Circular sparse rulers based on co-prime sampling for compressive power spectrum estimation

One of the main challenges in cognitive radio systems is the development of fast algorithms for spectrum sensing of wideband signals. To alleviate sampling requirements, several strategies have been proposed, relying on the fact that only covariance information is of interest. Under this premise, this paper solves the design of low rate covariance samplers by developing a multicoset sampling strategy based on co-prime samplers, achieving higher compression ratios than previous solutions. The designed sampling patterns constitute a family of circular sparse rulers, which can be implemented with only two analog-to-digital converters working in parallel, and operating at sampling frequencies significantly lower than the Nyquist rate.