Position Control of Active Orthotic Glove using 3D-Printed Integrated Sensors and Voice Control for Enhanced Musculoskeletal Rehabilitation

This paper presents the design, fabrication, and evaluation of a compact 3D-printed, voice-controlled orthotic glove intended for musculoskeletal rehabilitation. The system addresses key limitations of existing assistive hand devices, such as bulkiness, and complex actuation, by integrating a lightweight rack-and-pinion actuation mechanism with embedded 3D-printed flexible sensors and offline voice control functionality. The actuation mechanism uses servo-driven rack-andpinion systems mounted on the forearm, eliminating the need for long external tendon routing or bulky pneumatic components. Real-time position control is enabled via piezoresistive TPU-based flexible sensors embedded in a glove, allowing closed-loop feedback for precise motion. Additionally, a convolutional neural network (CNN)-based voice recognition system was locally deployed to recognize predefined commands without reliance on internet connectivity or cloud services. Experimental validation demonstrated accurate finger movement tracking with an average angular error within <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\pm 3^{\circ}$</tex>. Comparative evaluation of bult in MATLAB speech models and Python revealed a comparable performance and improved efficiency. The Python implementation achieved a 93.77 % validation accuracy while reducing training time by 75 %. The proposed glove demonstrates a promising step toward accessible and user-friendly rehabilitation technologies, offering customizable, real-time control suitable for home or clinical use.