Non-linear compensation of production inaccuracies and material drift by adjusting the sensor data fusion algorithms for shape sensing based on FBG-optical fibers

Shape sensing, where the shape of an object is estimated using fiber optical Fiber Bragg Grating (FBG) sensors, has gained increasing popularity in the last years. While the production process and the applications are different for the various research groups, the basic principle of shape sensing is the same: at certain cross-sections along the observed object, the information of three strain measurements is merged to one curvature information, and the information of these curvatures is then used to estimate the shape with various mathematical theories, depending on the application. Some effort has been made to calibrate and determine the accuracy of the shape sensor, but research on the influence of bad positioning of the Fiber Bragg Gratings seems relatively unattended. In this paper, one aspect of bad FBG positioning, namely inaccurate placement of the FBGs within one cross-section, and its influence on the reconstruction results is investigated. Furthermore, a modified approach for the reconstruction algorithm is presented, improving the reconstruction results for bad FBG placement compared to the conventional approach.