Johns Hopkins Bloomberg School of Public Health

This project focuses on advancing objective measurement of activities of daily living (ADLs) through wearable and implantable sensing and machine learning. Using multi-sensor data streams such as accelerometer-based movement and photoplethysmography (PPG) for heart rate, the team aims to extract single- and multi-sensor features that can both recognize specific ADLs and indicate ease or difficulty of performance. The analytical toolkit includes Lasso/LassoNet for feature selection and regularization, with consideration of complementary approaches such as artificial neural networks, Hidden Markov models, support vector machines, and movelet methods. The effort targets delivery of robust, bias-aware features packaged into a front-end software application for clinicians and researchers.The work is led by Amal A. Wanigatunga, PhD, MPH, an epidemiologist at the Johns Hopkins Bloomberg School of Public Health with expertise in aging, gerontology, and objective physical activity and function measurement. His experience designing and analyzing large epidemiologic datasets informs the project’s emphasis on reproducible feature extraction, linkage with established functional assessments, and practical tools that can support independent living assessments in older adults.