Speech Drives Safety Solutions for Cars

Article Featured Image

The automotive industry is facing the challenge of implementing technology while improving safety. Navigation, texting, and music management are the major areas of concern, as growing research shows that speech stands out as the key enabler in maintaining driver safety.

Driving is so basic that people don’t think of it as a complex task. But substantial physical coordination and analytical skills are required. The cognitive load of driving has not decreased over time. Increasing traffic levels; complex road systems, often subject to construction and constriction; and a much higher flow of information and infotainment to the vehicle have made ordinary driving a demanding task.

And with our connected lifestyles, enabled by mobile devices, driving has taken on new dimensions that can contribute to safety hazards.
Text messaging has become mainstream, with both adults and the younger generation guilty of texting while driving. This behavior can lead to a dangerous situation unless the user interface is radically changed.

Navigation also needs a robust and safe user interface. It makes no sense to type while operating a vehicle, especially using a knob that must be twisted and nudged until each target letter is highlighted, followed by a push (“knobbing”). Though it’s an awkward experience, knobbing sometimes is the only way to enter a destination into a navigation system, and the vehicle must be stationary.

The recent past has brought limited speech input capability—impressively, too—but not to the point where users can say a full address, business name, or category using natural speech. Ultimately, the ideal experience would be to say the destination as though the user is telling a cab driver where he would like to go.

Not surprisingly, texting and navigation system entry are the most common sources of driver distraction. And everyone—the public, industry, media, researchers, and legislators—has taken notice. The focus should be on how to promote safe driving with automotive user interfaces that depend little on vision. Speech stands out as an enabler.

Speech solutions—like outbound texting, destination entry, weather, traffic, location sharing, and route planning—are trending toward a mixture of on- and off-board technologies made possible by the connected car. In future implementations, the vehicle will connect to an off-board system through a wireless voice channel. During each interactive session with a driver, a server-based speech technology (IVR) is used with speech resources accessed through a data channel—the speech cloud. The user experience includes interactions with multiple speech recognizers within one dialogue session.

The connected car speech system can include the vehicle, an off-board IVR, and the speech cloud. The interfaces for the connected car services include an interaction of embedded and off-board speech technology. The driver initiates a connection to the off-board speech system by pushing a button, touching an icon on a display, or speaking a command to the on-board speech system.

Once the connection has been made, the IVR executes a dialogue with the driver. Depending on the driver’s intent, spoken utterances are automatically converted to text strings that may represent a text message or a destination. For example, a spoken utterance can be free-style text, an address, a business name, or a business category. To achieve high accuracy, the speech recognition process includes passing the audio over a data channel to one of many specialized speech recognizers, referred to as speech cloud recognizers. The IVR selects the appropriate cloud recognizer based on the driver’s intent.

Information, such as target destinations and routing, to support connected car services can be accessed remotely and downloaded to the vehicle. Navigation is a key service involving off-board connectivity and on-board voice interaction with a driver. For example, a driver may connect to an off-board service designed to capture a spoken destination, such as an address or business name: The driver speaks a destination, which is downloaded to the vehicle’s navigation system. The interactive navigation driving experience follows.

In the future, look for connected car solutions to leverage the speech cloud directly, allowing dialogue to be managed from within the vehicle. Indeed, the industry is moving in the right direction, as innovative speech solutions allow drivers to text and enter destinations safely while driving, making these common behaviors practical and easy. Most importantly, automotive speech interfaces can enhance the driving experience by safely enabling functionality that requires text entry.

Thomas Schalk is vice president of voice technology at ATX, a Cross Country Automotive Services company that provides telematics services to the automotive market. A member of the AVIOS board and a former president, Schalk can be reached at tschalk@axtg.com.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues