Flexible Activity Tracking for Older Adults Using Mobility Aids — An Exploratory Study on Automatically Identifying Movement Modality

Abstract

Wearable activity trackers are inaccessible to older adults who use mobility aids (e.g., walker, wheelchair), because the accuracy of trackers drops considerably for such movement modalities (MMs). As an initial step to address this problem, we implemented and tested a minimum distance classifier to automatically identify the used MM out of seven modalities, including movement with or without a mobility aid, and no movement. Depending on the test setup, our classifier achieves accuracies between 82 % and 100 %. These findings can be leveraged in future work to combine the classifier with algorithms tailored to each mobility aid to make activity trackers accessible to users with limited mobility.