Abstract:Continuum robots in robot-assisted minimally invasive surgeries provide adequate access to target anatomies that are not directly reachable through small incisions. Achieving precise and reliable shape estimation of such snake-like manipulators necessitates an accurate navigation system, that requires no line-of-sight and is immune to electromagnetic noise. Fiber Bragg Grating (FBG) shape sensing, particularly eccentric FBG (eFBG), is a promising and cost-effective solution for this task. However, in eFBG sensors, the spectral intensity of the Bragg wavelengths that carries the strain information can be affected by undesired bending-induced phenomena, making standard characterization techniques less suitable for these sensors. We showed in our previous work that a deep learning model has the potential to extract the strain information from the eFBG sensor's spectrum and accurately predict its shape. In this paper, we conducted a more thorough investigation to find a suitable architectural design of the deep learning model to further increase shape prediction accuracy. We used the Hyperband algorithm to search for optimal hyperparameters in two steps. First, we limited the search space to layer settings of the network, from which, the best-performing configuration was selected. Then, we modified the search space for tuning the training and loss calculation hyperparameters. We also analyzed various data transformations on the network's input and output variables, as data rescaling can directly influence the model's performance. Additionally, we performed discriminative training using the Siamese network architecture that employs two CNNs with identical parameters to learn similarity metrics between the spectra of similar target values. The best-performing network architecture among all evaluated configurations can predict the shape of a 30cm long sensor with a median tip error of 3.11mm in a curvature range of 1.4m −1 to 35.3m −1 .

Machine Learning-Assisted Dual Fiber Bragg Grating-Based Flexible Direction Sensing

Structural Instability-Enabled Mechanical Sensors Using Fiber Bragg Grating

Shape sensing using neural network of radial basis function based on Fiber Bragg Grating

Shape Sensing of Optical Fiber Bragg Gratings Based on Deep Learning

Sensitivity-enhanced two-dimensional bending sensor based on single-mode fiber measurement

Simultaneous Estimation of Shape and Force along Highly Deformable Surgical Manipulators Using Sparse FBG Measurement

Deep Learning-Enabled Two-Directional Stretchable Strain Sensor Based on a Single Multimode Fiber

Wearable Flexible Optical Fiber Sensors for Human Motion Intention Recognition

Femtosecond Laser Inscription of Fiber Bragg Grating in Twin-Core Few-Mode Fiber for Directional Bend Sensing

Machine learning-augmented multi-arrayed fiber bragg grating sensors for enhanced structural health monitoring by discriminating strain and temperature variations

Self-Healing Fiber Bragg Grating Sensor System Using Free-Space Optics Link and Machine Learning for Enhancing Temperature Measurement

Fibre Bragg Grating Inscription into a Seven Core Fibre and Its Application As a Vector Bending Sensor

Distortion Tolerant Method for Fiber Bragg Grating Sensor Network Using Estimation of Distribution Algorithm and Convolutional Neural Network

Research on vector bending sensors based on taper-drawn seven-core fiber Bragg grating

Robust Three-Dimensional Shape Sensing for Flexible Endoscopic Surgery Using Multi-Core FBG Sensors

Fast and High-Precision Shape Sensing based on Dual-Comb Fiber Bragg Grating Array Demodulation

A Three-Dimensional Fiber Bragg Grating Force Sensor for Robot

3-D Source Location by Neural Network for FBG Acoustic Emission Sensors

FBG Tactile Sensing System Based on GAF and CNN

Wearable and Skin‐Mountable Fiber‐Optic Strain Sensors Interrogated by a Free‐Running, Dual‐Comb Fiber Laser

In-Fiber Thermally Diffused Coupler and Fiber Bragg Grating Inscribed in Twin-Core Fiber for Sensitivity-Enhanced Vector Bending Sensing