Vision Transformer-based Local Physical Fatigue Assessment using Electromyography (EMG) Signals for Construction Worker Health Monitoring

Construction workers frequently endure high levels of physical strain, which can lead to fatigue and long-term musculoskeletal disorders.Traditional methods of fatigue assessment often fail to provide timely and accurate diagnoses, particularly in dynamic work environments.In this paper, we present a Vision Transformer-based approach for assessing localized physical fatigue using Electromyography (EMG) signals.This method leverages the high learning capacity of Vision Transformers to capture both spatial and temporal features from EMG data, enabling precise detection and prediction of fatigue.By focusing on specific muscle groups that are prone to fatigue during construction tasks, our model delivers detailed insights into the onset of fatigue at a localized level.Experimental evaluations conducted on roofing tasks targeting the biceps brachii and brachioradialis demonstrate that the proposed method can predict fatigue to a localized level.The results demonstrate the potential of this approach for enhancing on-site health monitoring systems and establishing a foundation for proactive interventions aimed at reducing the risk of fatigue-induced injuries among construction workers.