Figure 3 - uploaded by Claude A. Goodman
Content may be subject to copyright.
Source publication
The goal of our project was to build a prototype activity tracker that would detect and
track a subset of activities of daily living, create alerts about meals and medications, and
to infer changes in the health status of an individual over time. In particular, we used
these data to calculate and track the risk of falls over time. In this technical...
Contexts in source publication
Context 1
... build, test, and evaluate detection and assessment, we used actual, longitudinally collected sensor data from a single individual living in a barrier-free environment. The placement and types of sensors in the reference residence are shown in Figure 3. We assumed that medication data would be provided by the person monitored. ...
Context 2
... observed an actual fall during monitoring. The red circle in Figure 13 shows when the aide entered the apartment in response to the call button. The red circle in Figure 15 highlights how the detection rule segmented the data, such that the fall from the bed to the floor appears to the left of the actual sleep segment. ...
Context 3
... oracle is described in Figure 23, which defines activities that occur, activities we can detect, their baseline values, and problems we can detect. In addition, we include observations about problems we could detect with additional sensors. ...
Context 4
... from the point of view of the data and activity tracked, it was important to extract information while remaining agnostic with respect to diagnosis, since this is information that the tracker, as a tracking system and not a diagnostic system, was not designed to process. Figure 23 contains our reference scenario, or oracle, for our canonical subject, Martin (Default) McGee, and 81-year-old male, diagnosed with diabetes, hypertension, COPD, and arthritis, taking 12-17 medications, among these, subcutaneous insulin. Figure 23. ...
