基于柔性光波导的感知袖套设计与研究

Design and Research on Sensing Sleeve Based on Flexible Optical Waveguide

摘要: 针对传统刚性传感器在上肢运动感知中信号失真与机械不匹配的瓶颈，本文设计并制作了两种分别以Dragon Skin 20和弹性软管为外壳、Clear Flex 30为核芯的柔性光波导传感器，利用形变引起的光能量损失实现灵敏传感。在此基础上，将两种传感器集成于感知袖套，构建包含前臂4 × 4交叉阵列与肘部条形传感器的分布式传感架构，并设计了光电转换与数据采集模块。通过采集16种姿态下11个通道的传感器信号，结合支持向量机(SVM)与随机森林算法进行训练与实时识别，两类模型均表现出较高的识别精度与良好的泛化能力。该系统在运动康复与人机交互等领域具有重要应用潜力。

Abstract: To address the limitations of traditional rigid sensors—namely signal distortion and mechanical mismatch—in upper limb motion sensing, this paper designs and fabricates two types of flexible optical waveguide sensors. These sensors feature Dragon Skin 20 and elastic tubing as outer casings, respectively, with Clear Flex 30 as the core, and utilize light energy loss caused by deformation to achieve sensitive sensing. Building on this, the two sensors were integrated into a sensing sleeve to construct a distributed sensing architecture comprising a 4 × 4 cross-array on the forearm and a strip-shaped sensor at the elbow, along with a photoelectric conversion and data acquisition module. By collecting sensor signals from 11 channels across 16 postures and training and performing real-time recognition using Support Vector Machines (SVM) and Random Forest algorithms, both models demonstrated high recognition accuracy and good generalization capabilities. This system holds significant application potential in fields such as sports rehabilitation and human-computer interaction.