Computing and Information Systems - Theses

Permanent URI for this collection

Search Results

Now showing 1 - 1 of 1
  • Item
    Thumbnail Image
    Designing Proactive Smart Speakers and Their Application in Voice Surveys
    Wei, Jing ( 2023-05)
    Smart speakers are an emerging technology in recent years. They provide an interface that allows users to request information or control smart devices via natural languages. Most existing smart speakers exhibit a passive interaction mode. They need to be activated by voice commands first before acting upon user requests. This passive way of interaction limits the potential application scenarios of smart speakers. If smart speakers can proactively approach users, they can actively engage users through offering suggestions, sending well-timed reminders, and assisting users to collect self-reports for reflection by conversations. Hereby, this thesis investigates the implementation of proactive smart speakers and their use case in data collections through voice surveys. We build a proactive smart speaker prototype for this research and investigate the opportune moments and user perception of proactive interactions. We further explore the interaction errors and user obstacles with proactive smart speakers through an in-depth quantitative analysis of various in-situ interaction data. Drawing upon our own findings and prior work, we deploy voice surveys on smart speakers and evaluate the reliability and validity of different question types. Through building a proactive smart speaker prototype, evaluating user experience in the wild, and developing a voice survey application, this thesis presents findings from empirical data of proactive smart speaker and voice applications and discusses corresponding design implications.