Abstract: The portstarboard ambiguity in the conventional single towed linear array sonar is one of the most deceiving obstacles which exists in the way of development of spatial spectrum estimation. This paper proposes a spatial spectrum reconstruction estimation method based on sparse Bayesian learning using multiple signal models of the maneuvering towed linear array. Firstly, the signal overcomplete sparse representation model of the maneuvering towed linear array is established. Then, based on the principle of sparse Bayesian learning, the sparse characteristics of target angle is described implicitly by hypothesis of hierarchical prior. Finally, the change of spatial spectrum is modeled using a hiddenMarkov model, the objective law of slowlyvarying of spatial spectrum is applied to the calculation of probability density of the signal hyperparameters and the sparse reconstruction model based on multiple array signal models is established. Simulation and sea trial results demonstrate that the proposed algorithm has evident advantages in ambiguity suppression ratio and accuracy of reconstruction and achieve superior spatial spectrum estimation for towed linear array.

Key words: towed linear array, spatial spectrum estimation, sparse Bayesian learning, target portstarboard ambiguity, expectation maximization