Background and Motivation        

Multi-Modal MapQuest will provide the full range of MapQuest consumer services to mobile users on handheld and in-car devices. The interface will achieve a new level of accessibility, reliability and usability by tightly integrating spoken dialog with a graphical user interface.

Motivation

MapQuest is the world leader in online, voice and wireless mapping and driving directions, providing destination information solutions to business partners and consumers anytime, anywhere. MapQuest users can retrieve door-to-door driving directions and helpful travel information such as the nearby hotels, restaurants, ATMs, gas stations and much more through the MapQuest Locator.

MapQuest is also a key component of the AOL Anywhere strategy. The company already delivers mapping information via standard web browsers through HTML, via WAP phones and other handheld devices through WML, and over the phone via VoiceXML.

The primary goal for the application is to allow mobile users full access MapQuest's location-based services through a variety of devices including in-car telematics and hand-held devices.

Interaction will flexibly integrate voice and grahics, allowing any step of the dialog to be carried out by voice input and output or by pen input and graphic output or any combination of the two input and output modes. Voice output from the device can be muted for situations in which a user wishes to have a silent interaction. The application will be fully usable without ever looking at or touching the screen, allowing access in situations where a user's hands or eyes may otherwise be occupied, such as when driving or carrying a briefcase or other package.

Aesthetics

As one of the most recognized and trusted brands on the Internet, any MapQuest application must be accessible, reliable, responsive, and above all, usable. Users should have a strong feeling that MultiModal MapQuest facilitates the tasks they need to do, when they need to do them, with whatever device they have available.

Every month, 1 in 3 internet users access MapQuest content for maps, driving directions and other location-based information. Given MapQuest's strong web presence, Multi-Modal MapQuest should reinforce MapQuest's brand through a consistent look and feel that's shared as much as possible with the unimodal graphical and voice interfaces already available.

Like the web site and voice-only applications, Multi-Modal MapQuest must be highly responsive. Although there is no getting around transmission times for maps, directions and other data-intensive services, the application should always respond quickly to user input, and it should always be clear if database lookup is happening, how long it is expected to take. In the spoken portion of the dialog, the system will operate at conversational speed, avoiding unnecessary wordiness, and using colloquialisms where appropriate. At the same time, the application will not feel rushed, and the user will always be able to back up and/or correct previous inputs. The voice will have a calm intonation, and pauses will be inserted appropriately to space out vital pieces of information for optimal comprehension by the user. This will be particularly critical for synthesized output.

There will be two levels of verbosity in prompting, one minimal set for mixed-mode usage, and one verbose set for voice only usage.

Output (including feedback on user input) will always be displayed both graphically and verbally. The user will have the ability to mute the sound and operate the application entirely through pen input and graphical output. Conversely, the application should be usable without looking at the display. Mixing and matching of the modalities will typically be seamless and the user can employ best mode for any task given the perceptual and production resources they have available.

For display, we assume a touch screen of at least 160x160 pixels. This is consistent with existing telematics devices in cars running proprietary operating systems, and with existing high-end hand-held devices such as the Color Palm Pilot running the Palm OS or the Compaq iPAQ running PocketPC. It should also work on third generation (3G) mobile phones when they become avaiable.