Talk to me: Natural voice recognition in the car

July 26, 2017
Svenja Borgschulte

"Soon you'll never have to look at your phone in the car again”, says Thrillist. Originally built into fighter jets, the head-up displays (HUD) which are now available in high-end cars, are designed to keep the eyes on the road. But do they really help with distracted driving or just add more of a cognitive overload?  

One thing is for sure: Since it’s earliest days mankind relies on speech as the most natural and fastest means of communication. If the whole history of mankind would be one day, speech would be around 24 hours while text (and such HUD) only a couple of minutes or seconds, said Hector Ouilhet, Head of Design for Google Search at Tech Open Air Berlin.

That’s why products with integrated voice control are becoming increasingly popular. Tech companies like Apple, Google and Microsoft are making calculated bets that intelligent personal assistants are the future. And they’re right: in the future, infotainment systems in the car are no longer served with buttons, but with gesture control – or by voice command.

Assistants hear you, but do they get you?

So far the associated technology of the speech recognition is still in an early stage. In the future, however, it will offer us a whole new world of possibilities, especially in cars and other vehicles. Voice recognition for the end user has existed since the end of the 90s – at that time in the form of a dictation software. Since 2011, voice control has also been available on the smartphone. First, Apple launched Siri on the market. It was followed by competitors like Google Now and also Cortana from Microsoft. Most of these are started with a keyword like "Ok Google" or "Hey Siri".

But there is still even greater potential. The aim of the technology is to recognize and process natural language. And although the speech technology is already able to classify natural language even semantically, the error rate is still significant. Dialects and neologisms for example currently present challenges for developers. However, the technology is constantly being perfected.

Natural voice recognition is closer to reality than you think

In the future, technology will be able to understand human speech and respond naturally. One pioneer in this field is not a well-known language assistant like Siri, but Baidu, a Chinese search engine. The company of the same name has developed a software for speech recognition, which recognizes not only rigid command forms, but the spoken word in the natural language flow. Especially in Chinese, with its many dialects and few syllables, this is a remarkable achievement.

What works with Chinese must also be possible with other languages, right? Of course, and it is more important than it might seem. In-car speech recognition systems have become almost a standard feature in all many new vehicles on the market today. But hands-free voice recognition technology can only be the answer if they are not too overbearing. If you are forced to dictate commands to speak to the system (which don’t always work) it still distracts you from driving. And if it doesn’t understand your dialect or voice, it is totally useless.

Developers are already taking those challenges to the drawing board to work out perfect systems for the car. Systems which understand human speech and respond naturally. So  you soon will never have to look on your phone in the car again.

Svenja Borgschulte
Svenja Borgschulte
Story & Social

More featured articles