Attentive user interfaces are user interfaces that aim to support the user’s attentional capacities. By sensing the users’ attention for objects and people in their everyday environment, and by treating user attention as a limited resource, these interfaces avoid today’s ubiquitous patterns of interruption. Focusing upon attention as a central interaction channel allows development of more sociable methods of communication and repair with ubiquitous devices. Proposed methods are analogous to human turn taking in group communication. Turn taking improves the user’s ability to conduct foreground processing of conversations. Attentive user interfaces bridge the gap between the foreground and periphery of user activity in a similar fashion, allowing users to move smoothly in between. The work is focused on motivating the need for attentive user interfaces and on the practical implementation of some of them. In particular 4 different attentive interface are described based on eye tracking techniques (eye contact detection). The first solutions supports device control into the home environment by detecting eye contact, thus detecting the control context and providing only the needed interfaces. For example an attentive lamp is proposed which listens the on and off commands only when looked by the user. The second solution aims at improving attention detection by sensing physiological parameters such as ECG or EEG (more intrusive). The third solution is an attentive headset composed of a noise cancellation headset and two eye contact detectors. The headset cancels the surrounding noise except when the wearer is in eye contact with another person also wearing the same kind of headset, thus allowing communication. Eventually an attentive wall is described, built using a Liquid Crystal Window that is normally obscured and can become transparent when user attention is detected, thus allowing inter-cubicle communication.

The paper describes a control interface that uses gaze tracking to enable severely disabled people to control electronic devices easily. Instead of offering a centralized interface, controllable by eye, the proposed system is able to detect the object laying in the user field of view and to present only those controls suitable for the identified device. Two possible implementation are investigated one using a head mounted tracker and the second using multiple remote trackers. The former solution is more accurate but has several drawbacks: the tracking apparatus has to be worn (becoming stressful for long periods of usage), the calibration shall be done at every new use session, i.e. each time the user wears the system, the system is very sensitive to bright lights and daylight and eventually the allowable range of movements is quite limited (around 50°). The latter is more complex to implement and less accurate, however it has several advantages as it only requires one calibration per user, it is not much sensible to light conditions and the range of acceptable head movements is wider that for the head-mounted one (170°). Home control is achieved by means of an X10 RF controller which interacts with X10 actuators distributed in the house environment. Unfortunately, this control network, although cheap, is very primitive and supports only very simple automation, thus limiting the possibility for a user to control its own home. As a solution, the authors suggest the adoption of more sophisticated networks such as insteon or z-wave.