Data Fusion for Unobtrusive Assessment of Mobility

Abstract

Mobility is a core function affecting quality of life for the elderly. In addition to the obvious implications for the ability to perform activities of daily living, certain aspects of movement, such as the number of steps taken in a given time interval or the speed of walking, may be used to predict a person’s future cognitive function. In this project, we will investigate novel techniques for the assessment of elders’ mobility in real life by fusing data gathered through unobtrusive, ubiquitous sensing and computing technology. Our specific aims are 1) to develop models and corresponding computational algorithms that will process and integrate large quantities of data from unobtrusive, continuously monitoring sensors and provide manageable summaries of these data; and 2) to develop algorithms to infer clinically relevant measures from the representation in Specific Aim 1 capable of detecting changes and trends and supporting the generation of predictions of future cognitive status. The resulting algorithms will enable the reduction of within-subject variability, and the detection of subtle changes in mobility that can be used to predict future health status.