Proceedings of the
9th International Conference of Asian Society for Precision Engineering and Nanotechnology (ASPEN2022)
15 – 18 November 2022, Singapore

Digital Twin Smart Home Using Robust Mat Monitoring System with Multi-Modality Deep Learning Analysis

Yanqin Yang1,3, Qiongfeng Shi1,2,3, Zixuan Zhang1,3, Tianyiyi He1,3, Xuechuan Shan2,4, Budiman Salam2,4 and Chengkuo Lee1,2,3,5,a

1Department of Electrical and Computer Engineering, National University of Singapore, Singapore 117583, Singapore

2Singapore Institute of Manufacturing Technology and National University of Singapore (SIMTech-NUS) Joint Lab on Large-area Flexible Hybrid Electronics, National University of Singapore, Singapore 117583, Singapore

3Center for Intelligent Sensors and MEMS (CISM), National University of Singapore, Singapore 117608, Singapore

4Printed Intelligent Device Group, Singapore Institute of Manufacturing Technology, Agency for Science, Technology and Research (A*STAR), Singapore 637662, Singapore

5NUS Graduate School - Integrative Sciences and Engineering Programme (ISEP), National University of Singapore, Singapore 119077, Singapore


Digital twin attracts numerous attentions in the creation of Metaverse because of its capability of real-time projecting the physical world into the digital counterpart, thereby providing a more intelligent, immersive and comprehensive communication between the physical and virtual world. To construct a digital-twin smart home, versatile Internet of Things (IoT) sensors are required to monitor real-time home-related information. Conventional IoT sensors are mostly based on resistive or capacitive mechanisms which rely heavily on power supplies. Self-powered sensors based on triboelectric mechanisms can be potential candidates to address this issue. However, the practical deployment of triboelectric nanogenerator (TENG) based self-powered sensors are still hindered by some intrinsic limitations, such as unstable output in response to the changes in environmental conditions and user behaviours. In this work, we develop a robust triboelectric mat with high tolerance to environmental and user behavior variations. The triboelectric mat consists of an interdigital electrode (IDE)-designed environment-insensitive in-home mat array and a two-channel entry mat at device level. At data analytics level, time-domain analysis and multi-modality deep learning (DL) are utilized. Furthermore, leveraging the comprehensive sensor information and VR technique, a digital-twin smart home is successfully visualized.

Keywords: Digital-twin, Smart home, All-triboelectric mat system, Self-powered sensor, Robustness, Scalability, Multi-modality, Deep learning

PDF Download