Almeida-2017a

PhD thesis

In this work, we propose, design and implement a framework enabling easier development of multimodal interaction. Our proposal fully decouples the interaction modalities from the application, allowing the separate development of each part. The proposed framework already includes a set of generic modalities and modules ready to be used in novel applications. Among the proposed generic modalities, the speech modality deserved particular attention, attending to the increasing relevance of speech interaction, for example in scenarios such as AAL, and the complexity behind its development. Additionally, our proposal also tackles the support for managing multi-device applications and includes a method and corresponding module to create fusion of events. The development of the architecture and framework profited from a rich R&D context including several projects, scenarios, and international partners. The framework successfully supported the design and development of a wide set of multimodal applications, a notable example being AALFred, the personal assistant of project PaeLife. These applications, in turn, served the continuous improvement of the framework by supporting the iterative collection of novel requirements, enabling the proposed framework to show its versatility and potential.