Smartwatch-based Fine Vital Signs Extraction for Power Workers

The working environment in the electrical industry is complex, and monitoring the physiological state of workers is crucial for ensuring operational safety. However, photoplethysmography (PPG) signals collected by smartwatches are highly susceptible to motion artifacts in dynamic environments, compromising the accuracy of vital sign measurements. To address this issue, this paper proposes a lightweight motion artifact detection and removal method to enhance the reliability of PPG signals in dynamic conditions. First, a one-dimensional convolutional neural network (1D-CNN) is employed to efficiently detect motion artifacts in PPG signals. Then, accelerometer signals (ACC) are incorporated to extract motion-related information, and power spectral analysis is utilized to eliminate motion artifacts, thereby restoring high-quality PPG signals. Experimental results demonstrate that the proposed method effectively suppresses motion artifacts, improving the accuracy of key vital sign parameters such as heart rate and blood oxygen levels. This approach provides a reliable technical foundation for intelligent health monitoring of electrical workers.