Smart sensing and analytics for cognitive health care

The rapidly growing elderly population in many parts of the world has started to impact private and public interests — emotionally, socially and economically. Specifically, current demographics, medical and social trends in the U.S. indicate a large population of older adults being at risk of early cognitive impairment and dementia. This motivates us to design innovative technology solutions for cognitive health assessment, care and well-being in elderly people in their own environments of daily life. Among many significant challenges involved in such technology assisted solutions, an important one is multi-modal sensing and complex sensor data analytics on resource constrained wearable and regularly used daily objects, particularly in the presence of uncertainty and noisy environments. An additional critical challenge is how to transfer the extracted knowledge into actionable information to be effectively used by family members, doctors and caregivers. This keynote talk will present our ongoing research on elderly cognitive care with the help of complex at-home activities recognition with smart wearable, and user emotional behavior signatures detection with smart chair. For activity recognition, our innovative methodology is the first of its kind that detects 21 fine-grained and complex at-home activities by wearable devices, as compared to typical 6–12 at-home activities recognized by relevant works in the existing literature. The smart chair based novel framework can accurately detect user functional and emotional activities, in addition to static and movement based sedentary postures that are more common in current works. In collaboration with Phelps County Regional Medical Center, the proposed solutions are being validated with clinical data and patient studies.