Shape Sensing of Hyper-Redundant Robots Using an AHRS IMU Sensor Network

The paper proposes a novel approach for shape sensing of hyper-redundant robots based on an AHRS IMU sensor network embedded into the structure of the robot. The proposed approach uses the data from the sensor network to directly calculate the kinematic parameters of the robot in modules operational space reducing thus the computational time and facilitating implementation of advanced real-time feedback system for shape sensing. In the paper the method is applied for shape sensing and pose estimation of an articulated joint-based hyper-redundant robot with identical 2-DoF modules serially connected. Using a testing method based on HIL techniques the authors validate the computed kinematic model and the computed shape of the robot prototype. A second testing method is used to validate the end effector pose using an external sensory system. The experimental results obtained demonstrate the feasibility of using this type of sensor network and the effectiveness of the proposed shape sensing approach for hyper-redundant robots.

Influence of Two-Plane Position and Stress on Intensity-Variation-Based Sensors: Towards Shape Sensing in Polymer Optical Fibers

Shape reconstruction is growing as an important real-time monitoring strategy for applications that require rigorous control. Polymer optical fiber sensors (POF) have mechanical properties that allow the measurement of large curvatures, making them appropriate for shape sensing. They are also lightweight, compact and chemically stable, meaning they are easy to install and safer in risky environments. This paper presents a sensor system to detect angles in multiple planes using a POF-intensity-variation-based sensor and a procedure to detect the angular position in different planes. Simulations are performed to demonstrate the correlation between the sensor’s mechanical bending response and their optical response. Cyclic flexion experiments are performed at three test frequencies to obtain the sensitivities and the calibration curves of the sensor at different angular positions of the lateral section. A Fast Fourier Transform (FFT) analysis is tested as a method to estimate angular velocities using POF sensors. The experimental results show that the prototype had high repeatability since its sensitivity was similar using different test frequencies at the same lateral section position. The proposed approach proved itself feasible considering that all linear calibration curves presented a coefficient of determination (R2) higher than 0.9.

Fiber Optic Distributed Sensing Network for Shape Sensing-Assisted Epidural Needle Guidance

Epidural anesthesia is a pain management process that requires the insertion of a miniature needle through the epidural space located within lumbar vertebrae. The use of a guidance system for manual insertion can reduce failure rates and provide increased efficiency in the process. In this work, we present and experimentally assess a guidance system based on a network of fiber optic distributed sensors. The fibers are mounted externally to the needle, without blocking its inner channel, and through a strain-to-shape detection method reconstruct the silhouette of the epidural device in real time (1 s). We experimentally assessed the shape sensing methods over 25 experiments performed in a phantom, and we observed that the sensing system correctly identified bending patterns typical in epidural insertions, characterized by the different stiffness of the tissues. By studying metrics related to the curvatures and their temporal changes, we provide identifiers that can potentially serve for the (in)correct identification of the epidural space, and support the operator through the insertion process by recognizing the bending patterns.