WING DYNAMIC SHAPE-SENSING FROM FIBER-OPTIC STRAIN DATA USING THE KALMAN STATE ESTIMATOR WITH PROCESS AND MEASUREMENT NOISE CHARACTERIZATION

A novel method for shape sensing based on strain data and the use of the Kalman state estimator is presented and studied experimentally. Several recent studies proposed different methods in which strain data, measured in optical fibers, can be used to estimate the deformed shape of a flexible wing. Fiber optic sensors typically provide accurate, high resolution, strain data, yielding detailed prediction of a wing’s deformed shape. However, in real life applications there could be moments in which the strain data is imperfect (for example, due to saturation or temperature effects), corrupted, or missing. The current study proposes to use a Kalman state estimator that weighs in strain-data and simulation output of an aeroelastic plant model to estimate the wing deformations. The method is demonstrated with a flexible wing model that is excited by prescribed control surfaces deflection in a wind-tunnel test. Strains over the front and rear wing spars are measured by Fiber Bragg Grating sensors, embedded in two optical fibers, and used to estimate the wing deformations. The latter are compared to wing deformations measured by a motion recovery camera system. Results show the advantages of the use of the Kalman state estimator when the strain data is corrupted. Additionally, with this approach, wing’s modal velocities are estimated together with the wing deformations, resulting in smooth and accurate modal velocities that are readily usable by the vehicle’s control system.