Abstract
Vaidyanathan and Pal [IEEE Trans. Signal Process. 2011] proposed the use of Coprime Sensor Arrays (CSAs) to sample spatial fields using fewer elements than a Uniform Line Array (ULA) spanning the same aperture. A CSA consists of two interleaved uniform subarrays that are undersampled by coprime factors M and N. The subarrays are processed independently and then their scanned responses are multiplied to obtain a unaliased output. Although the CSA achieves resolution comparable to that of a fully populated ULA, the CSA beampattern has higher sidelobes. Adhikari et al. [Proc. ICASSP, 2013] showed that extending the subarrays and applying spatial tapers could reduce CSA sidelobes. This paper considers the problem of designing a CSA for the North Elba Sea Trial described by Gingras [SACLANT Tech. Report, 1994]. The experimental dataset consists of receptions recorded by a 48-element vertical ULA in a shallow water environment for two different source frequencies: 170 Hz and 335 Hz. This paper considers all possible coprime subsamplings for this array and selects the configuration that provides the best tradeoff between number of sensors and performance. Results are shown for both simulated and experimental data. [Work supported by ONR Basic Research Challenge Program.]