SPI studies multimodal interaction in mobile and pervasive computing settings. In particular, we study speech-based and auditory interaction combined with other modalities. Our research focuses on interaction models, interaction techniques and software architectures for interactive systems. Our research covers both human-technology and human-human interaction. We do constructive and applied research combining human-technology aspects with software engineering solutions. The interaction level issues are studied by developing novel interaction techniques and applications and studying how they are used in ecologically valid settings. Developed interaction techniques are integrated in the generic interaction toolkit and architecture. Our research aims at developing solutions that take into account also challenging user groups, such as visually-impaired people.