Nudging Occupants for Energy-Saving through Voice-Based Proactive Virtual Assistants

With the advancement of the Internet of Things (IoT) technologies, smart homes have promoted human-building interaction and sustainability in occupants' daily life. The rise of voice-based AI-powered virtual assistants has brought new potentials to provide occupants with a convenient and intuitive interface to interact with smart homes. Aiming at enhancing the human-building bi-directional communication, voice-based proactive virtual assistants integrated with smart home ecosystems-that is, Smart Home Assistants (SHAs)-were investigated in this study. A comprehensive data collection was conducted through an online experiment, in which 307 valid questionnaire responses with participants' demographic background information and their feedback to proactive SHAs were collected. Occupants' perception of the proactive SHAs was evaluated among different groups of users with various demographic backgrounds. Five occupants' socio-demographic background features were identified to have a significant impact on their acceptance level to the energy-saving suggestions by proactive SHAs, including gender, age, education level, employment status, and the number of occupants in a residence. By utilizing these demographic features, ensemble learning models can predict occupants' perception of proactive SHAs with good performance (accuracy in the range of 0.69-0.75). Findings in this study will provide a valuable reference for academic researchers and industry practitioners in the development of personalized proactive smart home systems for human-building interaction.