A new study by the Virginia Tech Transportation Institute found that interactive voice systems used for outbound text messaging reduced levels of driver distraction compared to manually texting on a smartphone.

The study was commissioned for telematics services provider ATX Group of Dallas, which developed the in-vehicle voice technology system.

Thomas Schalk, ATX's vice president of voice technologies, said the research was conducted in tandem with the company's development of voice-activated, voice-to-text, and navigation system destination entry features that will be part of Hyundai Motor America's forthcoming Blue Link telematics program.

"The VTTI study supports interactive voice applications for in-vehicle tasks rather than manual interaction and visual focus on the part of the driver, with the voice applications reducing those periods when hands are off the steering wheel and eyes are off the road," Schalk said.

Key conclusions from the VTTI research included:

• Eye glance data indicates that the ATX voice interface is safer than the equivalent manual interface for entering text messages. For manual texting, the number of eye glances to the mobile phone was measured to be 20 times the number of glances to the initiating button required for the outbound voice texting. Data also revealed that eye glance durations and the percentage of time a driver's eyes were off the road were shorter using the ATX voice interface as compared to comparable tasks performed using the handheld phone.  Eye glance data similar to findings observed while texting were captured for destination entry tasks using the ATX voice interface, again supporting the goal of driver safety while entering text.
• Handheld tasks required significantly more than twice the mental workload demand than the ATX voice-based tasks. Similar findings were also recorded for frustration levels, a measurement that correlates with stress and annoyance in the performance of tasks.
• Drivers' self-perceived awareness of their surrounding environment was higher when engaged in voice-based tasks as compared to those completed manually. Reliance on voice-based systems didn't affect drivers' ability to complete the tasks they were asked to perform.

"While we can't conclude at this point that voice interfaces eliminate driver distraction, the data indicates that voice can certainly reduce the risk," Schalk said. "With the improvements in acoustically modeling the vehicle cockpit, extending language models, voice interface metrics, and simplified dialogue design, the ATX solution provides our clients with an exciting solution."

The study was conducted with 24 participants who were familiar with texting and owned a smartphone. Variables used to assess the handheld and voice systems included task duration, glance frequencies, glance durations, total glance time as a measure of percentage of eyes off road time, speed maintenance, steering variance, lane deviations and time out of lane, and subjective ratings of mental demand, frustration level and situation awareness.