Our research centers on applications, systems, and networking aspects of mobile computing. We are currently exploring the following research directions.
Today's smartphones are equipped with a rich set of sensors (e.g., audio, GPS, accelerometer, Bluetooth, etc). We study the use of these sensors to track user behaviors and activities, and infer their behavioral pattern and mental wellbeing.
Our current project StudentLife examines the correlation between smartphone automatic sensing data and students' behavioral trends, mental health, and academic performance. We conducted our study on a group of undergraduate students in a smartphone programming class over 10 weeks of Spring 2013. We have released our dataset via this link. Please refer to our papers at UbiComp'14, UbiComp'15, and UbiComp'16 for detail.
We seek to address the impending spectrum crunch problem by exploring the visible light spectrum band. Its unique propagation characteristic presents new challenges, requiring new thinking on the mobile system and networking design, and on its interplay with human-computer interaction (HCI).
One of the past projects HiLight is on screen-camera communication. It presents a new way for screens and cameras to communicate without showing any coded images (e.g., barcode) on the screen. Our initial design appeared in ACM VLCS'14 and was awarded the Best Paper Award. Our full paper will appear in MobiSys'15. Check out the demo video clips at the HiLight project website.
Our latest project DarkLight studies the feasibility of maintaining light-based communication even when the LED light emits extremely low illumiance. For more, please refer to our HotMobile'16 and MobiCom'16 papers.
On the application side, our recent projects LiSense, StarLight turn ubiquitous lights into a sensing medium that senses what we do and how we behave. You will find more in our MobiCom'15 and MobiSys'16 papers.
Energy efficiency is a critical design goal for any mobile systems. It is increasingly important in the post-smartphone era, as our mobile devices become diversified and miniaturized.
Our project WiScan focuses on Wi-Fi scan energy efficiency. Wi-Fi connectivity is highly desired in our connected world. However, seeking maximal Wi-Fi connectivity faces a key obstacle -- the battery drain of frequency Wi-Fi scans. We set out to address the root cause of the energy inefficiency of today's Wi-Fi scan implementation. Our prototype implementation shows significant energy saving. Please check out our UbiComp'15 paper for detail.