NP-doped Fiber Smart Tendon: a Millimeter-Scale 3D Shape Reconstruction with Embedded Distributed Optical Fiber Sensor System

This paper presents the development of a Nanoparticles (NP)-doped Smart Tendon for three-dimensional (3D) shape reconstruction applications. The proposed smart composite structure is based on the combination of NP-doped optical fibers immersed into a polyurethane (PU) matrix with a silicone cover. In the proposed NP-doped optical fiber, the interaction between optical signals and nanoparticles mainly occurs in the mode evanescent field, which leads to the increase of backscattered optical power. In this context, the polarization-assisted Rayleigh Backscattering (RBS) signature is analyzed as a function of different angles at predefined rotation planes. The preliminary analysis shows a correlation between the return loss, spectral shift, s- and p-polarization components and the angles at different rotation planes. Then, the NP-doped optical fiber is integrated in the smart tendon for the mechanical characterization tests, where the viscoelasticity of the multifunctional system is analyzed as a function of different frequencies up to 50 Hz, which demonstrated an increase on the stiffness as a function of the frequency. Moreover, a feed-forward neural network (FFNN) is applied on the input data (namely spectral shift, return loss, s- and p-polarization components) and resulted in the XYZ components estimation with mean relative errors below 3% (maximum error around 5%), considering a positioning range from -100 mm to 100 mm. Therefore, the proposed approach is a feasible method for novel 3D shape reconstruction using a single optical fiber with millimeter spatial resolution that can be used in applications ranging from structural health monitoring, tendon-driven actuators and in aerospace..