2D Sparse Array Selection via Deep Leaning

This paper proposes a simulated annealing (SA) assisted deep learning (DL) based 2D sparse array selection approach. A SA-based optimisation method is used to assist in generation and annotation of dataset. The antenna selection problem is formulated as a multi-classification problem, and a convolutional neural network (CNN) is used to solve it. The SA-based dataset generation method enables optimal sensor distribution and reduces computation complexity related with dataset generation. Numerical examples verify the effectiveness of the proposed method in terms of configuration of the realised sparse arrays and DOA estimation performance.