Joint Array and Spatial Sparsity Based Optimisation for DoA Estimation

Traditional spatial sparse techniques for DoA estimation are implemented by full arrays. However in practice, it is desirable to use as few sensors as possible to reduce the cost for manufacturing the array or to counter against sensor failure. As a result, joint optimisation of sparse array and spatial sparsity becomes an ideal alternative. In most of existing methods, these two kinds of sparsity are studied separately. This paper proposes a joint model which achieves source detection in a subset of space using partial array sensors. The core idea is to use the weight coefficients obtained in sparse array optimisation to scale the model for the sparse reconstruction based DoA estimation. Compressive sensing based optimisation is used for both steps. The numerical results of DoA estimation for both stationary source and moving source are used to demonstrate the feasibility of this joint model.