On the solution of inverse obstacle acoustic scattering problems with a limited aperture

We present a computational methodology for retrieving the shape of a rigid obstacle from the knowledge of some acoustic far-field patterns. This methodology is based on the well-known regularized Newton algorithm, but distinguishes itself from similar optimization procedures by using (a) the far field pattern in a limited aperture, (b) a sensitivity-based and frequency-aware multi-stage solution strategy, (c) a computationally efficient usage of the exact sensitivities of the far-field pattern to the specified shape parameters, and (d) a numerically scalable domain decomposition method for the fast solution in a frequency band of direct acoustic scattering problems.