Sep. 26, 2012 at 8:27 AM ET
As personal electronic devices have proliferated in cars, and with the corresponding increase in crashes resulting from motorists trying to drive while using those gadgets, automakers have touted voice control as the solution that will permit drivers to safely text, tweet and update their Facebook pages to inform the world of the amazing fact that they are in the act of driving to work.
The problem with the notion of voice control as a panacea for distraction is that it assumes that attempting to command electronic devices by voice is not distracting. “It is a great concept, and used in the right way it can work well,” explained Bruce Mehler, a research scientist at the Massachusetts Institute of Technology.
The National Highway Traffic Safety Administration is aware that poorly designed voice interfaces are distracting, so the agency has announced that it will issue rules for automobile voice-command systems in 2014. Until now, no one has determined what kind of system is less distracting and what is more distracting.
MIT researchers are working with Toyota’s Collaborative Safety Research Center in Ann Arbor, Mich. to find the answer to the issue. Their first goal, according to CSRC senior principal engineer Jim Foley, is to establish the metrics that will measure the workload involved in the operation of a voice-command system.
This project’s overall goal is to produce some concrete information on what works and what doesn’t to help guide the federal rules when they come out.
To do this, the team has equipped a car with cameras to monitor drivers’ eyes as they watch the road. Test subjects are also wired up with a heart-rate monitor and an instrument that measures galvanic skin response (like a lie detector) to assess their stress levels. Researchers will study the motorists while they perform tasks such as changing radio stations. The scientists will monitor the drivers while they are using physical controls and voice control.
The researchers also have a test in which they read off a list of numbers and the driver has to repeat a specific number from the list. Repeating the latest number produces a low strain on the brain. However, asking a driver to repeat the number that precedes the latest number cranks up the load on the brain to “moderate”; when the driver is asked to recall the number that precedes that number, that results in a high cognitive workload.
This procedure allows researchers to measure the effects of hard work by the brain on driving skills, so when they find voice interface designs that apply a similar load, as demonstrated by the test subjects’ heart rates and galvanic skin response, they can predict the deterioration in driving ability that will result.
In a simple reaction time test that takes the average driver about two-thirds of a second to respond under normal conditions, the same driver takes nearly a full second to respond when that motorist is called on to repeat the number before the most recent number read to him. At 60 mph, a car travels 29 feet during that additional third of a second, which underscores the need for quick response times.