Adaptative Multimodal Architectures Managing Software Qualities

Multimodal interfaces for natural human-computer interaction involve complex architectures that should facilitate the process of matching IT to people. These architectures should react to events occurring simultaneously, and possibly redundantly, from different input media. In this paper, intelligent expert agentbased architecture for multimedia multimodal dialog protocols are proposed. The generic components of the multimodal architecture are monitored by an expert agent, which can perform dynamic changes in reconfiguration, adaptation and evolution at the architectural level. Software performance and usability are maintained by the expert agent via a scenario-based methodology. The expert agent’s behavior modeled by Petri nets permits a software quality tradeoff between attributes of usability and other software attributes like system’s performance.