Fiber Shape Sensing Using Deep Neural Networks

A Video by Samaneh Manavi, Planning & Navigation | The MIRACLE Project | Department of Biomedical Engineering | University of Basel
Edge-FBGs are a new generation of highly flexible fiber-based shape sensors designed by Fraunhofer institute in Germany, in which the FBGs are inscribed on the edge of the fiber’s core. In Edge-FBGs, the light intensity at the Bragg wavelengths contains the strain information at sensing nodes and can be interrogated with simple setups, including an uncooled SLED and a micro spectrometer.
However, such sensors are sensitive to changes of the spectrum profile caused by undesired bending-related phenomena. As the existing theories cannot accurately predict the spectrum profile in curved optical fibers, changes in the initial intensity that each Edge-FBG receives are not precisely known. These uncontrolled variations cause inaccuracies in shape predictions and make standard characterization techniques less suitable for Edge-FBG sensors.
We developed a new technique for modelling these sensors, in which deep neural networks are trained to directly extract the shape information from the full Edge-FBG spectrum. This way, the effect of bending-related phenomena that influence the spectrum profile are also taken into account. These findings open new possibilities for real-time shape sensing applications based on low-power devices and inexpensive interrogators. The designed model can predict the shape of a 30 cm long sensor with mm-range accuracy.
More: dbe.unibas.ch/...