Building the robotic driving instructor

Nothing will send a shiver down your spine like handing the keys to your car to a teen driver.

But what if a robotic instructor could turn every ride into a driving lesson, or give warnings about dangers ahead while suggesting safer, alternate routes?

“I would love it,” says Laura Holmquist, a photographer and mother of a 15-year-old about to get a permit. “Many of the rules of the road have been updated from when I learned to drive. I feel the robot's info would be current and accurate, so less pressure on myself.”

New technology that uses computer algorithms to monitor inputs like steering wheel position, accelerator position and vehicle location could be used to train new drivers, and intervene when they do something wrong. The technology is still in a research phase, but automakers are already thinking about potential uses.

At Carnegie Mellon’s Quality of Life Technology Center, driving assistance is a major focus of study. Researchers have developed adaptive navigational systems that analyze the routes drivers normally take and then suggest alternatives that avoid interstates, unguarded left-hand turns, busy intersections, or roads that are narrow to the point of being treacherous.

Brian Ziebart, a researcher on the project, says the next goal is to predict where a driver might be heading and suggest alternative routes if there is a potential hazard, such as inclement weather. For now, the team is working on commercializing the product for use by automakers. The research has expanded into a new company called NavPrescience.

Another system, called the Masternaut Sustainability in-cab coach, is used by fleet managers to teach their drivers about “green” driving. A light bar in the cab indicates how well the driver is doing – if they are braking too much, idling too often, or accelerating too fast. The system can monitor these driving behaviors over time and generate a report.

One implication: the same in-dash system could teach any driver about better driving practices, tapping into the navigation system to capture information about upcoming intersections and current speeds, and provide driving tips.

“Coaching is becoming more common,” says Jeff Greenberg, a senior tech leader with Ford Research.

At Ford, new technology that uses the Google Prediction engine suggests directions based on where you normally drive and the time of day. “Future vehicles could include a ‘coach me’ mode where the vehicle asks the driver to perform a maneuver – for example, accelerate to 40 mph and then back to 0 mph -- that analyzes fuel economy performance and provides feedback on how to improve,” says Greenberg. “A heads-up display could use augmented reality to virtually paint lane lines and make them more visible.”

Some existing Ford vehicles already have robotic teaching aids, like the Ford Focus Electric, which uses a coaching feature that trains drivers to brake with less pressure to save on fuel economy. Several Infiniti and Mercedes-Benz models can nudge you back into your lane and adjust your speed automatically if you’ve become distracted, setting off a warning to remind you to pay attention.

Of course, all of this additional in-car technology could itself lead to more distracted driving. Greenburg says there will be a fine balance between mild audible cues and visual feedback for a potential “coach me” mode in the car, whereas a collision warning is always loud and attention-grabbing.

But one thing is for certain: a computer will never become irritated like a human driving instructor can or suggest unsafe (or illegal) driving practices.

“I think [my daughter] would concentrate on a robot’s instruction and it could help her stay focused,” says Holmquist. “It could possibly help eliminate some fighting between us.”

Click here for more from Autos