Most of the wearable devices in the market are consumer fitness tracker for recording heart rate, sleep behavior and step count etc. Even though some devices can monitor the blood pressure (BP), which only provide a snapshot measurement. They cannot serve as medical devices to perform long-duration and high accuracy monitoring of vital signs for daily-use, especially cardiovascular-related vital signs. It is important to measure BP continuously for nighttime monitoring and accurate diagnosis of different cardiovascular disease symptoms and vascular dementia for elderly. In view of this, CUHK team develops an unobtrusive health-monitoring system combining multimodality sensing and machine learning technologies by integrating novel sensor designs, signal generation and processing methods (included multi-wavelength photoplethysmogram and multi-sensor coordination) in a medical wearable platform, in order to achieve early detection of cardiovascular diseases, effective remote health monitoring and prevention of dementia for the elderly.