Workwear with Loosely Coupled IMU Sensors for Posture Classification During Assembly Tasks: A Pilot Study

Workplace risk assessment is essential to prevent overexertion injuries and ensure safe working conditions. In recent years, wearable sensors have been increasingly used to capture postures and assess physical risk factors in the workplace. However, many of these approaches are inconvenient for practical applications because the sensors need to be firmly attached to the body, which can be uncomfortable. In this context, a workwear garment with integrated loosely coupled sensors (Body Sensor Network: BSN) has been for reliable posture detection. A pilot study was conducted to evaluate the reliability of the BSN compared to a state-of-the-art system (Xsens) in terms of biomechanical joint angles. Six participants performed assembly tasks under experimental conditions. Kinematic data for posture classification was recorded simultaneously with the BSN and the Xsens systems. Kinematics were calculated using the OpenSense software tool and MVN Studio. Time-series and discrete joint angles values were compare to determine the validity of the loosely coupled approach. With a highest RMSD deviation of 25 and a lowest RMSD of 1.5 in the joint angles, the comparison of both systems shows clear differences. Nevertheless, loosely coupled sensors in workwear showed great potential for recoding postures at the workplace and could contribute to a safer and healthier work system design in the future.