Autonomic User Interface

The object of autonomic computing is essentially to minimize human supervision, i.e. the computer system must at some level manage all its processes and seek human attention only to resolve high level issues. In today’s computer systems, interfaces are by far the most demanding aspect of computers– we spend more time using the “backspace” key than changing broken disks. Thus for a computer system to be truly autonomic it must possess an “autonomic user interface (AUI)”. An autonomic user interface must provide users with a much higher level of service than today’s interfaces while at the same time being self-aware, aware of its environment, adapting to changes and being self-healing. While tasks like wordprocessing and programming can be carried out efficiently with today’s inflexible keyboard and mouse interface, many tasks that involve querying, controlling and instructing a computer can be completed more easily with an autonomic, multi-modal interface, including speech and visual inputs, and more complex output modes than the traditional monitor. The paper discusses a natural interface to a computer through the use of multiple cameras and microphones as sensors while focussing on ways of achieving autonomic characteristics in such interfaces through the use of multiple sensors, cross modality in input and output, learning algorithms and models of system architecture, the user and the environment.