TY - JOUR AU - Liu, Lu AU - Liu, Hengxiang AU - Li, Shudan AU - Wei, Xinlu AU - Yao, Jiantao AU - Han, Bo AU - Zhou, Hao AU - Wang, Dianyu AU - Xu, Yundou AU - Jia, Xiaoli TI - High-performance, noninvasive flexible sensing array for gesture and handwriting recognition assisted by machine learning JO - Soft Science PY - 2026 VL - 6 IS - 2 SP - EP - 34 SN - ISSN 2769-5441 (Online) AB -

Wearable handwriting interfaces represent one of the most challenging frontiers in human–machine interaction as the precise decoding of subtle neuromuscular coordination is required. In this study, a machine-learning-assisted flexible force myography (FMG) array is presented for gesture and handwriting recognition. The system is constructed based on a 3 × 3 microstructured polydimethylsiloxane/multiwalled carbon nanotube array with a mass of 0.15 g and a thickness of 0.43 mm. The system is characterized by its high sensitivity (~1.364 kPa-1), low hysteresis [4.75% full scale (F.S.)], rapid response and recovery times (12 and 16 ms, respectively), and robust mechanical stability under bending, overload, and water-immersion conditions. Compared with conventional surface electromyography or electret-based FMG systems, quantitative mapping of the spatiotemporal deformation of wrist tendons is achieved, enabling the handwriting sequence, stroke count, and continuity to be decoded with an accuracy of 99.54% ± 0.16%. Mechanistic analysis indicates that pressure modulation patterns are governed by the coordinated activation of the coordinated activation of the flexor carpi radialis, flexor digitorum superficialis, and flexor digitorum profundus, and the flexor carpi ulnaris–flexor digitorum profundus groups, thereby providing a physiological basis for character-specific responses. This work extends FMG sensing from gesture-level motion capture to fine-grained handwriting interpretation, and a compact, noninvasive, and scalable platform is established for wearable handwriting recognition and human–machine interaction.

KW - Flexible sensor array KW - handwriting correction KW - human-machine interaction KW - gesture recognition KW - wearable device DO - 10.20517/ss.2025.143 UR - https://dx.doi.org/10.20517/ss.2025.143