Eigenstructure based towed array shape estimation using sources in unknown locations

The performance of direction-of-arrival (DOA) estimation algorithms depends on the knowledge of the array shape. However, the assumption of a straight line for a towed array is not always satisfied in practice. In this paper, we propose a new array shape estimation method based on one of the eigenvectors spanning the signal subspace. In our approach, we include the array deformation in the initial DOA estimation resulting in the algorithm converging rapidly and with good accuracy. Simulations demonstrated the efficiency of this algorithm and the performance is analysed by numerical experiments.