TY - GEN
T1 - YouHome System and Dataset
T2 - 8th IEEE International Symposium on Smart Electronic Systems, iSES 2022
AU - Pan, Junhao
AU - Yuan, Zehua
AU - Zhang, Xiaofan
AU - Chen, Deming
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Recent research has actively investigated technolo-gies, such as artificial intelligence (AI), energy-efficient computing, and privacy protection, to bring us more advanced smart home experiences. However, existing work has failed to integrate these key technologies to provide a complete smart home solution to understand user behaviors, protect user pri-vacy, and provide personalized suggestions. In addition, there are not many available smart home datasets that can support thorough studies of user behaviors in household environments. To address these critical problems, we present YouHome: the first smart home framework for developing integrated smart systems with a strong emphasis on usability, scalability, user privacy, edge/cloud coordination, and a comprehensive activities-of-daily-living (ADL) dataset to assist the smart home research. YouHome includes a flexible framework with three levels of system abstraction to collect sensor data, recognize user behavior, and generate personalized suggestions. We also establish multi-stage approaches in this work to protect user privacy effectively from threats such as data leaks. Finally, we implement an end-to-end prototype system to demonstrate the efficacy of our proposed framework. The accompanying YouHome ADL dataset is the first annotated dataset with multiple sensing data types tailored for smart home research and applications, including video, illuminance, temperature, humidity, motion, and sound. It covers two household settings, 20 users, and 31 different activities. We have released our comprehensive dataset to enable numerous research opportunities in both smart home and machine learning communities. We also intend to open-source the framework in the future.
AB - Recent research has actively investigated technolo-gies, such as artificial intelligence (AI), energy-efficient computing, and privacy protection, to bring us more advanced smart home experiences. However, existing work has failed to integrate these key technologies to provide a complete smart home solution to understand user behaviors, protect user pri-vacy, and provide personalized suggestions. In addition, there are not many available smart home datasets that can support thorough studies of user behaviors in household environments. To address these critical problems, we present YouHome: the first smart home framework for developing integrated smart systems with a strong emphasis on usability, scalability, user privacy, edge/cloud coordination, and a comprehensive activities-of-daily-living (ADL) dataset to assist the smart home research. YouHome includes a flexible framework with three levels of system abstraction to collect sensor data, recognize user behavior, and generate personalized suggestions. We also establish multi-stage approaches in this work to protect user privacy effectively from threats such as data leaks. Finally, we implement an end-to-end prototype system to demonstrate the efficacy of our proposed framework. The accompanying YouHome ADL dataset is the first annotated dataset with multiple sensing data types tailored for smart home research and applications, including video, illuminance, temperature, humidity, motion, and sound. It covers two household settings, 20 users, and 31 different activities. We have released our comprehensive dataset to enable numerous research opportunities in both smart home and machine learning communities. We also intend to open-source the framework in the future.
KW - Smart home
KW - cloud computing
KW - data set
KW - machine learning
KW - privacy protection
UR - http://www.scopus.com/inward/record.url?scp=85148080249&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85148080249&partnerID=8YFLogxK
U2 - 10.1109/iSES54909.2022.00091
DO - 10.1109/iSES54909.2022.00091
M3 - Conference contribution
AN - SCOPUS:85148080249
T3 - Proceedings - 2022 IEEE International Symposium on Smart Electronic Systems, iSES 2022
SP - 414
EP - 420
BT - Proceedings - 2022 IEEE International Symposium on Smart Electronic Systems, iSES 2022
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 19 December 2022 through 21 December 2022
ER -