Shijia Pan
University of California Merced
Lecture Information:
- October 23, 2020
- 2:00 PM
- Zoom: Contact SCIS coordinator (Hadi Amini) for zoom credentials if you did not receive the email.
Speaker Bio
Dr. Shijia Pan is an Assistant Professor at University of California Merced. She received her bachelor’s degree in Computer Science and Technology from the University of Science and Technology of China and her Ph.D. degree in Electrical and Computer Engineering at Carnegie Mellon University. Her research interests include cyber-physical systems, Internet-of-Things (IoT), and ubiquitous computing. She worked in multiple disciplines and focused on indoor human information acquisition through ambient sensing. She has published in both top-tier Computer
Description
The number of everyday smart devices is projected to grow to the billions in the coming decade, which enables various smart building applications. These applications, especially in-home long-term occupant monitoring, rely on the emerging device-free human sensing techniques. From the system perspective, general problems faced by existing technologies (vision-, sound-, RF-, and load-based methods) are the tradeoff between the information acquisition resolution and the sensing requirements (e.g., line-of-sight, high deployment density, calibration). We introduce an alternative non-intrusive indoor human information acquisition through ambient structural vibration, or ‘Structures as Sensors’. People’s interaction with ambient structures (e.g., floor, table, door) induces structures to vibrate. By capturing and analyzing the vibration response of structures, we can indirectly inferfine-grained (action-level) information about the people and their actions. However, due to the complexity of the physical world, sensing data distributions can change significantly under different sensing conditions. Therefore, from the data perspective, accurate information learning through pure data-driven approaches requires a large amount of labeled data, which is costly and difficult to obtain in real-world applications. We address these challenges by combining physical and data-driven knowledge to reduce the label data needed via multiple approaches.