The study of human activity recognition concentrates on classifying human activities and the inference of human behavior using modern sensing technology. However, the issue of domain adaptation for inertial sensing-based human activity recognition (HAR) is still burdensome. The existing requirement of labeled training data for adapting such classifiers to every new person, device, or on-body location is a significant barrier to the widespread adoption of HAR-based applications, making this a challenge of high practical importance. We propose the semi-supervised HAR method to improve reconstruction and generation. It executes proper adaptation with unlabeled data without changes to a pre-trained HAR classifier. Our approach decouples VAE with adversarial learning to ensure robust classifier operation, without newly labeled training data, under changes to the individual activity and the on-body sensor position. Our proposed framework shows the empirical results using the publicly available benchmark dataset compared to state-of-art baselines, achieving competitive improvement for handling new and unlabeled activity. The result demonstrates SAA has achieved a 5% improvement in classification score compared to the existing HAR platform.
7 April 2021 in International Journal of Precision Engineering and Manufacturing
BACKGROUND: Individuals with gait disturbances, such as that post-stroke, are discharged home to undergo outpatient rehabilitation. Rehabilitation in the community is not as effective as that in hospital, due to long travel times and short program duration. OBJECTIVE: This study analyzed rail unit structure, with the aim of assisting home indoor assistive mobility system (HIAMS) development, allowing patients to undergo gait-related rehabilitation training at home. METHODS: The HIAMS consists of a mobile rail running around the whole room, a turn-table for movement between rails, and a weight-supporting component. Structural analysis was performed using the Abaqus/CAE solution (Version 6.14, Dassault systems, Inc.) to verify device safety, according to the load applied to the rail and turn-table units. The load was applied vertically at 150 kg to reflect the weight of potential users. RESULTS: Structural analysis was performed on the weight-supporting components, which was consist of turn-table case, bearing components (center, left), connective bracket and rail rollers. The safety factors of each components were estimated as 1.31, 5.39 (bearing, center), 8.45 (bearing, left), 1.43 and 3.61 in sequence. CONCLUSION: We demonstrated a safety factor of ⩾ 1.3 for the key system units, suggesting this technology is safe for use in the home rehabilitation training of individuals with gait impairment post-ICU stay.