Toward Intelligent Monitoring and Assistive Technologies for Healthcare using Smartwatches
People with certain medical conditions (e.g., Stroke, Arthritis) have to follow a strict exercise regimen at home, and often require interactive support from a physical therapist to perform them correctly. Such in-home exercising and other daily health activities such as, frequent handwashing, taking care of mental health are often recommended by the experts for healthy people too. Recently, smartwatches have emerged as a ubiquitous and subtle way for monitoring the users' health activities and the emotional states using its advanced sensing capabilities. However, very little research has focused on building solutions that not only detect these health activities, but also provide interactive assistance based on the quality of the activity. An example of such interactive assistance would be if a patient misses an exercise step or a repetition, or does it incorrectly, a therapist sees it first and then interactively guides him/her. However, this is not possible at home. My research aims to fill this notable gap in technology by building interactive monitoring and assistive systems on smartwatch that integrate an intelligent dialogue exchange system and efficiently utilizes machine learning techniques on the smartwatch sensor signals. Moreover, the smartwatches have very limited processing power and battery, making many high accuracy machine learning solutions practically ineffective. We aim to build three technologies on a smartwatch that will be designed for high accuracy, assessing quality, providing interactive help and operate with low resource usage. First, we will develop a system that measures the quality of arm exercises performed by users and helps them to perform it correctly by multiple levels of interaction. Second, we will implement a system that robustly monitors the user's handwashing behavior and provides real-time feedback aiding proper handwashing. Third, a framework will be created that recognizes the usual emotions of the user as well as the unusual emotions which have similar physiological responses as the usual ones. The proposed research includes working with the doctors and the therapists from the UVA Center for Telemedicine and UVA Stroke Center, and involves real-life data collection and evaluation plans to show the effectiveness of our works.
- Hongning Wang, Committee Chair, (CS/SEAS/UVA)
- John Stankovic, Advisor, (CS/SEAS/UVA)
- Lu Feng (CS/SEAS/UVA)
- Seongkook Heo (CS/SEAS/UVA)
- Tariq Iqbal (ESE/SEAS/UVA)