
		AT&T Laboratories Cambridge

The Augmented Vehicle project

What do you mean by 'Augmented Vehicles'?

We believe that we can use technology to enhance the time spent in the car. To make journeys safer, more interesting, quicker and easier. Actually, the platform we are developing should be suitable for general mobile use - it might run on a PDA in a hiker's backpack, for example - but cars are our initial focus.

Why is a telecoms company like AT&T interested in cars?

The average U.S. driver spends nearly 10% of their waking hours behind the wheel. During this time, they are unable to make use of much that an internet services and communications company has to offer. People on the move tend not to access email, instant messaging, voicemail, traffic information, news, schedule updates, route guidance, weather reports, for a variety of reasons:

It is unsafe to do so while travelling
In-car versions of the applications are not readily available
The car has only sporadic connectivity to the outside world, despite the wide variety of networks now available.

Tell me more about these. How will you make it safer?

Firstly, we think that our systems will be safe enough if they provide no more distraction than the current in-car components such as speedometer, fuel gauge, radio and aircon. We should also be able to reduce some of the current in-car distractions. For example, you probably don't need a speedometer if you have something that warns you when you're going noticably faster than is typical or legal on this particular road. You shouldn't need to look at maps or road signs so often if you have a really good route guidance system. And the fuel gauge could be replaced by something which warns you when there are less than ten filling stations in range with your remaining fuel supply.

Secondly, we hope to detect the current driving situation, and modify the presentation of information accordingly. Information which is best displayed graphically when you are stationary may be better presented using speech when you are moving. Incoming phone calls might receive a 'busy' signal if you are currently negotiating a difficult junction or operating other controls.

What sort of applications will you build?

We have a mantra: "Sense, predict, automate". We think that current in-car systems are awkward to use because they don't monitor enough of the surrounding environment, and because they don't sufficiently predict what the user is likely to want. So we plan to build some more 'intelligent' apps which require less intervention from the user.

Route guidance systems, for example, should be able to use both live and historic traffic information and the driver's own route and driving style history to make intelligent guesses about which route he or she may prefer, where they are likely to stop on the way, and how long it will take them to get to the destination.

Similarly, in-car entertainment could be improved by systems which knew the types of music you had been listening to at home recently and which were your favourite radio programmes. The music, and the automatically-recorded programmes, could be downloaded to your car whenever it was in range of your home wireless network.

What sort of networks are you considering?

Most of our work so far has just used simple SMS messages to send small quantities of information to and from the car, and 802.11b when the car is close to the home or office. Data over conventional GSM has not been very convenient, though the data rate is starting to increase now. We expect to make use of GPRS in the near future, and we hope to participate in the upcoming Cambridge 3G trials. But one aim of the project is to make intelligent use of whatever network is available at any given time, depending on the quantity and urgancy of the data, and to present the options in a convenient way to the application developer.

Plenty of other companies are building in-car systems, aren't they?

Yes, but most of these are either car manufacturers, or audio/video specialists, who want to add value to their particular product by adding features. We think it is important to address the general problem of in-car systems from an architectural viewpoint, with particular reference to communications.