Unsupervised Learning of Wearable Sensor Data for Assessing Upper Limb Function

Assessment of upper limb function is integral to both rehabilitation and clinical care, guiding tailored treatment plans and monitoring patient progress. Recent advances in wearable sensor technology and unsupervised learning offer promising avenues for objective and continuous evaluation of upper limb function. This review paper offers an in-depth exploration of the application of unsupervised learning techniques to wearable sensor data for this purpose.Beginning with an examination of the limitations inherent in traditional assessment methods, the review highlights the growing demand for more sophisticated approaches capable of capturing the nuanced movements of the upper limb. Wearable sensor technologies, including inertial measurement units (IMUs), electromyography (EMG), and force sensors, are introduced for their potential to provide real-world movement data. Furthermore, the review elucidates the principles of unsupervised learning, focusing on clustering algorithms and dimensionality reduction techniques as key tools for analysing sensor data without the need for labelled training samples.A thorough review of existing literature showcases the diverse applications of unsupervised learning approaches in upper limb function assessment, ranging from motion pattern recognition to gesture classification and motor performance evaluation. Case studies and research findings demonstrate the effectiveness of these methods in identifying abnormal movement patterns, monitoring rehabilitation progress, and offering objective feedback to patients and clinicians.Addressing challenges such as data variability, model interpretability, and integration into clinical practice, the review underscores the need for further research and development in this field. It concludes by emphasizing the transformative potential of unsupervised learning and wearable sensor technology in advancing upper limb rehabilitation and clinical assessment.In summary, this review paper serves as a comprehensive resource for researchers, clinicians, and healthcare professionals interested in utilizing unsupervised learning techniques and wearable sensor data for objective upper limb function assessment, with implications for personalized rehabilitation strategies and improved patient outcomes.