Ford Edge), and you can "keep your eyes on the road and your hands on the wheel."
For instance, to make the cabin warmer, you can say "warmer," "increase temp" or "temp up." To hear a song, you can say, "play song" or "play track." To dial in a radio station, you can say "FM 89.5." You can just say an address, and the navigation system will program it in. You can even say "find dougnuts," and it will give you're a range of choices, all via synthesized voice. It can also guide you to the nearest chain stores.
Obviously, sophisticated voice recognition systems require considerable computing power, and increasingly it is onboard the car. Henri Richard is a senior vice president at Freescale, which makes the embedded silicon microprocessor, also known as the brain, of the SYNC system. "In all vehicles, the electronic content has been increasing very steadily," he said. "In high-end cars, there might be 80 micro-controllers, each as powerful as the one in your cellphone, each as powerful as your PC."
Richard said that voice recognition is a challenge for engineers, because despite a lot of computing power, cars have relatively small amounts of memory (512k to one megabyte, which tends to be embedded in the microprocessors themselves). "The software people have to very frugal at developing code, fitting it into a tiny bit of memory," he said.
According to Richardson, "As we've gained processing power and learned more about how drivers use the system, we've been able to refine the interface. Customers can do more and say more."
Another challenge is filtering out the ambient noise that is all around you in a moving car, and modern voice recognition is very good at that, Richard said.
As Richard points out, Ford has been able to leverage SYNC as a powerful brand in itself, one that actually influences car-buying decisions. I can see that, because some unique features (like the ability to play music off a hard drive) make it attractive to people like me. With voice recognition added, SYNC is cool, it's groovy, it's cutting-edge new technology.
Related:
Photo: Ford Motor Company