作者: Sandra Hellmers , Babak Izadpanah , Lena Dasenbrock , Rebecca Diekmann , Jürgen Bauer
DOI: 10.3390/S18103310
关键词:
摘要: One of the most common assessments for mobility older people is Timed Up and Go test (TUG). Due to its sensitivity regarding indication Parkinson's disease (PD) or increased fall risk in elderly people, this assessment becomes increasingly relevant, should be automated become applicable unsupervised self-assessments enable regular examinations functional status. With Inertial Measurement Units (IMU) being well suited analyses, we evaluate an IMU-based analysis-system, which automatically detects TUG execution via machine learning calculates duration. as duration single components. The complete was classified with accuracy 96% a rule-based model study 157 participants aged over 70 years. A comparison between durations determined by IMU criterion standard measurements (stopwatch automated/ambient (aTUG) system) showed significant correlations 0.97 0.99, respectively. classification instrumented (iTUG)-components achieved accuracies 96%, well. Additionally, system's suitability investigated within semi-unsupervised situation where similar movement sequence executed. This preliminary analysis confirmed that self-selected speed correlates moderately situation, but differed significantly from each other.