Abstract
PURPOSE: The purpose of this study was to determine the ability of a novel shoe-based sensor that uses accelerometers, pressure sensors, and pattern recognition with a support vector machine (SVM) to accurately identify sitting, standing, and walking postures in people with stroke. METHODS: Subjects with stroke wore the shoe-based sensor while randomly assuming 3 main postures: sitting, standing, and walking. A SVM classifier was used to train and validate the data to develop individual and group models, which were tested for accuracy, recall, and precision. RESULTS: Eight subjects participated. Both individual and group models were able to accurately identify the different postures (99.1% to 100% individual models and 76.9% to 100% group models). Recall and precision were also high for both individual (0.99 to 1.00) and group (0.82 to 0.99) models. CONCLUSIONS: The unique combination of accelerometer and pressure sensors built into the shoe was able to accurately identify postures. This shoe sensor could be used to provide accurate information on community performance of activities in people with stroke as well as provide behavioral enhancing feedback as part of a telerehabilitation intervention.