Destination Entry while Driving: Speech Recognition versus a Touch-Screen Keyboard

Omer Tsimhoni, Daniel Smith, and Paul Green

June 2002

To determine the effect of several destination-entry methods on driving performance, as a function of driving workload, 24 participants drove a simulator on roads with curves of different radii while entering street addresses into a simulated in-vehicle navigation system. Three destination-entry methods were examined: 1) speech recognition by word, 2) speech recognition by character, and 3) typing on a touch-screen keyboard. For each method, driving performance, glance behavior, task partitioning, and subjective evaluation were examined.

While driving, speech recognition by word yielded the shortest task completion time (15.3 s), followed by speech recognition by character (41.0 s), and then the touch-screen keyboard (86.0 s). Overall, driving performance degraded when entering destinations, especially with the touch-screen keyboard, and when driving workload increased. The participants made 21% shorter glances to the navigation display, and 40% longer glances at the road scene as a function of increasing driving workload. Similarly, touch-screen keyboard-entry behavior was significantly affected by an increase in driving workload and age. When driving, pauses between fields of an address increased 57% and pauses between groups of characters increased 27%. These results confirm the risks associated with destination entry using a touch screen and suggest the use of speech recognition. However, speech recognition systems with less than perfect accuracy and visual feedback are not risk free.

Graphical Abstract (.pdf) | UMTRI-2001-24 Full Report (.pdf)

 

Close This Window