Research
Human Behavior Analysis with Physiological Signals
Our current research focuses on leveraging the physiological signals from consumer-grade devices to recognize and analyze human behaviors, including his/her physical activites and mental states. Deep learning methods are designed to extract meaningful features from the raw physiological signals. Currently, we focus on the photoplethysmogram (PPG), electrocardiogram (ECG), and electrodermal activity (EDA) signals available on most of the off-the-shelf wearable devices.
Proximity-based Contact Tracing with Mobile and Wearable Devices
This project leverages the Bluetooth signals from mobile devices to detect the proximity between any two individuals. By logging the proximity information detected from everyday life, we can send a notification alert to individuals who are more likely to contract the virus when someone is diagnosed with the infected disease. Our contact tracing uses a decentralized approach to protect the privacy of each individual.
Bluetooth Low Energy Beacons
Bluetooth low energy (in short BLE) is a very popular short-range wireless technology. Among all the ble devices, ble beacon have been widely used for IoT development. Here are some examples of Beacons on the market that has been used to deliver various IoT applications, including proximity-based service in gallery, occupancy detection in hospital, providing smart parking services, and monitor the safety of workers. One important parameter that the receiver can measure upon receiving the advertising packet is the received signal strength (RSS). as shown here, we can see that the RSS decreases when the distance increases. Hence, how can we manipulate this RSS-distance information for IoT applications.