GTac-SE: A Low-Density Intention Sensing for Forearm Exoskeleton With Shear Force Sensors

This study explores the use of GTac-SE shear force sensors for detecting forearm rotation in upper-limb assistive exoskeletons for patients with arm weakness. The purpose is to evaluate these sensors as a reliable alternative to conventional methods, like electromyography (EMG) and inertial measurement units (IMUs), in real-world scenarios. Methods involved two experiments comparing GTac-SE sensors with EMG and IMUs. The first experiment assessed their ability to detect small forearm rotations mimicking weak patient motions, while the second evaluated sensor robustness during proximal joint (elbow) motions. Results showed that EMG struggled to detect subtle forearm rotations due to signal overlap between movement and rest, whereas IMUs and shear force sensors performed better. EMG was also sensitive to proximal joint movements, making intention detection difficult. While IMU gyroscopes displayed clear motion patterns, they were affected by other body movements, leading to unreliable joint angle estimations. In contrast, GTac-SE sensors, housed in a brace designed to isolate forearm rotation, outperformed both EMG and IMUs. GTac-SE signals were consistent across subjects with minimal interference during elbow flexion/extension. Additionally, GTac-SE required only simple moving-average filtering, while EMG and IMU signals needed complex filtering. In conclusion, GTac-SE shear force sensors provide a direct, reliable, and practical solution for intention sensing in daily-use exoskeletons. They outperform EMG and IMUs, particularly in scenarios where multijoint tasks are required for patients with weak arms.