User Modeling
Mobile devices are an inherently personal type of computing device. They are always-on, usually-connected, rich in their ability to sense the users context, full of user-specific infor- mation, and capable of autonomous and secure processing. We are exploring the creation of models of users behavior from observed mobile sensor time-series data. Built-in ac- celerometers, gyroscopes, magnetometers, GPS subsystems, WiFi receivers, cameras, and the microphone provide a rich source of information that can be mined in real time.Paper/Projects
|
Mobile Lifelogger/ Activity Recognition
DCAP: Social and Psychological Behavior Modeling Mobile sensors capture a users social interaction patterns from calling frequency, locations visited, and nearby devices. They also capture information such as voice, hand tremors, and typing speed that reflect a users psychological status. Using mobile sensing, we are now able to model a users social and psychological behavior, which enables the caregiver to identify psychological disorder at early stage and provides the opportunity for more effective interven- tion. The DCAP project (funded by DARPA) aims at using mobile sensing and behavior modeling to better understand and mitigate the psychological disorders of veterans suffering from PTSD and depression. |
|
Muscle Recognition
Virtual Medical Assistant Currently, providing quality care for elderly people, care professionals must stay with certain patients 24 hours a day, increasing cost and infringing on privacy. To alleviate these problems, we are developing a Virtual Medical Assistant (VMA) that ubiquitously monitors the medical condition of the patient. Learn more. |
|