A novel application of multimodal wearable sensors to detect mobility trajectories in older adults living in the community: Analytical approaches within the MacM3 project

Walking is the most common and accessible dimension of mobility in daily life. Methodological approaches for walking analytics, using body-worn sensors and GPS sensors, have been studied extensively. Individually, these sensors characterize walking in unique ways. Body-worn accelerometry is capable of deriving step-to-step characteristics which can uncover markers of disease or instability; however, these data often lack context that can affect performance, such as indoors vs. outdoors.
Conversely, GPS sensors can capture details about location and engagement with a person’s community to better contextualize mobility behaviours, but are limited in the accuracy of mobility modes and movement inside buildings. There is evidence to suggest that the combination of these sensors can improve characterization of mobility (e.g., determine mode of transportation); however, there is limited research combining these methods to describe and contextualize mobility with high accuracy in daily life. The proposed work aims to combine the strengths of these sensors to assess human mobility in people’s everyday lives.