Behavior-Based Indoor Navigation - Ambient Intelligence for Scientific Discovery - page 279

4.1 TetraNauta Smart Wheelchair

The behavior-based indoor navigation techniques previously presented have been
applied to our smart wheelchair, called TetraNauta (Fig. 10), which is able to
navigate in semi-structured environments (such as hospitals). The TetraNauta's
physical interface is supported by a mobile computer provided with input/output
devices and procedures fitted to the user's specific features. Its main goal is
to automate interaction tasks as much as possible in order to minimize user
intervention. Therefore, the type of devices that may be handled by the user to
communicate with the robot must be quite diverse. TetraNauta maintains the
joystick (or a similar device, mouthstick, headstick, etc., handled with any part
of the body with residual control) as the main input device. In addition one
or more push-buttons are used for start-stop and similar functions. The system
provides feedback to the user by means of a voice synthesizer and a graphical
display. Synthetic voice is useful for short messages with location information or
security warnings, however, when the information provided is more complex, a
graphical display of the results is more convenient.

Fig. 10. A prototype of the TetraNauta System.

The physical interface supports a user-system dialogue model directly re-
lated to the world model of both agents. The interaction is based on a number
of shared assumptions that configure a common vision of the world taking into
account that, usually, humans have a multilevel abstract description of world.
On the subject of spatial navigation, the user tends to think in terms of places
described by attributes such as activity performed there, proximity, people occu-
pying them, etc. Even if the physical interface is the same designed for a classical