An Integrated Fabric‐Pneumatic Soft Sensing System for Multi‐Site Muscle Activity Decoding

Force myography (FMG) decodes movement intent from muscle deformation rather than bioelectrical signals, avoiding electrode–skin coupling issues that constrain surface electromyography. Most FMG implementations rely on rigid force‐sensitive resistors with limited conformability. Here, we present an integrated fabric‐pneumatic wearable sensing system combining a soft textile chamber with a selected internal support architecture for multi‐site muscle activity decoding. The chamber, fabricated from thermoplastic polyurethane‐coated braided nylon, converts tissue deformation into air pressure changes with no electronics at the sensing site. Among four tested internal support geometries, the dense honeycomb configuration produced the largest pressure response and was selected for the subsequent wearable experiments. The chamber showed similar responses under controlled ±15 mm lateral offsets, while a preliminary single‐participant assessment showed temporal correspondence between pneumatic pressure and sEMG ( r = 0.89). Evaluation across three anatomical sites in four tasks with 16 participants yields within‐session accuracy of 75.8%–85.4% for classification and R 2 = 0.890 for continuous elbow tracking. Leave‐one‐subject‐out evaluation showed task‐dependent cross‐subject performance, with 75.0% accuracy for the three‐class static‐posture formulation and R 2 = 0.732 for continuous elbow‐angle estimation. The system results support the feasibility of fabric‐pneumatic transduction as a complementary mechanical sensing approach for wearable muscle–machine interfaces.