Designing Cars That Learn How You Drive

Engineers working for the automobile industry are currently working towards creating smarter cars that could literally learn your driving patterns. The new, highly-advanced systems would have the advantage of figuring out when you are not driving safely, and display a message, warning you of this. Already, existing luxury vehicles learn their drivers' preferred seat position, and the optimum setup of the steering wheel, and so the new improvements could be the next big step forward, ICT Results announces. A prototype of the new system is already undergoing testing.

In the case of human drivers, individuals who want to learn how to drive a car go to driving school, where they get a grip on how to anticipate and avoid emergency situations. They are also instructed on how to respond appropriately when such an event does occur. What researchers are now trying to do is apply the same approach in the cars themselves. The project, called DRIVSCO, is being funded by the European Union, and led by expert Florentin Worgotter. He envisions that state-of-the-art sensors, image processors, and learning algorithms, could make such an advanced vehicle a common reality within this lifetime.

Thus far, the DRIVSCO program includes a complex system, which is capable of observing a driver's motions. It then coordinates these recordings with data sets collected from its sensors that watch the road ahead, and creates patterns of behavior. It is therefore able to learn how its driver handles various situations, including going past other cars, and negotiating curves. When activated, it alerts the person behind the wheel when he or she exceeds the boundaries of these algorithms significantly enough. Data on what is happening on the road are collected using infrared headlights, stereo cameras, and advanced visual processing methods.

“What we wanted was a system that learns to drive during the day by correlating what it sees with the actions a driver takes. Then at night the system could say, ‘Slow down, a curve is coming up!' – a curve the human didn’t see. Now we have a prototype that does this,” the team leader adds. “How the visual front-end of DRIVSCO works was very much inspired by the visual cortex of vertebrates. The feedback mechanism, where higher-level modules interact with modules that detect simpler features, solves the very difficult problem of detecting independent objects even when you and they are moving at the same speed,” he concludes.