Geometrically Optimized FDM-Printed Conductive TPU Bend Sensors for Hand Rehabilitation

Abstract

Flexible resistive bend sensors are essential for monitoring human movement in smart rehabilitation and soft robotics. However, widespread adoption is currently hindered by a trade-off between the high cost of metal-film technologies and the performance degradation (significant hysteresis and non-linearity) of low-cost carbon/polymer composites. This study presents a geometrically customizable bending sensor fabricated from conductive thermoplastic polyurethane (TPU) using Fused Deposition Modeling (FDM) technology as an accessible alternative to commercial sensors. By parametrically optimizing physical dimensions—including trace width, layer thickness, and pattern geometry—the sensors were tailored to achieve target resistance values within a target window of 20–50 kΩ (achieved: ~44 kΩ nominal) for specific finger-joint applications. Electromechanical characterization revealed a negative gauge factor (GF), where resistance decreases upon bending or elongation due to conductive pathway formation and densification within the polymer matrix. This behavior cannot affect sensor operation, and required bend-resistance responses were acquired using geometrical optimization. To compensate for inherent viscoelastic-induced hysteresis and non-linear behavior, a third-degree polynomial modeling approach was implemented. This modeling approach yielded a coefficient of determination (R2) of approximately 0.90. Compared to standard commercial sensors, the proposed FDM-printed design successfully overcomes geometric limitations while offering a cost-effective, high-performance solution for tailor-made wearable technologies and smart rehabilitation gloves.