I. Introduction
Nonuniform linear arrays provide the ability to estimate the direction-of-arrival (DOA) of more sources than the number of physical sensors [1]–[6]. Recently, a new structure of nonuniform linear arrays, known as co-prime arrays, has been proposed [7], [8]. A co-prime configuration comprises two undersampled uniformly spaced subarrays with co-prime spatial sampling rates. Co-prime configurations have many advantages over other popular nonuniform configurations, including minimum redundancy arrays (MRA) [9], minimum hole arrays (MHA) [10], and nested arrays [11]. For a given number of physical sensors, MRAs and MHAs require an exhaustive search through all possible combinations of the sensors to find the optimal design [12], [13]. On the other hand, the positions of the sensors constituting the co-prime configuration have closed-form expressions. Although the same is true of nested arrays, the elements of one of the subarrays constituting the nested structure are closely separated, which may lead to problems due to mutual coupling between the sensors. Co-prime arrays reduce the mutual coupling between most adjacent sensors by spacing them farther apart [7]. Because of all of the aforementioned characteristics, co-prime arrays are finding broad applications in the areas of communications, radar, and sonar [14]–[20].