Optimal sensing matrix design with hardware friendliness

Compressed sensing, which is mainly composed of cheap sensing and costly recovery, has received much attention recently in the aspects of theoretic studies and wide applications. Prior researches reveal that the sparsity of a signal is upper bounded by the reciprocal of mutual coherence between the sensing matrix and sparsifying basis. Thus, some methods have been proposed to study the design of optimal sensing matrix from the viewpoint of reducing mutual coherence values. Nevertheless, the hardware friendliness of optimal sensing matrices is relatively ignored. In this paper, we focus on this issue and present a simple strategy to achieve optimal sensing matrix with entries of 0 and ±1 only. Extensive simulations demonstrate the usefulness of our design.