Shape Sensing for Continuum Robots Using FBG Sensors Array Considering Bending and Twisting

摘要

Continuum robots have been widely used in industrial and medical environments due to their dexterous abilities. However, the absence of robotic shape information limits their control accuracy and interaction ability in unstructured environments. In particular, sensing the shape of the continuum robot is challenging when the robot is subjected to material twist. This article presents a fiber optic shape sensor and related shape reconstruction algorithm to measure the curvature and material twisting angle simultaneously. Meanwhile, a calibration procedure in the initialization step is designed to identify the configuration parameters of the shape sensor. Then, the shape reconstruction performance of the robot under twisted and untwisted conditions is evaluated by conducting several different sensing shape tasks. The experimental results show that the mean distal position error is 1.26% of the length, and the mean Fréchet distance is 1.64% of the length. The results prove that our method can effectively evaluate the continuum robotic shape information and show the feasibility of an application. Besides, a sensing system is designed to monitor the shape of a continuum robot in real-time.