Otago Exercises Monitoring for Older Adults by a Single IMU and Hierarchical Machine Learning Models

TL;DR

The paper tackles unobtrusive monitoring of Otago Exercise Program (OEP) adherence in older adults using a single waist-mounted IMU for HAR. It introduces a two-stage hierarchical system: Stage 1 uses a CNN-BiLSTM on a $10$-minute window to separate OEP from ADLs, followed by Stage 2 using a RF on a $6$-second window to classify six Level-1 OEP sub-classes and, subsequently, 15 Level-2 sub-classes. Data from lab and home environments show Stage 1 achieving $f1 > 0.95$ and $IoU > 0.85$, while Stage 2 achieves $f1$ above $0.8$ for four sub-classes in the home setting, demonstrating feasibility for real-life OEP monitoring with a single sensor. The work contributes a hierarchical labeling strategy to reduce class complexity, demonstrates the potential of a minimal wearable in daily life, and provides a benchmark across lab and home contexts; limitations include limited training data for Stage 2 and sensor placement calibration, suggesting avenues for larger datasets and end-to-end models like TCN in future work.

Abstract

Otago Exercise Program (OEP) is a rehabilitation program for older adults to improve frailty, sarcopenia, and balance. Accurate monitoring of patient involvement in OEP is challenging, as self-reports (diaries) are often unreliable. With the development of wearable sensors, Human Activity Recognition (HAR) systems using wearable sensors have revolutionized healthcare. However, their usage for OEP still shows limited performance. The objective of this study is to build an unobtrusive and accurate system to monitor OEP for older adults. Data was collected from older adults wearing a single waist-mounted Inertial Measurement Unit (IMU). Two datasets were collected, one in a laboratory setting, and one at the homes of the patients. A hierarchical system is proposed with two stages: 1) using a deep learning model to recognize whether the patients are performing OEP or activities of daily life (ADLs) using a 10-minute sliding window; 2) based on stage 1, using a 6-second sliding window to recognize the OEP sub-classes performed. The results showed that in stage 1, OEP could be recognized with window-wise f1-scores over 0.95 and Intersection-over-Union (IoU) f1-scores over 0.85 for both datasets. In stage 2, for the home scenario, four activities could be recognized with f1-scores over 0.8: ankle plantarflexors, abdominal muscles, knee bends, and sit-to-stand. The results showed the potential of monitoring the compliance of OEP using a single IMU in daily life. Also, some OEP sub-classes are possible to be recognized for further analysis.

Figures (10)

• Figure 1: An example of recorded acceleration from one subject
• Figure 2: An overview of the proposed system. NOTE: To better illustrate the process, the ratio of the sliding windows and signals in the figure does not correspond to the actual size. The post-processing stage is explained in the following sections.
• Figure 3: The architectures of the CNN-BiLSTM model. Each convolutional layer and LSTM layer was followed by a dropout layer, which is not shown in the figure. The hyperparameters of were tuned according to Table \ref{['tab:hyper']}.
• Figure 4: Some selected hyperparameters in each outer loop from CNN-BiLSTM in stage 1 (dataset 1+ dataset 2).
• Figure 5: The definition of IoU, TP, FP, and FN. There are eight cases shown in the figure. In case 4 and case 5, FN or FP depends on the length of the true and predicted segment. In case 7, if a true segment is predicted as some smaller segments, FP numbers increase. In case 8, if some separate true segments are predicted as one segment, FN numbers increase.