Towards Privacy-Preserving Voice Control in Smart Home IoT: A Taxonomy

Modern smart homes combine the Internet of Things (IoT) and voice control technologies to connect home hubs and devices, and provide the user convenient control over their environment. Smart homes are an instance of the Internet of Everything (IoE), where the IoT facilitates integration between home devices, and voice control allows user’s a natural interaction modality for device control and information seeking dialogues. Cloud-based processing is typically used in such setups to process device data, understand voice commands and generate responses, however the privacy and security implications of the cloud-side analysis and storage of user data has negatively impacted adoption of smart home technologies. This paper explores the design of a privacy-preserving voice assistant for smart homes that uses purely local computation to ensure people’s data remains private. The paper presents a taxonomic classification and analysis of the voice control technologies and IoT devices present in current commercial smart home systems. We follow with a discussion and comparison of open-source technologies and types of hardware that are applicable to the design of a cloud-free voice assistant for smart homes. By focusing on local processing techniques, our findings demonstrate the potential for more secure and trusted voice interactions with IoT environments, enhancing users’ privacy in the IoE.