A study released today shows that the use of in-car voice recognition systems has a positive impact on road safety, by allowing drivers to control stereo and navigation systems and cellular telephones using voice commands.
The 2008 In-Car Distraction Study, carried out by the Technical University of Braunschweig in Germany, is based on the Lane-Change-Task (LCT) test, an International Standards Organization (ISO) certified test that exposed 30 drivers to a range of driving skills assessments, challenging each to drive while performing various common tasks within the car. The test simulated driving and changing lanes, all while simultaneously selecting music on an MP3 media player, making phone calls, and setting the address on a satellite-based navigation system.
The most significant finding of the 2008 In-Car Distraction Study is that speech-recognition significantly reduces distractions and improves driving performance while selecting music, making calls and using navigation systems in the car. Further, the data shows that speech recognition delivers the most dramatic benefits in reducing distractions in the car.
Key findings of the study include:
- Using the Phone – While most people think they are adept in dialing a mobile phone while driving, the study showed that even with a car kit speech input improved the ability to maintain the ideal car position by 19 percent when compared to manual dialing. Better still; speech input was also approximately 40 percent faster in making a call, reducing the distraction period by the same amount.
- Selecting Music – With the growing popularity of using an MP3 player in the car, the safe use of these devices is of paramount concern. The study found that the average driver is 50 percent more distracted and takes more than two times longer for lane changes when selecting music manually, versus being able to simply say the artist and song title via a speech-based interface. Swerving within a single lane was even worse without speech input, with 600 percent higher levels of distraction.
- Setting Navigation Address – Not surprisingly, using a cumbersome manual interface to enter city, street and street number into a navigation system, results in significant safety risks. In contrast to entering information manually, voice destination entry resulted in 1,000 percent or ten times less swerving while staying in a single lane and 30 percent less distraction while changing lanes.
- In-lane Deviation – The study also measured how much drivers moved from the perfect lane position. The drivers showed significantly less deviation when controlling the tested devices by voice versus manual input – with speech input resulting in 60 percent less deviation from the ideal when selecting music and 50 percent less when entering a destination.
- Reaction Time – When required to change lanes, the reaction times when drivers were using voice commands was consistently better than when controlling devices manually. For example, making a music selection using speech was 66 percent better versus manual input.
- Swerving in Current Lane – The study also measured how much drivers had to correct their position when staying in the current lane. When compared to speech input, drivers using manual input swerved within the lane 800 percent more for music selection, up to 1,200 percent more for navigation entry, and 300 percent more when dialing a phone.
- Eyes on the Road – The analysis of drivers’ eye movements revealed that voice commands help drivers keep their eyes on the road, reducing driver distraction to almost zero for music selection and less than 10 percent for phone dialing and destination entry. On average, speech helps keep drivers eyes on the road 200 percent to 300 percent better than manual input.
- Natural Speech Input – The study exposed an interesting finding even within different speech recognition interfaces used, especially for navigation entry. It compared the impact of conventional destination entry by voice requiring individual confirmation of city name, street name and house number, and a more advanced user interface requiring a single confirmation. While the conventional speech input for city, street and address reduced reaction time by 24 percent compared to manual operation, the system supporting single confirmation reduced the reaction time by 47 percent.