Abstract
The continuous static contraction of the upper limb during overhead work often leads to progressive muscle fatigue, which is the main cause of work-related musculoskeletal disorders (WMSDs) and reduced operational safety. Traditional fatigue detection methods cannot capture subtle early fatigue characteristics under dynamic working conditions, and sensors are difficult to conveniently deploy in harsh environments. This study constructs an upper limb fatigue detection model based on wearable interface pressure signals and proposes a spatiotemporal dual-entropy fusion strategy. Sample entropy evaluates temporal muscle movement regularity, with a 10% incremental rate-of-change threshold for significant fatigue identification, superior fatigue sensitivity, and anti-interference capability. Information entropy reflects the spatial dispersion of pressure amplitude, adopting a 5% rate-of-change threshold for early fatigue warning. Combined with sEMG comparative verification and nonlinear fitting analysis, a hierarchical monitoring framework is established: single-index abnormality indicates slight fatigue, while dual-entropy synchronous elevation represents severe neuromuscular fatigue. The proposed method overcomes the limitations of single-index evaluation, accurately identifies multiple fatigue levels, and reliably monitors fatigue in real time, providing effective technical support for ergonomic monitoring and occupational safety protection in overhead operations.