Modelling interface agents for personality-based behaviour

Visually synthesising the metaphor of life-like visual personal service assistants (PSAs) is the focus of much current research. Often the complex visual requirements of the PSA means that little attention is given to the underlying infrastructure necessary to support the reasoning requirements of such a metaphor. This lack of support, from a software engineering perspective, means that many designs and implementations are application specific thus re-use of the designs and implementations are limited. To address this problem our research goes behind the scenes of visual stage representation to provide a technical infrastructure enabling the creation of more believable life-like interface characters based on a structured meta-representation called the Asset Description Language (ADL). The language is extended to enable agents to communicate emotions with other agents (software or human). Hence this explicit representation of emotions within the communication means that an agent can reason about these emotions and respond in an appropriate manner either visually to a human or through communication with another agent. This response permits an agent to exhibit personality. Through the provision of a structured meta-representation we aim to evaluate the effect personality-based behaviour on interactive user interfaces as well as on other service agents in a multi-agent environment. As the representation provides some basic general design primitives which are not application specific means that re-use of our infrastructure by other developers is possible