This paper presents the development process and the evaluation of compliant Fiber Bragg Gratings (FBG) based, normal and shear force sensing elements for biomechanical measurements with their effective novel calibration methodology and preliminary experimental results. The sensing elements are intended for human machine interaction such as in transfemoral prosthetic interface. The sensing elements are designed to utilize the generated deflections due to the applied force to stretch the optical fiber when a normal or shear force is applied on the sensor’s top surface. The performance of the sensor is evaluated through a series of experiments including both dynamic and static loading conditions. The experimental results show that the fabricated sensing elements have the ability to measure applied normal and shear force. Due to the deformable material used, the sensor exhibits slightly nonlinear behaviour between force and deformation. This has been addressed using a novel calibration procedure composed of a linear model to characterize the main sensor data and a nonlinear estimation model based on a nonlinear autoregressive exogenous (NARX) model to simultaneously estimate the errors from the input wavelength data in real-time. The results achieved from the proposed calibration method have revealed an improvement from an R-squared value of 93% to 100% when compared to a data obtained using a linear least squares method.

中文翻译：

---

用于生物力学应用的基于 FBG 的兼容剪切力和法向力传感元件的开发和表征

本文介绍了基于柔性光纤布拉格光栅 (FBG) 的、用于生物力学测量的法向力和剪切力传感元件的开发过程和评估，以及它们有效的新型校准方法和初步实验结果。传感元件用于人机交互，例如在经股骨假肢接口中。当法向力或剪切力施加在传感器的顶面上时，传感元件被设计成利用由于施加的力而产生的偏转来拉伸光纤。通过包括动态和静态负载条件在内的一系列实验来评估传感器的性能。实验结果表明，所制造的传感元件具有测量外加法向力和剪切力的能力。由于使用了可变形材料，传感器在力和变形之间表现出轻微的非线性行为。已使用由线性模型组成的新校准程序解决了这个问题，该程序由表征主要传感器数据的线性模型和基于非线性自回归外生 (NARX) 模型的非线性估计模型同时实时估计输入波长数据的误差。与使用线性最小二乘法获得的数据相比，从提出的校准方法获得的结果表明，R 平方值从 93% 提高到 100%。已使用由线性模型组成的新校准程序解决了这个问题，该程序由表征主要传感器数据的线性模型和基于非线性自回归外生 (NARX) 模型的非线性估计模型同时实时估计输入波长数据的误差。与使用线性最小二乘法获得的数据相比，从提出的校准方法获得的结果表明，R 平方值从 93% 提高到 100%。已使用由线性模型组成的新校准程序解决了这个问题，该程序由表征主要传感器数据的线性模型和基于非线性自回归外生 (NARX) 模型的非线性估计模型同时实时估计输入波长数据的误差。与使用线性最小二乘法获得的数据相比，从提出的校准方法获得的结果表明，R 平方值从 93% 提高到 100%。