The semantic web comes to cars

In-car navigation systems exist for some time now. But BBC News reports that a new German project, dubbed SmartWeb, will use the semantic web and peer-to-peer networks to interact with drivers. This system, which is currently in its development phase, will use speech recognition and human gestures as interfaces. And it will warn drivers about jams and dangers. For example, a car detecting slippery conditions will pass the information wirelessly to all the vehicles following it. The drivers will be informed via their dashboard screen or a GPS-equipped mobile device.

In-car navigation systems exist for some time now. But BBC News reports that a new German project, dubbed SmartWeb, will use the semantic web and peer-to-peer networks to interact with drivers. This system, which is currently in its development phase, will use speech recognition and human gestures as interfaces. And it will warn drivers about jams and dangers. For example, a car detecting slippery conditions will pass the information wirelessly to all the vehicles following it. The drivers will be informed via their dashboard screen or a GPS-equipped mobile device. But the SmartWeb will also transmit other kinds of information to drivers, such as parking availability or speed traps.

Before going further, below is a picture showing how a motorbike driver would be informed of a danger ahead by a car in front of him (Credit: Wolfgang Wahlster). This picture has been picked on page 31 of a presentation given by Wahlster at the "50 Years Artificial Intelligence Symposium" held in Bremen, Germany, in July 2006, "Three Decades of Human Language Technology in Germany" (PDF format, 36 pages, 1.72 MB). You also should take a look at page 28 for a picture describing a dashboard interface telling a driver where the next speed traps are.

SmartWeb in action on the road

This project is led by the Deutsche Forschungszentrum für Künstliche Intelligenz (DFKI) -- or German Research Center for Artificial Intelligence. You can find more information on the Intelligent User Interfaces page. And here is a link to the official website of the SmartWeb project, led by Professor Wolfgang Wahlster.

So how SmartWeb-equipped vehicles will be warned of a danger? Let's go back to the BBC News article.

For example, cars could spot oil on the road by combining temperature readings with wheel traction information, said [Dr. Anselm Blocher, the SmartWeb project manager.] Once a car detected this sort of danger, information about it would be generated and passed down the line of vehicles approaching the patch of oil. "When the motorbike comes after to the point of danger, information has been spread out by wireless network and the danger will be propagated to the driver in the motorbike," said Dr Blocher.

But how the information will be transmitted?

If a driver was executing a series of fast manoeuvres, such as a motorbike driver leaning to go fast round a bend, the system would not use a blaring alarm to warn them of the upcoming oil patch. Instead, he said, it might generate a warning on the dashboard of the bike or mark the danger point on a digital map. By contrast, if a driver was driving at low speed along a straight road, the system may use visual cues on a dashboard screen as well as telling the driver about the problem via a headset.

However, this project has some limitations as reminds us Nate Anderson at ars technica in "SmartWeb brings semantic web search to cars" (March 19, 2007).

Because the percentage of web pages that feature semantic markup is statistically indistinguishable from zero, SmartWeb currently attempts to mark up standard HTML pages automatically using its own "advanced language technology and information extraction methods."

So will the SmartWeb be successful? Will it be integrated with other European projects? We'll discover it in a few years.

Sources: Mark Ward, BBC News, March 17, 2007; and various websites

You'll find related stories by following the links below.

Newsletters

You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
See All
See All